Podcasts about 2X

Could your rental property earn 2X more each month? Kathy Fettke sits down with mid-term rental expert Jesse Vasquez to explain how investors are increasing cash flow by renting to corporations, traveling professionals, and insurance companies instead of relying on traditional leases or nightly Airbnb stays.   Learn how to find these opportunities, which markets work best, and why mid-term rentals are one of today's fastest-growing investment strategies.  

What if your 100 people could feel like 300? That's the bet Robin Daniels, Chief Business Officer at Zensai and author of The Human Success Playbook, is making - and it's working. His company doubled revenue over three years and has been named a great place to work three years running. He joins us to break down how to unlock the potential already sitting inside your organization, lead through genuine uncertainty, and build B2B content that people can actually put to work. About our guest, Robin Daniels Robin Daniels is Chief Business Officer at Zensai, a learning and performance platform competing in a market dominated by Oracle, SAP, and Workday. He previously served as CMO at Matterport and WeWork and held senior marketing roles at Salesforce, LinkedIn, and Box. He recently co-authored The Human Success Playbook - a narrative business book that follows a fictional CEO who refuses a board's demand to cut 40% of the company and sets out to prove there's a better way. The book comes with an AI-powered interface where readers can ask the authors questions and apply every framework directly to their own organization. What B2B Marketers Can Learn From The Human Success Playbook The biggest AI opportunity isn't efficiency - it's multiplication. Robin starts every conversation the same way: do you see AI as a way to cut costs and get more efficient, or do you see it as a way to unlock insane potential in the team you already have? He's made his choice - and built a company around it. Ian puts the math plainly: if you have 100 people, the far more exciting challenge is figuring out how to make them feel like 200 or 300 and generate the revenue to prove it. Robin connects it to something every leader already knows but rarely acts on: "I truly believe everyone has something they're really creatively good at. But if you can actually find time to focus on that, it can lead to new ways of growing your business that maybe you hadn't even thought of before." The companies that win the next five years won't be the ones who cut hardest. They'll be the ones who figured out how to multiply. You can't shrink your way to greatness. Robin credits Tom Peters for the line and means it completely. His argument isn't about shaming companies that make cuts - sometimes it's necessary to survive. The argument is about where the real upside lives: investing in people, giving them new tools and skills and goals, and then letting them find growth paths nobody had imagined yet. "I've never seen a company shrink their way to greatness," he says. "You have to grow." Ian adds the frame: the goal isn't to replace anything - it's to generate more revenue. "How are we going to 2X our revenue?" is a fundamentally more energizing question than "how do we get more efficient?" And it tends to produce more interesting answers. Lead with a vision even when you don't have all the answers. Robin is direct about what employees actually hate most: not bad news, not hard feedback, but uncertainty. "If you don't know what you're working towards, it's really hard to do great work." In The Human Success Playbook, the fictional CEO Evan has eight months to show a different path or get fired - and the whole tension of the story is whether he has the courage to set a direction without knowing exactly where it leads. Robin's advice for real leaders in that position: paint the vision, commit to it, and stay flexible enough to adjust as the world changes. And then, with characteristic directness: "Being a leader has always been tough. Suck it up, buttercup." The future of B2B content is a framework people can make their own. Robin didn't just write The Human Success Playbook - he and his co-authors built an AI-powered interface on top of it so readers can chat with the book and apply its frameworks to their specific company. A 500-person manufacturer in Germany gets different advice than a mid-size bank in the Midwest. Ian connects it to a bigger shift in how B2B content works: "You're building to say: here's the framework, your AI already knows where you want to go as a business, layer this on top and execute." The white paper people download and never read is being replaced by something that actually does the work. Robin is explicit about why: competing against Oracle, SAP, and Workday with a smaller budget means doing something genuinely different. The book is that thing. "I truly believe everyone has something they're really creatively good at. But if you can actually find time to focus on that, it can lead to new ways of growing your business that maybe you hadn't even thought of before." - Robin Daniels Time Stamps [4:00] Meet Robin Daniels, CBO at Zensai and Author of The Human Success Playbook [4:39] Why Robin Wrote the Book - and Why Now [6:00] The Multiplier Choice: Unlock Potential vs. Cut Costs [8:00] It's Just a New Technology - Adapt and Grow [10:00] Robin Has Seen This Before: Internet, Cloud, Mobile, AI [11:00] The Book's Setup: CEO Evan, a 40% Cut Demand, and a Better Way [12:00] The Villain Isn't AI - It's Ignorance and Fear [14:00] Internal vs. External Motivation - and Why Inner Fire Wins [15:00] The Mountain Analogy: What Leaders Owe Their Teams in Uncertain Times [17:00] How to Lead When You Don't Have the Answers [19:00] Suck It Up, Buttercup - Leadership Has Always Been Hard [22:00] Your Whole Job Is to Make Stuff and Sell It [25:00] What Taylor Swift and the Vinyl Revival Teach Us About AI and Creators [31:00] What Should the CEO Actually Do? Reorganize Around Projects, Not Hierarchy [33:00] Invest in People's Capabilities - Measure Productivity, Not Efficiency [35:00] The Human Success Score - a Sleep Score for Your Organization [38:00] Replace "Replace" With "Multiply" in Your Vocabulary [40:00] "You Can't Shrink Your Way to Greatness" - Tom Peters [42:00] The Courage to Invest in People When Everyone Else Is Cutting [44:00] The Future of Books: Build a Framework, Let AI Make It Specific [49:00] Demo: Chat With the Book at Zensai.com [51:00] The "Talk to a Human" Button - the Future of B2B Content [52:00] Final Thoughts + Connect with Robin on LinkedIn Links Connect with Robin on LinkedIn: https://www.linkedin.com/in/robindaniels/ Learn more about Zensai: https://www.zensai.com Chat with The Human Success Playbook: https://zensai.com/report/thehumansuccessplaybook The Human Success Playbook on Amazon: https://www.amazon.com About Remarkable! Remarkable! is created by the team at Caspian Studios, the premier B2B Podcast-as-a-Service company. Caspian creates both nonfiction and fiction series for B2B companies. If you want a fiction series check out our new offering - The Business Thriller - Hollywood style storytelling for B2B. Learn more at CaspianStudios.com. In today's episode, you heard from Ian Faison (CEO of Caspian Studios) and Meredith Gooderham (Head of Production). Remarkable was produced this week by Meredith Gooderham, edited by Jon Goldberg, and our theme song is "Solomon" by FALAK. Create something remarkable. Rise above the noise. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Send us Fan Mail“Enjoy your self-discovery process now, not later, to welcome self-esteem and self-determination.” -Matteo TuriExclusive Insights from This Week's EpisodesSimilar profits can produce radically different valuations. Valuation Expert Matteo Turi reveals why systems, intellectual property, leadership depth, and transferability create the gap between a 2X and 10X business.Episode Highlights[00:07:00] Why revenue growth does not automatically create valuation[00:10:00] The three pillars inside the High Valuation Triangle[00:14:00] How founders can turn personal expertise into transferable intellectual property[00:22:00] Why control obsession destroys value during the scale-up phase[00:27:00] The emotional execution mistakes that quietly derail strong companies[00:31:00] Why a founder's absence can become one of the company's most valuable assets[00:41:00] How identical financial statements can hide a valuation gap from 2X to 10XFull show notes, transcript, and resources for this episode:https://podcast.deepwealth.com/565The Deep Wealth PodcastMost entrepreneurs do not fail.They just carry too much for too long.The business grows. Pressure grows faster. Profits get harder to predict. Decisions cost more energy. Over time, focus slips and health takes the hit.The Deep Wealth Podcast and Deep Wealth Mastery are built from real experience. We're the only system based on a 9-figure exit. This system exists because guessing gets expensive.

Watch the full episode on YouTube:We first covered Baseten last year when DeepSeek mania was at peak hype. Now they have raised a monster $13B round and become one of the new cohort of AI Infra decacorns that are (with Nvidia, Intel, and the semis complex) chief beneficiaries of the Inference Inflection. We return to Baseten at the peak of the 2026 edition of Open Weights debate. Ali has published a viral breakdown of Kimi K3:And since you last saw him, Philip has spoken at AI Engineer and written the definitive book on Inference Engineering spotted all over SF:Three years ago, inference engineering barely existed as a category.Today, it is one of the most critical disciplines in AI. Inference engineering inherently tackles a different question than standard model training: “How do you turn those weights from training into a product that is fast, reliable, and affordable at scale?” Focusing on these creates an entirely new optimization problem.In one recent GLM-5.2 experiment, quantizing more of the model actually preserved its benchmark quality while increasing throughput by 20%, because the errors introduced in different layers could cancel each other out.Inference is no longer just the final step after training. It is becoming its own engineering discipline, with its own research problems, infrastructure, and increasingly specialized roles.In this episode, Baseten's Philip Kiely and Ali Taha join swyx and Vibhu to explain what actually happens after a new open model is released and what it takes to turn “we generated a token” into a fast, reliable, production-ready API.We go deep on cache-aware routing, disaggregated prefill and decode, quantization, speculative decoding, KV-cache movement, model parallelism, GPU kernels, and the race to make frontier models up to 10× faster. Philip and Ali explain why inference optimizations can still produce gains of 20%, 100%, or even 200%; how quantization errors can cancel one another out; why identical weights can behave differently across clusters; and how Baseten grafted a Kimi vision encoder onto GLM-5.2 without changing the underlying language model.The conversation then expands beyond LLMs into NVIDIA Dynamo, mega kernels, Rubin, AI-specific chips, local inference, video generation, diffusion versus autoregressive models, and the enormous compute barrier to generating coherent long-form video. Finally, we explore the convergence of training and inference, continual learning through persistent KV cache, and the emerging loop where models help optimize the infrastructure that runs them.We discuss:* What happens when a 200,000-token request enters an inference system* Cache-aware routing and reusing previously computed KV cache* Why prefill and decode are increasingly handled by different GPUs* When dedicated deployments become cheaper and more reliable than shared APIs* How speculative decoding uses a smaller model to accelerate a larger one* Tool calling, structured outputs, and what LLMs actually do* What it takes to support a new open model on day zero* Grafting Kimi's vision encoder onto GLM-5.2* Retrofitting inefficient model layers with components from other architectures* Why models sometimes collapse into repeating the same token* How hardware, kernels, and race conditions create nondeterministic failures* Preserving model fidelity while making inference faster* How quantization errors can cancel each other out* Why inference optimizations still deliver gains of 20%, 100%, and 200%* How optimized serving can make a model up to 10× faster* NVIDIA Dynamo, KV-aware routing, and distributed model serving* Speculative decoding the speculative decoder* Why local AI is about making models less dumb while data-center AI is about making them less slow* Tensor, expert, and pipeline parallelism across GPUs* Hardware-aware model design, auto-tuning, and the case against mega kernels* Rubin and why inference is becoming a systems problem* Whether modern GPUs are evolving into programmable AI ASICs* Why enormous models like Kimi K3 require GB300-class hardware* Why open-source video generation still trails Veo, Kling, and other closed models* The quadratic attention bottleneck behind long-form AI video* Autoregressive video, real-time generation, and compounding quality drift* Why future video systems may combine autoregressive and diffusion architectures* Training for inference and inference for training* Continuous post-training, deployment, evaluation, and improvement loops* How GLM-5.2 helped optimize the kernels serving GLM-5.2 itself* Why faster networking could unlock dramatically faster decoding* Continual learning, KV-cache compaction, and persistent model memoryShow Notes* How to build a day-0 API for Kimi K3* 22580: From GPT2 to Kimi3, ExplainedPhilip Kiely* LinkedIn: https://www.linkedin.com/in/philipkiely* X: https://x.com/philipkiely* Inference Engineering: https://www.baseten.co/inference-engineering/Ali Taha* LinkedIn: https://www.linkedin.com/in/aliestaha/* X: https://x.com/waterloointernTimestamps00:00:00 Introduction and the 200K-Token Prompt00:03:18 Dedicated Deployments, Speculative Decoding, and Tool Calling00:11:26 Launching Production-Ready Open Models00:19:06 Model Retrofits, Failure Modes, and Nondeterminism00:28:22 Quantization and Canceling Errors00:32:15 The Race to 10× Faster Inference00:40:48 Dynamo, Speculation, and Local vs. Data-Center AI00:50:18 Model Parallelism, Auto-Tuning, and Mega Kernels01:00:55 Rubin, GPUs vs. ASICs, and Custom AI Chips01:10:03 Giant Models and the Limits of GPU Memory01:12:42 AI Video, Quadratic Attention, and Autoregressive Generation01:21:47 Audio, Images, and Diffusion Models01:27:32 Training, Self-Optimizing Models, and Continual Learning01:40:06 Closing ThoughtsTranscriptIntroduction: Baseten, Waterloo Intern, and Inference EngineeringSwyx [00:00:00]: Okay, we're here in the studio with Philip, old friend from Inference Engineering, the book, as well as Baseten and everything that you've done, you and I have done before, as well as Ali. Welcome.Ali [00:00:15]: Pleasure to meet you.Swyx [00:00:15]: Waterloo intern.Ali [00:00:16]: Waterloo intern, always.Swyx [00:00:17]: When did you get “Waterloo intern” as a handle?Ali [00:00:19]: As a handle? Oh.Ali [00:00:20]: I think the rebranding happened mid-March. When I saw it was open, I was like, “I have to take it. Up for grabs.”Philip [00:00:26]: The problem is that Ali is really good at his job and is not gonna be an intern much longer.Philip [00:00:30]: So we have to figure out who's gonna get the handle.Ali [00:00:33]: Well, I'll pass the torch over to the next intern.Swyx [00:00:34]: Oh, okay. It can be, like, you just pass it to another Waterloo grad.Ali [00:00:37]: To another Waterloo intern. No, bruh.Philip [00:00:39]: Yeah.Ali [00:00:39]: Intern.Swyx [00:00:40]: Intern, yeah.Ali [00:00:40]: And no.Philip [00:00:41]: You gotta get an intern from Waterloo.Ali [00:00:42]: Yeah, I've gotta get an intern from Waterloo.Swyx [00:00:44]: Right.Ali [00:00:44]: But they have to follow the path.Swyx [00:00:45]: Oh, it could, but it could come from Baseten, so it's like whoever Baseten gets from Waterloo.Ali [00:00:48]: Right.Swyx [00:00:49]: Has the title of Waterloo.Ali [00:00:50]: It stays in the ecosystem.Philip [00:00:51]: Exactly.Ali [00:00:52]: Halfway through the internship, you either get it or you're out.Philip [00:00:55]: You should also do, like, a big graduation ceremony where you change the handle.Ali [00:00:59]: Just say it.Philip [00:00:59]: For everybody.Swyx [00:01:00]: You guys are good at ceremonies, clearly. We had a nice launch of the book, very successful. But before we get into all that, I wanna start off with a fun question for you. Okay, you're an expert inference engineer. What happens when I send a long query, say two hundred thousand tokens into Baseten's inference? What's the process of query through GPU model routing, balancing, all that? What is all the stuff that we don't think about?Long Context Requests, KV Cache, and Cache-Aware RoutingPhilip [00:01:26]: With a long query specifically, the first thing that I'm gonna ask is, “Have you sent me this query before, or at least part of it?” and I really hope you have, because it's gonna be a lot easier for me and a lot cheaper for you. So the first thing that we're gonna look at is some cache-aware routing, where we're going to see, we probably have a number of instances, a number of replicas up serving whatever model you're hitting. We want to send this one to something with, number one, available prefill workers, and number two, ideally some cached input already there so that we can skip prefill on at least part of these two hundred thousand tokens. If you're doing two hundred thousand tokens, it's probably coding or a multi-turn agent or something where you would expect to have that cached. If you don't, we're gonna have to send it to a prefill worker. We've at least on certain models disaggregated prefill and decode, so you're going to have one set of GPUs that's solely going to process the input, create the KV cache, and get you your first token, and then that's going to be passed over to a separate set of GPUs, which is going to run decode. We're going to iteratively make those tokens. We're probably going to have some speculator model in front of that. I'm going to assume that you're doing coding, and because of that, our speculator model, which assumes you're doing coding, is gonna have a high draft token acceptance rate. If I'm wrong and you're asking me to summarize every Harry Potter book, it's gonna be slower. And then we stream that output to you and account for it, charge you, a couple of pennies and say, “Hey, would you like to send another one?”Swyx [00:03:04]: Except Baseten doesn't charge by pennies.Philip [00:03:07]: Well, yeah, we charge. I'm assuming that we're talking about the public model APIs. If you are setting up a dedicated deployment, then yeah, it's not pennies.Public APIs vs. Dedicated DeploymentsSwyx [00:03:18]: Yeah, one of the key differentiators when I was talking with Baseten initially was that people who want very high volume just need to rent by the box, ‘cause then it's up to you to figure out how to saturate the box.Ali [00:03:31]: And more often than not, it's, like, way cheaper if you're pushing, like, millions of tokens per hour, if you just pay per hour instead of pay per token.Philip [00:03:37]: Yeah, they do. I think that we've increasingly seen a lot of demand for the pay per token APIs, just because everyone wants to try open models, and then once they find a use case that's really sticky, then they move over to dedicated.Swyx [00:03:51]: Is there a best practice on when it's time to swap over?Philip [00:03:54]: Couple reasons. Yeah, reliability, that's a big one, right?Ali [00:03:57]: Like, if they have a very specific use case, they want you to train something specifically for them, like they want their own spec dec, for instance, for their own traffic.Swyx [00:04:04]: Spec dec is speculative decoding.Speculative Decoding and Custom SpeculatorsAli [00:04:05]: Speculative decoding, yeah.Swyx [00:04:07]: You have to explain.Ali [00:04:07]: Sorry. Like, speculative decoding is like, if you have a huge model, right? And so the model is going to be generating one token at a time every single turn, every single forward pass. So we attach, like, this little, like, parasite, like this layer that goes on top of the model, and this model just has to predict. It does three very fast autoregressive forward passes, and it will predict, like, three certain tokens, and then you do one forward stage over the entire original model in order to see if those predictions were correct or not, and then you accept them or you reject them. Now, this draft model is traffic specific, so if you, like, Philip said, if you're summarizing Harry Potter books, I can train exclusively that draft model on Harry Potter books, and I can guarantee you that I'm gonna accept the three tokens every single time. And so with that case, I increase your decode speed. I wouldn't be able to provide this to you if you're a shared endpointSwyx [00:04:53]: YeahAli [00:04:53]: ‘cause I have no idea if you're doing Harry Potter, if you're doing coding, if you're doing English. We don't know. Also, there was a thing in the book that mentioned that if they really cared about a specific threshold, chapter four, I think. Do you remember that?Philip [00:05:06]: Yeah. The things that you can do is you can set a specific, like, batch sizing, a specific, like, parallelism strategy if you're trying to optimize for, like, throughput versus latency. You can. Maybe a NVFP4 quant doesn't pass your benchmarks and you wanna run a model at higher precision, you could do that. There's just a bunch of reasons why you might wanna have your own endpoint and the biggest one, of course, just being, like, you don't have to deal with someone else doing a hundred million tokens of benchmarking traffic at the endpoint when you happen to be trying to serve your users.Swyx [00:05:40]: Yeah. I think one thing that is. That is a classic journey. Like, it's people is asking the, what happens when you type Google into the browser. Tool calling, is that just, you're generating JSON or is there more complication beyond that?Tool Calling, JSON, and Structured OutputsAli [00:05:58]: Certain customers that we have, they have their own post-trained models, and so they demand a tool calling that's not just, like parse a file or go find the weather. It's something that's very specific and you have to do post-training on this. And if the post-training on the model is not good or if the quantization after the post-training to get the inference to be fast, the model will struggle reading the JSON file and reading the tool calling. But it doesn't require its own like sandbox. It's not like it's going to use that tool calling to like escape a sandbox or like it doesn't have to be contained. It can just be a normal dedicated deployment. The challenge with tool calling more and more seems to be that the companies want certain tool calling which is a very sensitive thing to train. And because you're dealing with all of the JSON outputs, if it doesn't like close the end of the request in a very certain manner, you end up with a model that did the tool calling and like the thinking and so as a result of that, it didn't see the result and just hallucinated the result as it decoded. That seems to be the most challenging thing with tool calling, not really the sandboxes model.Philip [00:06:56]: Yeah, that's a challenge on the training side and then on the inference side, there's work that you can do to scope the possible output. So we published this at this point close to two years ago, the solution to this problem which is you make a state machine and you use that to constrain the output to a specific format. So this is the structured output problem. If you remember backSwyx [00:07:27]: Yeah, the specific grammar is,Philip [00:07:29]: Yeah, exactlySwyx [00:07:30]: GML had this thing.Philip [00:07:31]: Yeah. So it's like the old-school “make sure this is only JSON”, return only JSON orSwyx [00:07:38]: YeahPhilip [00:07:38]: Grandma's gonna die type of prompts.Swyx [00:07:39]: Is it BNF grammar? At some point OpenAI had released a thing that was like, yeah, if you want to constrain your output, write BNF grammar, back as NOR.Philip [00:07:47]: In our inference system, it's just a specified output format. And you get the guarantee that your output's gonna be structured along that format. And so applying that to tool calls can like help cut down on. You can still call the wrong tool or call no tool. It doesn't solve the certainty problem but it at least solves the output structuring problemSwyx [00:08:10]: YeahPhilip [00:08:10]: Within tool calls.Swyx [00:08:12]: And MCP is just another form of tool, right.Philip [00:08:14]: Yeah, exactly.Swyx [00:08:15]: As far as there's no special thing there.Philip [00:08:16]: The thing I'm always like explaining to people is the LLM is not capable of doing anything. It's only capable of making suggestions of what to do and then if those suggestions are formatted in a certain way and applied to a system that knows what to do with them, then an action occurs.Swyx [00:08:32]: Yeah. Part of the fun stuff is, this is solved outside of tool calling too. Like in an agent loop if the output is not correct or you're right, like reasoning, tool calling was done in the reasoning trace, just be like, “Oh, I don't know what to do. Let me just try again.” And it might get there after a few tries. And on your point of training, sometimes this is harder in smaller models, so you don't have the same exact quality outputAli [00:08:56]: Right.Swyx [00:08:57]: When you just swap from a big model, right?Ali [00:08:59]: Yeah. I will say that, before, I think we need to go back to inference engineering proper.Ali [00:09:04]: But, I had expected that something would replace JSON because it's hard to stream JSON ‘cause JSON must be complete and you must have open and close brackets and everything. So it's hard to parse something or validate something while it's being streamed. So people invented all sorts of things that are like, I forget the name of some of these alternatives, but it's something like TOML, something like YAML. But JSON seems to be dominant still.Philip [00:09:30]: The JSON outputs aren't that long, right? Like you could have a long-- ‘cause tool calls also contain the arguments in them and perhaps for a certain tool you might pass like a very long argument. But my impression of the median tool call is that it's a relatively small number of tokens, right? So I would expect that speculators are generally fairly good at something as formatted as JSON. And so you would have like a pretty fast decode step there and that the streaming wouldn't be as valuable, but maybe I'm wrong about that.Ali [00:10:02]: I think you're also bounded by the software or that the model is gonna integrate with if the software is built with JSON for the tool calls or if the company that you'- if your customer says that this is how our software works and our tools are interfaced with JSON, you can ask them to like, change their software and say like, “Yeah, this is gonna be better for the model.” but like with the right training shouldn't be that much of a difference. Also more profitable if it outputs more tokens probably.Swyx [00:10:25]: Depends on your business model.Swyx [00:10:27]: It really depends. But I will say that, as a writer with like experience a lot with generated output, I do try to move from text to JSON text which is very long JSON, right? Like there's paragraphs in every field because I'm trying to structure it, right?Philip [00:10:44]: Right.Swyx [00:10:44]: I want you to first make factual statements, then make opinions then make bullet point summaries, have dates, have entity references have your sources for references, all these things. Anyway, so these are things that like I think people who really experiment with structural output have to really care about. But, let's, let's recurse up the stack a little bit. Before we started recording, you mentioned something really cool, which is that there's a lot of engineering that-- inference engineering that goes on when a new model provider releases a new model, right? So let's call it GLM-5.2, Kimi K3. I had previously assumed, especially if it's like, well, GLM 5 to 5.1 to GLM-5.2, like that you've supported them before. Is it that much work?What It Takes to Support a New Open ModelAli [00:11:26]: It's a lot of work.Swyx [00:11:28]: Yeah. Okay. So like, a lot of people, all you guys, right whenever a new model launch like, people rush to say like, “Oh, Hugging Face supports this, Fireworks supports this, Spacetime supports this,” and I'm like, “Yeah, of course we support it.” But what goes into that? What goes intoPhilip [00:11:40]: I think it's more than just support it too, right? It benefits the consumer a lot. Like I think it was with Kimi K2.5 or GLM-5.2 the latest, there was an inference war, right? X provider is at 90 tokens a second. The next day we're at 150. The nextSwyx [00:11:55]: I kinda kicked that off with the GLM-5.2.Swyx [00:11:58]: I wrote a Twitter article about. It got like half a million views,Ali [00:12:02]: Based on being numberSwyx [00:12:03]: YeahAli [00:12:04]: Or it's for something else.Swyx [00:12:05]: Yeah. Which,Ali [00:12:06]: Oh my GodSwyx [00:12:07]: Which then got everyone really excited about, hey, how can we, bend tracks a little bit further and,Philip [00:12:14]: There's a difference between support the model, as in I can make a token out of this model, and support a model, as in I have a production-ready API from this model.Philip [00:12:26]: Getting to the point of I can make a token out of this model is not that hard because generally the, open source inference engines, vLLM, SGLang of the world oftentimes even receive weights ahead of time, maintainers do, or the people making the model merge PRs to ensure support. So you generally can, just get it working on the standard open source stack without too much pain in most cases. The challenge is, every inference company is gonna have own proprietary stack. Some open source components, some in-house stuff. And for any arbitrary model, there's going to be some new stuff. Sometimes you get lucky, like K, two five to two six was, like, pretty similar.Quantization, Speculators, and Production ReadinessAli [00:13:16]: Yeah. It was pure continued post-trainingPhilip [00:13:18]: YeahAli [00:13:18]: If I remember correctly.Philip [00:13:19]: Even in those cases, there's still stuff you have to do. You have to redo the quantization work. You're taking the model from. Generally, these models are not released in NVFP4, and we want them to be in NVFP4 for maximum Blackwell compatibility. So we have to perform that quantization, and, calibrate the quantization to make sure that we're not causing any regression in the model's intelligence. And then we also have to train the speculator, as we've talked about. Generally, we have. We have ZDR, zero data retention on our model APIs, so we don't know exactly the traffic that people are sending us, but we know what's popular. We know that coding use cases are popular. We know that agents, agentic use cases are popular. So we can get public data sets that are representative of that traffic and train general speculators. Now, with speculators today, you need to train the speculator using the base model itself because you're getting hidden states out of the model from running inference on these specific prompts, and that is the training data you use to create the speculator. So there's that process which you need the real model weights for. And then there's of course just the process of, standing up all the infrastructure behind it, loading all this stuff, testing it. And then when there's a new model with a newer architecture, I think that, like, the DeepSeek models tend to be the most challenging as they have, like, the most novel architectural stuff going on, model after model. But every new model has something. Kimi K2 had. Oh, sorry, GLM-5.2 hadAli [00:14:53]: Sparse attention.Philip [00:14:54]: Yeah,Ali [00:14:54]: YeahPhilip [00:14:54]: the DSA.Ali [00:14:55]: Right. Which is brought from DeepSeek.Philip [00:14:57]: Yeah. AndAli [00:14:59]: So you can copy-paste then?Philip [00:15:01]: It kindAli [00:15:01]: I don't know how this works.Philip [00:15:02]: So, like we had to, like, build support for that into our runtime. And you're right, like it is really interesting the way that all of these open source labs borrow from each other. For example, like GLM-5.2 doesn't have vision. So something that, Haley, a guy on our team, if we could take a look at this, he, like, grafted the Kimi vision encoder onto GLM-5.2.Retrofitting Vision into GLM-5.2Ali [00:15:27]: We'll be training the projector.Philip [00:15:28]: Exactly. So if you think about, like, the encoder, there's the encoder, which is the part that looks at the image and turns it into latent information, and then there's the projector which likeAli [00:15:38]: You can say latent space. It's okay.Philip [00:15:41]: And then there's the projector that maps it onto, the model itself, and then there's the model weights. You don't wanna mess with the model weights because you run a chance of making the model dumber at something else for the purpose of giving it vision. So instead, Haley started with just a projector, which is only a handful of millions of parameters.Ali [00:16:02]: That would be, yeah.Philip [00:16:02]: Yeah.Ali [00:16:03]: Can you show the training one?Ali [00:16:04]: Like the way it groksPhilip [00:16:05]: YeahAli [00:16:06]: Very interesting.Philip [00:16:06]: And maybeAli [00:16:07]: That right therePhilip [00:16:07]: Maybe Ali, you should take it from here. You've got a betterAli [00:16:10]: Ooh, double the sandPhilip [00:16:11]: Understanding of this than I do.Ali [00:16:11]: Yeah. You can see, like, he. The way he trained this is really cool. At the beginning, he was training it using just like, “Here's a picture of a mountain. Can you describe what's in this mountain?” And that caused it just like the first, learning walls. Like here you can see this all we're trying to teach it is to translate the encoded. Like it's already taken the encoder from Kimi K. It's taken the image. It'Philip [00:16:31]: Yeah. FrozenAli [00:16:31]: FrozenPhilip [00:16:32]: With adapter.Ali [00:16:32]: Exactly.Philip [00:16:33]: Yeah.Ali [00:16:33]: So the brain is frozen and the eyes are frozen. It's just we're tryingPhilip [00:16:37]: AlignAli [00:16:38]: Interconnect between the eye and the brain, right? So the projector. And so you take the tokens and then he's like, “Oh, can you describe what's in this image?” And he's like, “Oh, it's a mountain,” or it's a person or it's a human, whatever the case is. But that didn't cause complete understanding. So he changed it such that every image was associated with a data set of questions. Like, does this image have a white male? Does this image have birds in the top corner? Does this image have a scientist in it? All of that stuff. And it would have to answer questions correctly. And using not just training on describing an image, but being able to answer question, another question, answer over time. Like you can see the grokking, which is like genuinely insane, that retrofitting vision into a large LLM can learn to that extent. And even for images that it doesn't perform well on, for instance, if you ask it a picture of like Stephen Hawking, “Who is this?” Maybe it doesn't get it, but it will say something like, “This is Albert Einstein.” Like it still understandsPhilip [00:17:25]: Close enoughAli [00:17:26]: That this is a scientist who is a man who has, some significant achievements, all that stuff. So that's like really cool.Philip [00:17:32]: Yeah. So, we've covered Hao Tian before, who the author of the LLaVA paper that did this, a while ago. And I think that's very foundational work for anyone who hasn't done vision work before.Ali [00:17:41]: Same with the CLIP and MetaCLIP, where you go from just captioning to building out questionsPhilip [00:17:47]: RightAli [00:17:47]: Off the image and how much better you can get performance.Philip [00:17:50]: Right. Right. Right. Yeah. But what's, what's so exciting about this is if you look at a model like this. Now, this is a little bit more of a research project. It's not. It got to 56% on MMLU Pro, I think. So not quite frontier. But if you're running this model, you haven't suffered any loss on your GLM-5.2 quality. If you don't have an image, it'll just behave exactly the way it used to. And ultimatelyAli [00:18:14]: Which in the inference code you literally do not include the other part, right?Philip [00:18:18]: Yeah. You would just skip the encoder if you don't have an image input.Ali [00:18:22]: Okay.Philip [00:18:22]: Just confirming.Philip [00:18:23]: YeahAli [00:18:23]: Does it affect a lot on the overall inference side? Like you're not adding much, you're adding a very small vision encoder. These are typically likePhilip [00:18:30]: They're super fineAli [00:18:31]: Less than a billion parameters, right?Philip [00:18:32]: Yeah. It's, - There's a little bit less standardization among vision encodersSwyx [00:18:37]: YeahPhilip [00:18:37]: So the support matrix can be a little bit, sparser. But overall, yeah, it's a pretty, it's a pretty minor component of the overall system. And ultimately what you get out of the system is all of a sudden you have Kimi Vision, GLM weights, and DeepSeek attention all in one model.Open Source Model Grafting and Franken-MergesPhilip [00:18:56]: And that's, I think, a lot of the power and beauty of open source, is that you can take all of these different components and combine them together into a system that's better than anyoneSwyx [00:19:05]: YeahPhilip [00:19:05]: Can be individually.Swyx [00:19:06]: People used to say that you would also do Franken-merges where you would take likePhilip [00:19:10]: YeahSwyx [00:19:10]: Layers from each model.Swyx [00:19:11]: Does anyone do that anymore?Ali [00:19:13]: Well, to your point previously when you were mentioning like, the work that goes into supporting a model when it first comes out, like GLM-5.2 or MiniMax M3 or whatever the case is. Sometimes you do have to like, you do have to switch out some things. Like, for instance, the MiniMax M3 head uses full attention, and with full attention you end up with this like insane bottleneck in spec dec ‘cause you're doing auto-regressive token generation for three tokens, and you're doing this like N squared over all of the tokens that are in your sequence. Your KV cache is like very large because it's not sparse, it's not top K. So we find it better to like, okay, we're gonna replace this, we're gonna replace this layer with a layer from another model that's using like GQA, for instance. And then just with the right training, you can get it to have the same acceptance rate. So it is very possible to retrofit layers from other models and very much needed. If a layer is like inefficient, the training just becomes the challenge, like how do you ensure that you train it properly? Which again to your earlier point is like the mesh between training and inference. As in like you need very good training in order to do fast inference. That's like, I feel like more and more becoming true.Swyx [00:20:21]: Yeah. Anything else on the support side when you say like get it to fully production ready?Loop Detection, Race Conditions, and Non-DeterminismPhilip [00:20:26]: Yeah. I think that there's also a question of just, we can test a model to a pretty extensive degree, but we're trying to get it out quickly and then you see a bunch of other people test it and you get interesting results. There was an issue with, GLM briefly where we had some like mode collapses where it would just output the same token over and over again for certain prompts on certain temperatures. Like once you expose an endpoint to the real world, there's going to be, so many more varieties of things given to it that you're able to, discover and patch things. So it's not just a, day zero process, it's then like for the first week, for the first month, if a model remains popular, like how do you both fix bugs and then continue to push the envelope on performance?Ali [00:21:21]: What do you mean you don't want your model outputting S?Swyx [00:21:24]: Is there loop detection on that stuff, by the way? It still happens like quite a lot, which is surprising.Ali [00:21:30]: We have like we, in our endpoint, like if a model was to output the same token like four plus times, we just cut the generation. We say like, “Oh, sorry, this-- Like try again,” or like we will reprocess the request. ‘Cause we know then, like if it, like if, yeah, it's four times the same token, it's probably collapsed.Swyx [00:21:45]: Yeah. Is there a way to opt out in case I really want that?Ali [00:21:48]: You want that?Ali [00:21:50]: I think there's a way that we have to handle it. I'm not exactly certain, but I feel like in certain models, like when they output something like you can imagine, like a table for instance, and so they want, they wanna draw like 12 dashes and 12 dashes. Yeah, I think there's a way for that to happen. I think we only do it on certain tokens. Like we exclude certain special characters.Swyx [00:22:07]: Yeah.Ali [00:22:07]: So we only do it on like certain like S is the most common almost. GLM-5.2Swyx [00:22:11]: OhAli [00:22:11]: And I think it was DSV 4 as well. Like you'd just have like looping issues where like you literallySwyx [00:22:17]: ItAli [00:22:17]: Just have like S.Swyx [00:22:18]: Yeah. Is there a special, something special about S? No, just randomlyAli [00:22:21]: It just seems to be the one token involved.Swyx [00:22:23]: Yeah. And it'Philip [00:22:24]: Is thereSwyx [00:22:24]: And it's only temperature 0Ali [00:22:27]: NoSwyx [00:22:27]: Even at other temperaturesAli [00:22:27]: Even at like 0.9 or whatever, it will still, it will still collapse.Swyx [00:22:30]: That's weird, right?Ali [00:22:30]: It's, it is an inference problem to be honest, like a software problem. Like oftentimes, the image you run will-- like NVIDIA will release an image for instance, and if we will upstream the changes from their latest TensorRT-LLM image into our stack, we'll find that it fixes it. Or oftentimes this will only happen in an inference engine that you're using like SGLang. But if you were to switch to vLLM, that isn't the case. So it seems to be like an extremely like deterministic software issue and not really a model issue. It's not like a weights problem. Like I'- we'll say like, “Oh, it's a problem with the quant. We did PTQ wrong,” right? But that isn't, that doesn't make sense because the same weights used with a different inference engine does not repeat the problem. And sometimes it's, the kernels that are being used in the backend have like these very subtle sometimes race conditions, where if you were to use this model hosted on one cluster, you will never get this problem.Swyx [00:23:19]: Oh my God.Ali [00:23:19]: But if you host it on a different cluster, you will. And the reason is the KV cache transfer from a node to node in that one cluster is using a slower interconnect than the node to node in another cluster. So that exposes the race, whereas in another cluster it doesn't. So then you end up just like, okay, this model is not gonna be hosted on this cluster. We're gonna host it on, another cluster because that cluster exposed that problem. But then it ends up with like, okay, is it the software? Is it the model weights or is it the hardware?Swyx [00:23:42]: There is a thing about this with temperature 0 still not being deterministic, right?Ali [00:23:46]: Right.Swyx [00:23:46]: Mostly because of hardware. Even at temperature 0 same model, you won't always get the same output.Swyx [00:23:52]: Even-- But I'm surprised by the race condition one because, I thought PyTorch was a graph that like guarantees that you at least, execute things in the right order.Ali [00:24:02]: Well, yeah, true. Like I'm not, I'm not saying that there is. Like well, you have things like PTL optimizations where like you can start a kernel before the end of the previous kernel, and that's like ‘cause you want to do that because there'sSwyx [00:24:12]: It's like pipeliningAli [00:24:12]: Expense. Exactly.Swyx [00:24:13]: Yeah.Ali [00:24:13]: But it'- But you don't do it cleanly. Like you overlap a little bit of the execution. No, it is very possible that the kernel itself, like that one block that is supposed to be running in this instance of time, that kernel itself has a race condition. For instance, like a missing barrier. Like often if you're designing a kernel and you want it to make it to be very fast, if you don't test it extensively, you'll, you'll have certain threads access data points from registers before they've been written to by other threadsSwyx [00:24:36]: YeahAli [00:24:36]: For example, because like your barrier is wrong or your synchronization was wrong. But yeah, like the testing itself is very difficult in those like, andSwyx [00:24:42]: And there's no like borrow checkerAli [00:24:45]: What does that mean?Swyx [00:24:46]: Like Rust. Like the. If you're trying to have like memory safety It sounds like a comparable problem.Ali [00:24:52]: Well, yes, but you're working in CUDA, right, NVIDIA GPUs. Like- You just need a higher level language like modular Maybe that's what modular is supposed to do. I don't know.Quantization Quality and Vendor FidelityVibhu [00:25:00]: How do you see keeping quality of the model? So you talked about all these steps of, okay, you gotta do quantization, train your own speculative decoderAli [00:25:07]: RightVibhu [00:25:07]: Run on different hardware. Looking at other model providers, okay, you kicked off a inference speed race on the consumer end. What goes into keeping quality the same across them, right? Sure, you can run benchmarksAli [00:25:22]: YeahVibhu [00:25:22]: But, like, how do you determine how much quantization are there standards? What goes intoPhilip [00:25:27]: There's a few things on quality. Most inference optimizations are lossless. KV caching, for example. You are just recomputing or preventing recomputing the same values. Speculation, of course, if a draft token is wrong, it gets rejected. The main lossy optimization is quantization. And that really comes down to, number one, data format, number two, which parts of the model you choose to quantize, which layers, and number three, like doing a lot of calibration on the quantized weights, to ensure that you're preserving all the outliers. There's other tricks that you can do, though. A big one is long context, ‘cause one thing you asked at, right at the beginning is, “Oh, what's gonna happen if I send a 200,000 token request in?” So with a long input sequence, you need to, store a lot more information. You need to process a lot more tokens. And so even if a model has a context of a certain length, you might, as an inference provider, choose to build an API with a shorter context length, and of course a full length one as well. Because if someone doesn't need the full million token context, for example, you can get them better performance. I don't know if that's exactly like quality of the model. The way that I think about quality is to what degree are we faithfully serving the original model? If you think of a golden implementation of a model that performs exactly the way the model is designed to perform, I think of quality as how close are we getting to that, 100% fidelity of the model.Philip [00:27:13]: You can also, of course, think about quality from the training side and how do you push yourself past 100%. But when I think about purely inference optimizations, it's getting faster while staying as close to that 100% fidelity mark as possible. And certainly our standard internally is that, like you should not be able to tell the difference between our API and a, official API. I think Kimi in particular does a good job of vendor benchmarking hereAli [00:27:41]: YesPhilip [00:27:41]: Where they haveAli [00:27:42]: They released an actual vendor benchmark.Philip [00:27:43]: Exactly, yeah.Ali [00:27:44]: ‘Cause they accused, some people, Amazon? There was some provider that was not doing very well on Kimi's benchmark.Philip [00:27:50]: Yeah.Philip [00:27:51]: So, with Reflect we probablyVibhu [00:27:52]: This was a long time ago, right?Philip [00:27:54]: No.Ali [00:27:54]: Yeah, like threeVibhu [00:27:55]: They alsoAli [00:27:55]: Four, five months agoVibhu [00:27:57]: This also happened with, I don't remember which model, but they pulled out quite a few, and then they started a whole chart about this. It might have beenPhilip [00:28:03]: Kimi Vendor Verifier.Ali [00:28:04]: Yeah.Philip [00:28:05]: Yeah.Ali [00:28:05]: Yeah, ‘cause you, ‘cause you'd be pissed, right? Like if you'Philip [00:28:07]: Yeah.Ali [00:28:07]: If like if I'm a consumer and I'm using like Amazon's endpoint for instance, and I've used Kimi and I'm like, “Oh my God, like this is bad,” I'm not gonna say, “Oh, Amazon quantized the model in a bad way.” I'm gonna say, “Oh, Kimi sucks.” Right?Philip [00:28:17]: Yeah.Ali [00:28:17]: So it seems like that makes sense.Philip [00:28:19]: Yeah, they care. They care.Vibhu [00:28:21]: Justifiably.Ali [00:28:21]: Yeah, justifiably.Vibhu [00:28:22]: This is probably a stupid question, but just checking, has anything improved from main quantization?Philip [00:28:28]: Yeah.Vibhu [00:28:28]: Like, is quantization always strictly worse?Ali [00:28:30]: Well technicallyVibhu [00:28:32]: NoAli [00:28:32]: It's a lossy. QuantizationPhilip [00:28:33]: YeahAli [00:28:33]: Is a lossy, it's a lossy implementation.Philip [00:28:36]: Speed improvesVibhu [00:28:36]: Speed improves.Ali [00:28:37]: It the number, likeVibhu [00:28:38]: No, I' always look for inverse scaling laws.Philip [00:28:40]: Yeah.Ali [00:28:40]: Yeah.Vibhu [00:28:40]: This is something I learned from Noam Brown, where like things that normally act in one direction sometimes do.Philip [00:28:45]: Well, technically when you run a benchmark, because these models are deterministic, sometimes your,Ali [00:28:52]: YeahPhilip [00:28:52]: NVFP4 quant is like, two basis points higher than yourAli [00:28:56]: No, it's noise. It's noise.Philip [00:28:57]: Yeah, exactly. I'm like, yeah, it's, it's within. That's why I always say within margin of error.Philip [00:29:01]: And I stopped saying that because everyone assumes that what is, well, within some margin of error, we're barely inside of that to the worst, so we're saying. But yeah, sometimes it's just like, gives you a higher output score. But like Ali said, that's noise. To my knowledge, you're not necessarily making the results better. You're just trying to, again, like keep your fidelity as close to 100% to the original model.Layer Selection, KL Divergence, and Better QuantizationAli [00:29:27]: There is, to your point, research that we did on MP. I don't know if you are able to pullPhilip [00:29:31]: YeahAli [00:29:32]: A tweet we did. One of our research interns, Joshua, I think it's a tweet on how we have 20% better quantized GLM-5.2 than NVIDIA. Essentially what we found throughout like this month research is, okay, quantization is a lossy. It's. You're compressing the data from, occupying 16 bits to occupying, four bits, for instance. And so you're losing some information, and you're trying to minimize that. And so when I say that I'm gonna quantize the model, my job becomes how do I find the layers that I can quantize, and how to find the layers to not. For instance, with image models, I don't quantize modulation layers, and I don't quantize out projections because those two are. Like out projection is what you see as the user. Modulation is what the model sees or understands. Right, exactly. And so to his paper, do you have the. It doesn't have the. Yeah. It's a long paper. I don't know if I can findVibhu [00:30:25]: If there's a part to search or it's probably in the thread.Ali [00:30:28]: It's probably in the thread.Vibhu [00:30:29]: Yeah.Ali [00:30:29]: But the long and the short is it is very possible that quantizing more of the model makes the results. Like if I have a model that I quantize layers one, five, and 10, and another model where I only quantize layers one and It is possible that the model in which I quantized more information is going to perform better because the quantization errors have canceled out. And so what Joshua showed in his mathematical proof where he had like a verifier in, is that you can predict which layers are going to have quantization errors that will cancel out with each other, and you choose to quantize those layers. And so the result of doing this mathematical quantization is you end up with a model that's 20% more quantized than another provider, so you get 20% more throughput of it because there's more layers than running an NVFP4, and your quality is better than that other quant because the layers that you chose to quantize have their errors cancel out, like one layer skewed to the right one layer skewed to the left, one layer skewed to the right. Your final logits distribution is more similar to the original distribution of the model, so you have better fidelity. And so the way we proved this was with KL divergence. So instead of just scoring on the benchmarks, we scored the KL divergence between the logit distribution of the quantized model and the logit distribution of the original full precision model, and we showed that with this technique we get. If your probability distribution on the logits which token it wants to select is more of the same as the original model, you're probably gonna end up staying true to the original model. So yeah, so it seems like previously before this, it seemed like the industry was, well, the more you quantize, the worse it's gonna be, ‘cause the more loss you introduce. That's not exactly, not necessarily true. So yeah, doesn't improve it, but can cancel out.Philip [00:31:57]: I think it might be this, but reminds me a good bit about pruning where you can prune off certain layers.Philip [00:32:03]: But very interesting. Didn't know this was a whole paper you guys put out.Ali [00:32:06]: It's. Fun fact, it was originally 72 pages, this paper, and then we decidedPhilip [00:32:11]: WowAli [00:32:11]: We can't tell. We couldn't release it. So it's now 45.Swyx [00:32:15]: Still 39 pages, so very substantive. We talked about evals and all these things and, like what's possible in terms of speedup? Like it's like probably like the numberInference Speedups and BenchmarkingSwyx [00:32:25]: Thing that people do wanna care about, and it's something that you wrote about in your post. Like official API is 70 tokens per second, and you push it up to 90. Is that like a normal thing?Philip [00:32:36]: So what's cool about working in inference, the reason that I think inference is going to be a useful place to do engineering for a long time, is that if you look at highly optimized domains like, say, finance, if you're in finance, you measure how much better you got in basis points. It's like, “Oh, I got five basis points better, like twentieth of 1% better,” that's huge news because everything is so optimized. When we publish optimizations, it's 20%, it's 100% it's 200%. So there's still probably like a lot further to go, honestly. Like you'll, you'll know that inference is pretty much solved when researchers start publishing about how they got 1% faster at something.Swyx [00:33:19]: Which by the way, because I am from the finance background, in the ‘70s, that was the margin at the time. When you did quantitative finance research, you would findAli [00:33:27]: And like 20%, tens of percent.Swyx [00:33:29]: That's. Yes.Philip [00:33:29]: Yeah.Swyx [00:33:30]: And now it'Philip [00:33:31]: Tiny fractionsSwyx [00:33:32]: For those people interested, look up Andrew Lo's paper. He had a really interesting illustration of quant, stat arb, distribution, narrowing down from like those kinds of 20% differences in the ‘70s, down to nothing today, which is very cool.Philip [00:33:48]: Exactly, and we're at the beginning of the same type of thing. Now benchmarking is hard. I think anyone will tell you that, and benchmarking provider speeds is hard because there's so many variables that go into it. What hardware are you using? How much load do you have on the system? What's the exact nature of the prompts and input and output sequence lengths? All that stuff. But overall, when you start stacking these improvements, you're looking at multiples. You can look at it. The most common form, of course, is TPS, tokens per second, which is bad naming by us in the industry, ‘cause there's two tokens per second. There's tokens per second, the throughput number, and the latency number.Ali [00:34:31]: TTMT, yeah.Philip [00:34:32]: Like total tokens per second out of the, out of the GPU as a throughput number. Most people only care about tokens per second as the latency number, which we should call ITL, intertoken latency, but we don't.Philip [00:34:44]: Anyway, so you can imagine a standard API without many optimizations for a 1 trillion parameter model operating somewhere in the 30 to 50 tokens per second range for reasonable traffic profile. And we generally see the goal of, pushing to 10X that. But, not necessarily day zero, but by stacking enough optimizations, if you have, say like four optimizations, each of which doubles performance. Or sorry, three optimizations, each of which doubles performance, then you stack that up, that's an 8X gain. That's the order of magnitude that we're working with in this space. We're trying to make things substantially faster, not just go from like 70 to 90.Swyx [00:35:38]: Are you saying you've. You have done that?Philip [00:35:40]: So let's say you have as a reasonable baseline, 30 or 40 tokens per second. You can achieve 10X that. So like on GLM-5.2, if you run it unquantized, perhaps on H100s even, and you're just using an off-the-shelf inference engine with no particular optimizations, no speculator, nothing extra around like KV routing, no disaggregation, you're, you're probably, yeah, looking at that like 30 to 40. You think that's like a reasonable baseline?Swyx [00:36:12]: Right. Right.Philip [00:36:12]: To get to something like 10X, there's a lot of trade-offs that you're making. If we're running at more like a 300, 400 tokens per second range, you are using the best hardware possible. You have a optimized speculator. You have done all of your quantization work. You are Seeing a pretty high cache hit rate. You are running with a reasonably small batch size and a parallelism configuration that is tuned for latency versus throughput, but it is possible. So the spreads that you see if you, like, go on artificial analysis or you go on OpenRouter and you look at, the worst provider to the best provider, oftentimes can hit that range. 10X is of course very aggressive. It's oftentimes maybe more of a four to six times improvement. But that's the performance that makes us really excited, is when we can get these huge gains, not just go from 70 to 90 tokens.Stacking Optimizations: NVFP4, Speculation, and DisaggregationAli [00:37:19]: It's also, like, hardware dependent. Like, ifPhilip [00:37:20]: YeahAli [00:37:20]: If you have a thing where you're serving it on just, like, a node of H100s and then you throw, like, you shard the model across, like, four nodes of B200s. Like, you can definitely increase the speed with just throwing more hardware at it. Like, normalizing for the same exact hardware and the same number of GPUs.Philip [00:37:35]: Yeah. Then you're looking at, like, a two to 4X improvementAli [00:37:38]: Right. RightPhilip [00:37:38]: Depending on the inference optimizations. So yeah, it's. Some of it's, what's the call, and some of it's who's the driver.Vibhu [00:37:46]: If you break down the two to 4X, say the example is run GLM-5.2Ali [00:37:51]: YeahVibhu [00:37:51]: On B200sAli [00:37:53]: YeahVibhu [00:37:53]: Single node, right? What's, like, the cost trade-off for effort to get, like, the last bit of juice out versus what should people just think of, right?Ali [00:38:01]: Spectre quantization. Yeah.Vibhu [00:38:03]: Spectre quantization.Ali [00:38:04]: That's, that's, that's like 95%. LikeVibhu [00:38:06]: And how far does that get you? And how easy is that for the average person to do? So say right I wanna throw the weights of GLM-5.2 on a node of B200s, how easy is it to find speculative decoder- decoder model or already quantized model? How much work goes into it?Philip [00:38:23]: If you're doing it up front, it's quite a lot of work. If you're doing it today, there's going to be people who have published things that you can just, you can just grab some NVFP4 weights. You can grab a speculator. Yeah, if we're thinking about, like, what are the 2Xs we're stacking, going from, BF16 to NVFP4 is, it's not quite a 2X, right? It's like. I think it's about, like, 30 to 40%, from 16 to 8, and then another 30 to 40% multiplied from, 8 to 4. So that doesn't quite get you a 2X, but, like, roughly a 2X. Speculator, roughly a 2X. Disagg on top of that if you're able to get enough hardware and put enough traffic through it, another roughly a 2X. And then you add in some, double-digit percent increase from having just a better runtime with, the latest kernels and stuff behind it. And that's how it stacks up.Ali [00:39:21]: YeahPhilip [00:39:21]: So building each of those, like, building the, quantized weights is, for someone who really knows what they're doing, hours to days of work. Building the speculator, again, like, hours to days of work. And the, disagg setup, hours to days. Well okay, but like once you haveAli [00:39:39]: Once set up. Once set up. YeahPhilip [00:39:40]: Yeah, getting disagg working for the first time, I'm saying, of course, is very difficult.Philip [00:39:44]: The marginal implementationAli [00:39:48]: Like, if you're just grabbing, like if you are a person, like just a normal consumer who has access to, like, a node of B200s and you're wondering, “How can I just host it myself?” You don't need to quantize the model yourself. There's always gonna be, like, an open source quantized checkpoint. NVIDIA's gonna push one out if no one else does. You. Usually, the providers will have their own spec dec that they've trained as well. You don't need to train your own spec dec. You can just use that as well.Philip [00:40:09]: Yeah. Like, GLM-5.2 has its own MTP.Ali [00:40:13]: Right. Right.Vibhu [00:40:14]: What's multi token prediction?Philip [00:40:15]: Yes.Ali [00:40:16]: I'm justVibhu [00:40:16]: Can you explain that?Ali [00:40:16]: I'm just an expert.Ali [00:40:18]: I can do it for you in case I get it wrong?Vibhu [00:40:20]: No.Vibhu [00:40:21]: Yeah, you should correct if we're wrong, but their multi-token prediction can be used for self-speculative decoding.Ali [00:40:27]: I'm not sure. I'm not gonna correct that.Vibhu [00:40:28]: Okay. I'm semi-confident in thatAli [00:40:30]: Okay. YeahVibhu [00:40:30]: But someone can check. But it's useful to paint the story of, okay, not just the average person, but say a company wants to switch from serverless inference I wanna throw this up on. I wanna rent some GPUs, throw it up. These are the steps you take to do significantly faster than just put it behind vLLM.Ali [00:40:48]: Right.Vibhu [00:40:49]: I was waiting for a mention of Dynamo.Vibhu [00:40:51]: I feel like, that's supposed to be the baseline that you measure against.Dynamo, KV Routing, and Disaggregation ToolkitsPhilip [00:40:55]: I would think of Dynamo as less of a box system and more of a toolkit for building with. So when we talk about doing aware routing, when we talk about doing KV offloading, when we talk about doing, PD disaggregation, Dynamo fundamentally is. By the way, Dynamo is an open source library from NVIDIA.Ali [00:41:17]: We've done a pod with KylePhilip [00:41:18]: OkayAli [00:41:19]: Kyle Cranin.Philip [00:41:19]: Cool. So then your listeners know then that it supports all the different inference frameworks. And it is multi hardware, which is interesting.Ali [00:41:28]: But it's just a router, it's not like an optimizer layer.Philip [00:41:30]: Yeah. All it does, like, what Dynamo is good at, it is a library for moving information around your cluster, around your hardware. So if you have, KV cache on one place and you need it to be somewhere else, Dynamo coordinates NIXL for you to move that around.Philip [00:41:49]: That doesn't mean that, like, out of the box, you just say, “Pip install Dynamo,” and then you get, like, a massive performance speed up. It's more of a developer toolkit.Ali [00:42:01]: Yeah. I would have said it would. It comes with a set of defaults that you can then swap out.Philip [00:42:06]: It does. If the industry at large, I think, was, like, rolling out all of these deployments, standard, then I think it would be, like, a credible baseline. But, we've got to, we've got to benchmark against, like, what we're seeing in the wild.Speculative Decoding Methods: Medusa, EAGLE, n-Gram, and Spec-SpecVibhu [00:42:23]: I did wanna talk a little bit more about PD disagg, because that is probably, like, number three after quantized and speculative decoding. In your book though, I was just gonna pull out the book.Philip [00:42:31]: Yeah.Vibhu [00:42:32]: Like section 522 on Medusa, 523 on EAGLEPhilip [00:42:35]: YeahVibhu [00:42:36]: 524 on gram.Philip [00:42:37]: It's 55, would be disaggregationAli [00:42:42]: Yeah. Well, no, I just wanted to dwell a little bitPhilip [00:42:44]: YeahAli [00:42:44]: The other. Like, so what do you choose to include? What do you choose to not to include? Because there was all these other techniques.Philip [00:42:51]: Yeah.Ali [00:42:51]: Are these still relevant? Because I think they came out, like, a year and a half ago maybe.Vibhu [00:42:55]: Medusa is quite old.Philip [00:42:56]: Yeah, Medusa's old.Ali [00:42:58]: It was old.Vibhu [00:42:58]: But is it in the book as a good, here'sPhilip [00:43:01]: BaselineVibhu [00:43:01]: Baseline vanilla understand it?Philip [00:43:02]: Like you should know this.Vibhu [00:43:03]: Like I read the paper, I'm like, “ it makes so much sense.”Philip [00:43:05]: Yeah.Philip [00:43:05]: So with the book, I had a couple goals. One was to give people just a working vocabulary for the space as a whole, and the other was to give them some intuition about how each of these techniques works. As I mentioned in my AI Engineer talk, which is the first public addendum to this, the speculation space has moved much faster than everything else. So yeah, even at the time that I wrote the book Medusa, I very much included as a way for people to understand how the space evolved rather than what the most modern technique is. And now of course, there's DFlash, dSpark. There's, there's newer techniques even than EAGLE, although EAGLE is still very commonly used.Ali [00:43:51]: SpecSpecta.Philip [00:43:52]: Yes. Speculative decoding.Vibhu [00:43:54]: What canAli [00:43:56]: Oh, it's a paper by Tri Dao and it's like, it's doing speculative decodingVibhu [00:44:00]: HuhAli [00:44:01]: For the speculative decoder.Philip [00:44:02]: Oh, in spec- oh my God.Ali [00:44:02]: It's literally just an another. It's like, yeah, that's the most simple way to explain it, and it seems like he got trivial speed ups there. But it seems that the complexity with training, it's almost like in our mind at least, it's almost as complex as training GANs. Like it's like a very delicate balance and oftentimes you, it's just but yeah, it's literally speculative decoding on speculative decoding.Vibhu [00:44:21]: Speculative.Ali [00:44:22]: Yeah. We saw this paper.Vibhu [00:44:24]: It's interesting, right?Ali [00:44:24]: Yeah.Vibhu [00:44:24]: I wouldn't even expect it to be very particular to train, I wouldAli [00:44:29]: Right.Vibhu [00:44:29]: The naive part of me is like, okay, train speculative decoder.Ali [00:44:32]: But like, and it makes sense, like the whole idea of speculative decoding is you. It's like, it's like almost like the iPhone auto predict version but for a normal model, right? Like you're just, you're just, generating three tokens and you're like, okay, I'll do prefill on them. And so you save those three turns for your original model. Now your speculative decoder is doing three turns of auto regression, so why not just have an even smaller model?Ali [00:44:53]: The other question there is what are the size of speculators? So say forPhilip [00:44:58]: Right. It's like a billion parameters.Ali [00:45:01]: Like for MiniMax, it's. Yeah. It's like one layer. It's like one 60th of the original model usually.Philip [00:45:06]: Yeah. I think we should do a paper when we get back to the office.Philip [00:45:10]: SpeculativeAli [00:45:11]: SpeculativePhilip [00:45:11]: Decoding.Ali [00:45:13]: No, it's, it does seem like how, when do you stop? But then it also seems like if you're able to train spec-spec decode for instance, right? Like if you're able to have a small model that is accurately predicts what the intermediate speculator is gonna predict, that is able to predict what the original target model's gonna predict, then why not just use that smallest model directly, right?Vibhu [00:45:34]: Yeah. This isAli [00:45:35]: Like it seems likeVibhu [00:45:35]: Adjacent to the routing problem.Ali [00:45:36]: Right.Vibhu [00:45:36]: Yeah.Ali [00:45:36]: Right.Philip [00:45:37]: The thing with speculators is one of the practical constraints on using them is that you do have to run a small model on the same hardware that you're running the big model on. There is a orchestration and resource competition problem inherent in that, and that is one of the constraints on speculation in general, is that draft tokens cost resources to create and cost software complexity to manage. And so if you have like infinitely recursive speculators, you add in quite a bit of that complexity on the actual implementation within the inference engine as well, not just in the training process.Vibhu [00:46:17]: I was gonna say, I would wonder if you could do similar, like distillation and pruning of, it's the same thing, it's just a model. Can we not just distill a lot of the weights, quantize the speculator, out of my domain? The question that also comes up is, this is all for big server workloads, right? How much of this applies to, say I have this MacBook, I wanna run Gemma really efficiently. Similar problems, not the same?Local AI vs. Data Center InferencePhilip [00:46:45]: Pretty different. I talked to Selo, about this on his podcast a couple weeks ago. The difference between inference engineering for the data center and for production workloads versus inference engineering for local AI, is that we start with fundamentally like different constraints and different goals. With local AI, it's how do I fit this model onto my hardware and then make it less dumb? And with data center influence, it's how do I load this model and then make it less slow? And we care about less dumb, and they care about less slow. But the local AI inference engineering ecosystem, I think has a lot for us to learn from in the data center space. They are experts in various forms of quantization, including dynamic quantization that we just don't touch, in the pruning, in the distillation, in the, layer removal. There'Ali [00:47:42]: Layer removal matters less.Philip [00:47:43]: Yeah. There'Ali [00:47:44]: No one loves pruning really.Philip [00:47:45]: Yeah. Well, but the, but they doVibhu [00:47:46]: Which is surprising, right? But that's, that's a whole different thingPhilip [00:47:48]: Just to fit something on the laptop.Ali [00:47:50]: Right.Philip [00:47:50]: So yeah, it's a, it's an interesting, it's an interesting space. Not necessarily that like their techniques make sense for us to do in the data center, because we have different resources and different goals, but more that the process as well as the openness of that field is something to, admire.Ali [00:48:12]: Yeah. Like to your point, like, certain optimizations that would. Like for instance, Turbo Quantum Sharper, like it made such huge hype on that and we did like a whole deep dive on Twitter and like said, what is it? How does it work? Why is it good or not? And it took off and it was implemented on local devices because your memory bandwidth is so slow on like a MacBook, for instance. But try putting the same thing on like an NVIDIA GPU on a B200 Turbo quant would not be. Like, it would not be used. Like, NVIDIA - Like, NVIDIA made it clear that this is not a good optimization, and we've seen it firsthand where the overhead of doing dequantization, quantization of, in the kernel itself with turbo quant kernel, each end is much slower than the time that you save from doing the bandwidth. ‘Cause on the B200s, you have like 3.5 terabytes per second. You don't need decrease the storage that much. You don't need to do, FP4 KV cache. You don't need to use a requant. There's, there's, there's better optimizations to be made. But on Edge devices, it's extremely important, it's extremely useful. So, seems to be, like, different optimizations there, but then they're all uniquely combined with like all you wanna quantize the model, you wanna do speculative decoding, like certain common prefixes with bothPhilip [00:49:18]: Principles.Ali [00:49:19]: Yeah, exactly. Exactly. Exactly.Philip [00:49:20]: They also do a lot of work on, model parallelism, especially over, heterogeneous topology, where you have, some sparks and they are wired together with, Ethernet, DGX sparks.Ali [00:49:35]: Yeah, this is the Exo Labs guys.Philip [00:49:36]: Yeah. You have, a nu

Lisa Cole, a four-time Chief Marketing Officer and current CMO at 2X, discusses the challenges marketers face in doing more with fewer resources and the misconceptions around AI in marketing. She explains that while AI can automate tactical tasks, human judgment remains essential for creating unique, high-quality content that differentiates companies from competitors. Lisa emphasizes that marketers should be intentional about what tasks to keep in-house, outsource, or automate, rather than relying solely on AI to handle all marketing functions. Lisa and John discuss the challenges marketers face due to AI, including the pressure to create content that is both understandable to humans and AI models across multiple channels. Lisa explains that marketers now ns to consider both human and AI audiences, leading to increased workload and expectations. She notes this situation is creating tension in organizations of all sizes, with some facing pressure to reduce headcount while others are expected to do more with less. Lisa advises companies to focus on understanding and mapping their current marketing workflows before implementing AI, rather than immediately acquiring more AI licenses. She recommends conducting a "hackathon" to reimagine and optimize existing processes, aiming to reduce time from idea to execution and eliminate manual review cycles. For John's small business helping people buy and sell businesses, Lisa suggests asking their preferred AI model for specific advice on how to apply AI to grow their LinkedIn following and engage with executives.You can reach Lisa at www.2X.marketing or Lisa.cole@2X.marketing.John MartinkaJessica MartinkaContact us via either website or give us a call and be sure to check out our blog pages with new posts weekly.https://nokomisadvisory.com/https://www.martinkaconsulting.com/https://www.gddpodcast.buzzsprout.comhttps://www.youtube.com/c/JohnAMartinka/videos 425-515-4903 

In today's Cloud Wars Minute, I examine why Google Cloud's explosive growth is redefining the competitive landscape ahead of Microsoft and AWS earnings. Highlights 00:03 —We have earnings results coming up later this week from both Microsoft and AWS. Microsoft on July 29, and AWS on July 30. So, that'll give us some nice comparisons to the stunning Q2 that Google Cloud reported last week. Google Cloud will be outgrowing AWS by somewhere between two and a half times at the high end, and at least two times on the low end. 01:02 — Over the last several quarters, Jassy has alluded to Google Cloud, not by name on earnings calls, but he said, "We've got some companies throwing up these fancy high growth rates." What we see with these growth rates is the customers expressing their vote. They are the ultimate arbiters in this decision about who's the hottest company. That's expressed in growth rates. 02:23 — I'm comparing Q2 here for Google Cloud versus Q1 for AWS because, right now, the Q2 numbers aren't out yet. AWS grew at 28%. Google Cloud grew 82%. Say AWS reports this massive acceleration to 40%. They would still be outgrown by Google Cloud by 2X. My guess is AWS's growth rate will come in somewhere around 30% or 31%. 03:22 — It is becoming a little tiresome for these bigger companies to say, "We're bigger than them. Of course, they have a higher growth rate." The customers are voting here. These growth rate disparities are so important because they tell us how the different vendors are doing in the marketplace through the eyes and wallets of customers. 04:01 — In the Cloud Wars, the customers are always, always the big winners. Microsoft and AWS are growing their cloud businesses in the range of 30%, which is astonishing. But Google Cloud posted 82% revenue growth and backlog growth of almost 400% to over $500 billion. By the numbers, Google Cloud is the king of the hill right now. Visit Cloud Wars for more.

Everybody thinks 10X takes 10 times the work. It doesn't. 2X is the grind. You just do more reps, more hours, more volume until it works. But you can't out-hour your way to 10X. There aren't enough hours in the week. 10X isn't a math problem you solve with effort. It's a different game entirely, and almost nobody plays it right.In this episode I break down why 10X is actually easier than 2X, and why the reps and leaders who get it stop adding and start cutting.What we cover:    •    Why 2X is linear and 10X is exponential (and why that changes everything)    •    The 80/20 rule most people quote but nobody actually applies    •    Why 10X is a subtraction game disguised as a growth game    •    The 3 things every 10X move comes down to: identity, time, and leadership    •    How to spot the 20% of activities driving your results, and kill the rest    •    The team application: stop spreading your time evenly across reps    •    The individual application: what to delegate and what to guard with your lifeIf you run a sales team or you're trying to break your own ceiling, this one will change how you spend every hour of your week.Drop a comment with the one thing you're cutting this week.New episodes of the Rainmaker Podcast every week. Subscribe so you don't miss the next one.Connect with me:Instagram: @nicknsales (https://www.instagram.com/nicknsales/)Youtube: @nicknsales (https://www.youtube.com/@nicknsales/)#SalesLeadership #Rainmaker #DoHardShit

Bilt Bulletin July 2026 | Palladium Card Value, Alaska Transfers, Bilt Cash Expiration & More Justin Vacula of the Hurdy Gurdy Travel Podcast is joined by returning co-host Darren covering the latest Bilt Rewards updates, Bilt Palladium Card strategy, Bilt Cash expiration concerns, and tools from NextCard that help Bilt users. Justin and Darren explain how Bilt Rewards works as a loyalty program with credit cards, transfer partners, and ways to earn rewards on rent and mortgage payments. They focus on the Bilt Palladium Card, including its 2X earning structure, value from Bilt Cash redemptions, hotel credits, and the ability to earn points paying rent or mortgage. Justin shares his first transfer from Bilt to Alaska Airlines Atmos Rewards, using points to book Philadelphia to Savannah flights for 7,500 points each way. Darren also discusses heavy Alaska mileage usage for family travel and a business-class redemption. They also review progress toward Bilt Platinum status, spending stats, and whether Bilt Cash expiration at year-end may reduce late-year card spend. Justin and Darren discuss why cardholders may want to redeem Bilt Cash aggressively before expiration rather than letting value go unused. Timestamps 00:00 Podcast Intro 01:25 Bilt Program Basics 02:37 Palladium Card Highlights 05:27 Earn Points on Rent And Mortgage Payments 06:37 Real Redemptions: Alaska Flights 09:28 Spend Stats And Status Progress 13:49 Bilt Cash Expiration Concerns 19:33 Support The Show 22:13 NextCard.com Tools For Bilt 25:43 Upgrade And Downgrade Questions 30:00 Rakuten Card Earns Bilt Points 33:01 Wrap Up And Listener Feedback 34:00 Outro And Credits —

If you're job searching in 2026 and following every resume rule you've been told, one page only, match every keyword, avoid two-column layouts, optimize for the ATS, this episode is going to change how you think about your resume and your entire job search strategy. Sam Wright runs the data side at Huntr, and he's analyzed over 1.7 million job applications, more than 240,000 resumes, and over a million job postings to find out which resume rules actually get people hired, and which ones are just guesses everyone kept repeating. The results are surprising. Two-page resumes outperform one-page resumes across the board. Resume tailoring is 2X more effective at driving interviews. Referrals convert at a massively higher rate than cold applications. And most of what you've been told about the ATS is a myth. Scott and Sam break down the resume myths costing you interviews, the tailoring strategy that actually works, why your target job title matters more than your resume format, and the data-backed job search moves that separate the top 10% of applicants from everyone else. What you'll learn: Why the one-page resume rule was never actually tested (and what the data shows about two-page resumes) The 5-step resume tailoring process that doubles your interview rate Why "optimizing for the ATS" is a myth (and what actually gets your resume seen) How to write a resume summary that stands out (without sounding AI-generated) Why referrals convert at a massively higher rate than cold applications, and how to build them The exact application strategy Sam recommends: how many jobs per week, how fast to apply, and how to target the right roles Connect with Sam Wright: https://www.linkedin.com/in/samuel-b-wright/ Try Huntr: https://huntr.co   Our book, Happen To Your Career: An Unconventional Approach To Career Change and Meaningful Work, is now available on audiobook! Visit  happentoyourcareer.com/audible to order it now! Visit happentoyourcareer.com/book for more information or buy the print or ebook here! Want to chat with our team about your unique situation? Schedule a conversation   Free Resources What career fits you? Join our free 8 Day Mini Course to figure it out! Career Change Guide - Learn how high-performers discover their ideal career and find meaningful, well-paid work without starting over.   Related Episodes Changing Careers (When You Don't Know Your Next Job Title) (Spotify / Apple Podcasts) Discover Your Strengths to Find Your Ideal Career (Spotify / Apple Podcasts)

What if the worst professional moment of your life turned out to be the best thing that ever happened to you? Lyndsay Dowd spent nearly 30 years in leadership — 23 of them at IBM — before leaving on her own terms for a new opportunity. Six months in, she was fired. No warning, no explanation. The shame and embarrassment took her out… until she asked herself three questions that changed everything: What am I really good at? What do I love to do? How can I help people the most? The answers became Heartbeat for Hire, a company that helps brands ditch toxic, top-down leadership and build cultures people actually want to work in. In this episode, Wendy and Lyndsay — a speaker, founder, 2X bestselling author (Top Down Culture and Voices of Women), Harvard guest lecturer, and globally ranked podcast host — talk about reinvention at 50+, betting on yourself, the “power skills” every leader needs, why trust is your currency, and the permission slip every woman at a crossroads deserves to hear. Fierce, funny, and full of heart.

America has between 32 million and 130 million empty bedrooms sitting unused every single night. The housing shortage is estimated at just 1.8 to 7.5 million units. If you can do basic math, you already know what that means, and Atticus LeBlanc, founder and CEO of PadSplit, is here to break down exactly why room rentals are the fastest, most profitable way for landlords to close that gap.In this episode, Atticus walks us through the exact room rental conversion that took his net operating income from $1,100 to $2,400 a month. He shares why the traditional buy, renovate, rent playbook does not pencil the way it used to, and why landlords looking for cash flow, financial freedom, or a way to make an underperforming property profitable again are turning to co-living.Room Rental Inventory on Furnished Finder:https://www.furnishedfinder.com/search/furnished-roomsCheck average room rental pricing in your markets: https://www.furnishedfinder.com/statsIn this conversation you will learn:Why room rentals can deliver a 2X increase in net operating income compared to traditional single family rentalsWhere co-living demand is significantly outpacing supply in a variety of US marketsWhy PadSplit residents save an average of $317 a month, adding up to over $200 million in collective savingsHow the Furnished Finder and PadSplit partnership brought 1,000 room rentals across seven markets onto the platformThe real story of Tiffany, a night shift security guard who saved $600 a month, bought her own car, and moved into her own apartment within six monthsWhat zoning, HOA restrictions, and Golden Girls bills mean for landlords considering the co-living modelWhy seasonality in room rentals looks nothing like the short term rental world, giving landlords steadier year round cash flowFurnished Finder and PadSplit together offer over 95,000 room rentals nationwide, giving landlords and investors the largest combined footprint of monthly co-living inventory in the United States. Whether you are a mid-term rental landlord, a short-term rental host adjusting to new regulations, a long-term landlord looking for more cash flow, or a new investor exploring your first property, this episode gives you the data and the strategy to think about your next move differently.List Your Property on Furnished Finder: https://www.furnishedfinder.com/list-your-property Use code LLD10 for $10 off new listings.Timestamps0:00 Welcome to Landlord Diaries1:30 What is PadSplit and the problem it solves3:20 The impact of room rentals on the housing shortage4:30 Millions of US bedrooms sit empty every night5:00 The barriers that block market rate apartments7:35 Why co-living is a 4 sided win for communities8:30 What landlords meet PadSplit brand standards10:10 How room rentals move the needle on affordability11:20 10,000+ affordable units created in Atlanta metro12:25 Who is leasing and renting by the room14:35 1/3 of US renters earn less than $35k a year16:20 Why demand stays stable in every economy18:45 The huge cash flow opportunity in room rentals21:40 What NOI is and why it matters for landlords23:40 Turning your dining room into a rentable bedroom26:55 Top room rental markets and how to create supply30:30 Why Furnished Finder and PadSplit partnered up34:55 Remote workers can test cities for 3 to 4 months36:40 A real underperforming property that transformed37:45 Tiffany's story: from motel to her own apartment39:10 Misconceptions landlords have about co-living41:35 Do room rentals have a seasonality problem44:10 Zoning and HOA restrictions to watch for48:15 Golden Girls bills and the fight to change laws49:20 The first step to take this week50:15 Rapid fire: surprise markets and biggest mistakesMeet your hosts Kelly Bailey & Katie Lyon, plus catch up on past episodes:https://www.furnishedfinder.com/resources/podcastTopics covered: mid-term rentals, room rentals, co-living, PadSplit, Furnished Finder, net operating income, real estate investing, housing shortage, affordable housing, cash flow strategy, landlord tools, monthly rentals The Landlord Diaries is brought to you by Furnished Finder, where you can list your property for one low price and pay zero booking fees.

¿Te han vendido "gas" en un bote? Descubre la verdad sobre el Óxido Nítrico y cómo maximizar el PUMP (hiperemia) con base científica. En este video te explico de forma clara (para dummies) por qué tus músculos se hinchan al entrenar y qué suplementos tienen evidencia real para potenciar el crecimiento muscular. Analizamos la fisiología detrás de la citrulina malato, la arginina y los nitratos de la remolacha. No tires tu dinero: aprende qué dosis necesitas y cómo funcionan las vías de señalización anabólica (¡hola, integrinas!). MI PRIMER LIBRO: "Hormonas, Metabolismo y Rendimiento" https://amzn.to/3QNPjnq ✍️ Me ayudas muchísimo dejando una reseña en Amazon. Al ser autopublicado, cada comentario ayuda a que el libro llegue a más gente. Código PROFECLAUDIO → https://belevels.com/collections/todo... _________________________________ Suscríbete aquí / @profeclaudionieto / profeclaudionieto https://www.profeclaudionieto.com/ Mi libro: https://amzn.to/3QNPjnq Si dedicas más tiempo a tu canal que a tu negocio, tienes un problema. Nosotros lo solucionamos → crecimientoenyoutube.com ___________________________ ⚠️ Este contenido es divulgativo y tiene fines educativos. No sustituye el asesoramiento médico, diagnóstico o tratamiento profesional. Antes de tomar decisiones sobre tu salud, consulta con un profesional cualificado. El uso que hagas de este contenido es bajo tu propia responsabilidad. 00:00 - ¿Qué es realmente el "Pump" o Hiperemia? 01:11 - La mentira del Óxido Nítrico en suplementos 02:17 - Fisiología básica: ¿Por qué se hincha el músculo? 03:23 - Espacio intersticial y miocitos: Ciencia fácil 04:44 - El secreto de las Integrinas y el anabolismo 06:23 - Vías metabólicas: Fosfágenos y Glucólisis 07:29 - Fibras tipo 2X y Aquaporina 4 08:46 - Óxido Nítrico: El gas vasodilatador 09:12 - Vías NOS dependiente vs. independiente 11:18 - Citrulina Malato: El suplemento estrella 12:50 - El papel del Malato en el Ciclo de Krebs 13:42 - ¿Qué dice la ciencia sobre la dosis efectiva? 16:01 - Arginina: ¿Por qué no es la mejor opción oral? 17:44 - El poder de los Nitratos (Zumo de remolacha) 19:54 - Resumen final: ¿Cuál elegir para tu objetivo?

► Want to work with Randy directly: https://randygage.com/breakthroughu/ ► Prefer audio? Get the podcast https://randygage.com/podcast Most people assume the bigger the business, the harder it is to build.Randy Gage has spent decades proving the opposite is true.In this video, Randy makes a claim that stops most founders cold and then spends the entire video proving it with real examples, real clients, and a framework built from over $23 billion in generated revenue.He also breaks down something nobody in the business world is talking about: what AI is currently destroying, silently, in companies of every size. And why the businesses that feel the safest right now may be the ones in the most danger.If you're a founder, operator, or builder, whether you're at $100K, $1M, or $10M, this video will change how you think about growth, risk, and what it actually means to build something that lasts.The good news: everything Randy covers is fixable.But only if you're willing to hear something most business advisors will never tell you.Videos Randy mentioned in this video► Video: https://youtu.be/Ec-aTTtKSPw?si=I67P4OnQfsVjtRcc► 10X better than 2X: https://a.co/d/0b6iU1wl► Who not how: https://a.co/d/0icNDJNgWho is This Guy?Randy isn't some influencer who listened to a couple podcasts and started posting hot takes. He's the streetwise founder of Prosperity Factory, Inc., who has been building and scaling businesses for more than 40 years. Randy has authored 16 bestselling books, translated into 25+languages, including his latest—Wealth Without Apology—and spoken to more than 2 million people across 50 countries. He's been inducted into both the Speaker Hall of Fame and Direct Selling Hall of Fame.But none of that is why people follow him. They follow him because he calls BS…and says what most people are too afraid to admit. When he's not rocking the stage or building his next project, you'll probably find him coaching a softball team somewhere.Connect with Randy:Instagram: randy_gage Twitter: randy_gage Facebook: randygage

You didn't leave the corporate world to build a different cage. In this episode of The Fulfillionaire, Austin Netzley, author, investor, and entrepreneur behind Operations.com, breaks down exactly how to scale your revenue, protect your cash flow, and build a business that thrives without you. Most businesses are running at 40% to 60% waste. Austin's first move with every CEO is always the same: simplify ruthlessly, nail the strategy, then build systems that live outside anyone's head. If your business slows or stops the moment you step away, you don't own a business. You own a job.  Stage four is a business that runs, grows, and thrives without you. That is when you stop managing and start building real enterprise value and real freedom. Stop building a business that owns you. Tune in now to How Accountability Turns Failure Into Unstoppable Success with Austin Netzley. When you are ready to align your work, wealth, and life around the freedom you actually want, visit fulfillionaire.com. Austin Netzley is an author, investor, and entrepreneur who helps service-based business owners scale predictably and profitably without the chaos that typically comes with growth. As the founder of 2X, a back-to-back Inc. 5000 fastest-growing company, he has built a coaching methodology that has generated over $310 million in client results. His work focuses on helping entrepreneurs between $500K and $5M in revenue reclaim their time, gain clarity, and transform their business into a system that runs without them. He has been featured in Forbes, Entrepreneur, Business Insider, and the Washington Post, among others.  Personal Website: https://austinnetzley.com/  Company's Website: https://operations.com/  Instagram: https://www.instagram.com/netzley/  Facebook: https://www.facebook.com/scale2x/  LinkedIn: https://www.linkedin.com/in/austingnetzley/  JP Newman is the founder of Fulfillionaire and CEO of Thrive FP, known for helping high-achievers align financial success with deeper human connection and purpose. With over $2 billion in real estate transactions and hundreds of investors coached, he brings a powerful blend of strategy, psychology, and emotional intelligence to the world of investing and negotiation. JP teaches that the best deals are built by understanding people, energy, and intention. Through his Fulfillionaire™ movement, he helps leaders stop operating from fear and start making decisions rooted in clarity and alignment. His approach redefines negotiation as a human-centered skill that turns insight into influence and lasting success. IG: https://www.instagram.com/jpnewman_/  LI: https://www.linkedin.com/in/jp-newman-45a1ba/     

This is a free preview of a paid episode (43 min), exclusively available on our subscriber-only premium feed. Become a premium subscriber to tune into the full episode: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://cubicletoceo.co/podcast⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Questions about our premium podcast subscription? Send us a DM ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠@cubicletoceo Continuing our series on “Revenue → Returns (How My Money Makes Me Money),” today we're giving you the detailed low-down on angel investments as a money multiplier. Returning to our show to transparently dissect her own angel portfolio is Jaclyn Johnson. Jaclyn is a 2X exited entrepreneur (selling her previous companies for over $23M) and prolific angel investor creating real impact for underrepresented founders through both financial and relationship capital. She's the founder of Create & Cultivate, the largest media and events company for ambitious women, the current CEO and co-founder of Cherub, a platform connecting entrepreneurs and angel investors, and runs Receipts VC, her umbrella for 20+ portfolio investments spanning beauty to tech. Jaclyn walks us through her investing thesis and recent deals, what she looks for before writing a check, how much of her business profits she allocates annually toward angel investments, and her early-stage investment in luggage company Away that turned a $10K bet into $1.2 million. Listen to our last episode with Jaclyn: https://open.spotify.com/episode/76d1DtICFUJagqWxExM5fq?si=s9UTTRWnTvK1-bjt64Ez5A Connect with Jaclyn: Apply to attend Cherub's Creator Summit: https://creatorsummit.investwithcherub.com/ ⁠Take a look at Jaclyn's private room: https://receiptsvc.investwithcherub.com/#/join  www.receiptsvc.com www.createcultivate.com  www.investwithcherub.com IG: @jaclynrjohnson If you enjoyed today's episode, please: Post a screenshot & key takeaway on your IG story and tag us ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠@cubicletoceo⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ so we can repost you. ⁠⁠⁠Subscribe to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠our premium feed⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ for case-study style interviews every Monday.⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Learn more about your ad choices. Visit megaphone.fm/adchoices

Do you want to know God's vision for your life but feel like you keep shrinking your dreams down to what you can figure out on your own? In this teaching, Travis Lee Peters breaks down what vision, mission, and purpose actually mean for the believer, and how to hear what God wants to do in your life clearly, then walk it out through daily obedience. You'll learn why most of us settle for a vision that's "too small" (just get out of debt, just get comfortable), how to stop dreaming without God involved, and why His vision for you is always bigger than anything you could engineer in the natural. We cover the spiritual version of "10X is easier than 2X," how to get quiet and let God drop the vision in your heart, and how that vision turns into a mission you live out one obedient step at a time. If you've ever felt stuck, scattered, or unsure of your purpose, this one is for you. If this blessed you, don't just watch and leave. Get plugged in today so you can begin to experience God's Promise for Increase on new and exciting levels:

Fourteen deals worth mentioning in a single week. And those are just the ones that surfaced.The week before Cannes, the dam broke. Christian and Ayelet break down the deepest deal review we've done yet — anchored by a transaction Ayelet's team actually advised on the sell side: Residence acquiring GateMaker, a female-founded creator and influencer agency with a blue-chip beauty roster.Plus a sponsor-to-sponsor recap in commerce services (Bluebird Group + Bertram Capital), and a rapid-fire run through 12 more deals across creator, beauty, luxury PR, B2B, and commerce.One platform investment. One deep dive. Twelve quick hits. Under 20 minutes.⏱️ TIMESTAMPS0:39 — Welcome to Market and Deals Friday — and a 14-deal week1:12 — Did everyone wait until the week before Cannes to announce?1:37 — Coming up: a special edition with Chris Erwin of RockWater on the Accenture/Whalar deal2:00 — Market update: Bluebird Group partners with Bertram Capital2:20 — Bertram's buy-and-build model and the Bertram Labs tech advantage3:00 — Reading the deal size from a $1.6B control fund with a 43% IRR4:00 — Why a relationship-driven commerce services business resists AI disruption5:35 — Deep dive: Residence acquires GateMaker — a sell-side deal Ayelet's team ran6:05 — GateMaker's founders, blue-chip beauty roster, and creator economy pedigree7:20 — Did Residence already have creator capability? (No — this was the capability buy)7:50 — Second acquisition in under five months: Residence is now a 9-agency network8:44 — The Gemspring-backed platform build and why Residence is now an active acquirer9:28 — The "anti-holdco" model — and Christian's pushback on the framing10:42 — Why creator and influence relationships command a premium right now11:30 — The cross-industry pattern: do-no-harm PMI for people-heavy businesses11:56 — Advisors: Palazzo and Speed M&A on the sell side12:32 — Brinkley the deal-finding agent and a 20-deal week13:20 — Quick hit: Front Row acquires Carbon Beauty (second deal this year)13:57 — Quick hit: Mazarine acquires Bacchus — luxury PR and UHNW access14:10 — Quick hit: Huge acquires Rotate — composable commerce14:27 — Quick hit: Motion Agency acquires LKHNS — B2B and video (Kim Everl's 7th)15:30 — Quick hit: Akeneo acquires Pricing Hub — PIM moves into pricing16:33 — Rapid fire: Mile Marker/Lyfe, Legion Advertising, Factual/Intelsio, Everything Branding/Darlington17:44 — The week's only disclosed number: 2X acquires KnownWell at a $400M combined valuation18:09 — Quick hit: Scorpion acquires One SEO Digital18:32 — 14 deals, one disclosed price: the lower middle market buying capability quietly18:52 — Don't miss the Erik Huberman interview (Ep. 71) + the Chris Erwin special coming up

In this episode of Million Dollar Flip Flops, Rodric sits down with private equity investor Steve Walsh for a conversation about what it really takes to grow a business beyond the next plateau.Steve shares how he helps entrepreneurs think bigger, build stronger companies, and avoid the common mistakes that keep them stuck at two, three, or even ten million in revenue. The discussion explores mindset, customer understanding, culture, capital, and why the right investor can bring far more than money to the table.From there, Rodric and Steve dive into the difference between smart capital and dumb capital, the importance of hiring A players, and why communication with investors matters more than most founders realize.This is a practical, high-level conversation about growth, leadership, and building a business with the right people around you.In This Episode, You'll LearnWhy many entrepreneurs get stuck before their next levelHow to think bigger than your current revenue targetWhy mindset is often the first bottleneck to growthThe difference between smart capital and dumb capitalWhy the best investors bring doors, systems, and strategyHow to build a culture that attracts A playersWhy customer understanding matters before capitalWhat entrepreneurs often get wrong about their investorsHighlights & Timestamps[00:00] Tell your investor the truth Steve opens with a blunt truth: the things you least want to tell your investor are often the things they most need to know.[01:00] Meet Steve Walsh Steve shares that he works in private equity and focuses on investing in entrepreneurs and helping them reach the next level.[02:00] Why 10X thinking matters He explains why he believes 10X growth is not harder than 2X growth—it just requires a different mindset.[03:00] Special snowflake syndrome Rodric and Steve discuss how entrepreneurs often believe their market is different, when in reality the core principles of business still apply.[04:00] Culture can solve hiring problems Steve shares an example of a company that believed it could not find talent, and explains how better culture can attract people away from competitors.[05:00] Shiny object syndrome The conversation shifts to the trap of spending capital on things that do not move the business forward.[06:00] Start from the end and work backward Steve explains why entrepreneurs should begin with the end goal in mind and then reverse-engineer the path.[07:00] The Magnet and sales fundamentals Rodric connects the conversation to sales and marketing, explaining how many founders neglect the basic “magnet” that attracts customers.[08:00] Why builders struggle with sales They discuss how many builders and contractors are excellent at building but weak at sales, marketing, and customer messaging.[09:00] Smart capital vs. dumb capital Steve explains the difference between money alone and strategic money that brings expertise, relationships, and systems.[10:00] Partnership, not transaction The best investors, Steve says, are true partners who want to help founders succeed beyond the check they write.[11:00] Tell your investor the hard stuff He emphasizes that many founders hide problems from investors when those are exactly the issues that need to be surfaced.[12:00] Why the right investor changes everything Rodric and Steve compare smart capital to strategic help like networks, expertise, and market access.[13:00] A players need a big mission Steve explains why top performers need a meaningful challenge or they will get bored and leave.[14:00] The book and where to find Steve Steve shares where listeners can find his book and connect with him.[15:00] The family man question Steve answers the signature question: businessman with a family or family man with a business?[16:00] Energy, time, and personal alignment He shares how he looks at both business life and personal life when evaluating entrepreneurs.[17:00] Closing thoughts on priorities Steve closes by reinforcing that personal and professional priorities always show up in the calendar and the bank account.Notable Quotes“Most of the things you do not want to tell your investor are the things you need to be telling your investor.” – Steve Walsh “10X it is just as easy as 2X it.” – Steve Walsh “The bigger picture is always bigger than what they see.” – Steve Walsh “Start from the end. Start from where you want to go and work backwards.” – Steve Walsh “Smart capital opens doors you probably never could open on your own.” – Steve Walsh “If you want A players, you need a big goal.” – Steve WalshConnect with Steve Walsh

Matthew Evangelo sits down with host Matti McBride in this episode of the Progressive Dairy Podcast to chat about milking Holsteins and Jerseys, mastitis management and vermiculture. Learn about Evangelo's career in dairy, current management practices and future plans at Bar E Dairy and their commitment to sustainability.  Episode breakdown: 1:00 – Rapid-fire questions 2:00 – Evangelo's background 3:20 – Bar E Dairy 4:25 – Milking cows 2X vs. 3X 6:25 – Milking Holsteins and Jerseys 9:45 – Youngstock management 10:40 – Milking protocol at Bar E 11:50 – Mastitis treatment protocol 13:00 – Advice for improving somatic cell counts 15:00 – Cow comfort 17:00 – Vermifiltration system

By Doug Green “We're a partner-first company. It is in our DNA,” says Patrick Sheehan, Vice President of Channel Development and Distribution at Intermedia. In this episode of Technology Reseller News, recorded for the CCA community, I spoke with Patrick Sheehan, Vice President of Channel Development and Distribution at Intermedia, about a new approach to helping channel partners grow recurring revenue while reducing operational complexity. Intermedia describes itself as an intelligent cloud communications provider, bringing voice, video, messaging, contact center, collaboration and related services into one seamless, AI-powered platform. Sheehan noted that Intermedia supports thousands of partners and more than 150,000 businesses, with a strong focus on helping partners look good and keeping customers happy. The central topic of the conversation was Intermedia's Co-Op Partner Model, a program designed for partners who want to retain control over customer relationships and pricing, while Intermedia handles many of the back-end operational burdens that can slow growth. For many partners, cloud communications, AI, Microsoft Teams integration, and contact center services represent significant opportunities. But traditional models can also add complexity to billing, taxation, collections, support and administration. The Co-Op model is designed to remove much of that friction. With Co-Op, partners can maintain ownership of the customer relationship while Intermedia provides the operational infrastructure behind the scenes. That allows partners to focus on selling, serving customers and expanding existing accounts into new recurring revenue streams, with the potential to earn up to 2X more in profit compared to traditional models. Sheehan also discussed where partners may be leaving money on the table. Existing customer relationships often contain opportunities for voice, collaboration, AI-enabled communications, customer experience tools and Microsoft Teams-related services. By simplifying the path to offer those services, Intermedia is encouraging partners to revisit accounts they already know well. The conversation also covered Intermedia's broader partner-first strategy, including its focus on customer service, technical support and reliability. Sheehan highlighted Intermedia's ninth consecutive J.D. Power recognition for assisted technical support, along with the company's financially backed 99.999% uptime service-level agreement. For partners that have not engaged with Intermedia recently, Sheehan's message was direct: the opportunity has changed. Cloud communications is no longer just about replacing phone systems. It is about helping customers modernize communications, improve customer experience, adopt AI-enabled tools and create more flexible ways to work. The Co-Op model gives partners another way to participate in that opportunity without having to rebuild their own operations. Learn more about becoming an Intermedia partner. Read more about Intermedia's Co-Op program.

Jenna is joined by Mississippi State pitching coach, All American pitcher at Arizona, 2X national champion and ESPY winner, Women's College World Series Most Outstanding Player, and former pro, Taryne Mowatt-McKinney! They talk about the Bulldogs' first WCWS in program history, her own historic run in OKC as a player, coaching multiple pitchers who went pro, full circle playing for and now coaching with the Chicago Bandits, the AUSL, how the game has grown at every level, living life around softball with wife Brittany and son Lincoln, and more. 00:00:00-00:12:06 Intro/Covering Our Bases 00:12:06-01:25:06 Interview 01:25:06-01:26:32 Bring It Home/Outro IG: @bleavinsoftball X: @BleavInSoftball Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

It's official: Knownwell is now part of 2X, and together we're building the leading human-agentic GTM services company, unifying B2B marketing, sales, and customer success. In this special in-person episode of AI Knowhow, Knownwell CEO David DeWolf and 2X founder Dom Colasante join host Courtney Baker at 2X headquarters to break down the news and what drove the acquisition, including: why software and services are converging into a single category and what humans uniquely own in an agentic world. In this episode: The big news: 2X acquires Knownwell to build the first human-agentic GTM services company The trend permeating the services world: customers buy outcomes, not software or services Go-to-market fragmentation, and why marketing, sales, and customer success must run as one motion The 90/10 retention paradox every revenue leader should know Why all work should have a human in the loop Moving Knownwell's commercial intelligence up the stack into sales and marketing Show Notes: Read today's press release: https://2x.marketing/press-release/2x-acquires-knownwell/ Connect with David DeWolf: https://www.linkedin.com/in/ddewolf/ Connect with Dom Colasante: https://www.linkedin.com/in/domeniccolasante/ Connect with Courtney Baker: https://www.linkedin.com/in/courtbaker/

If you've ever undercharged a client because you needed the money, over-delivered on an offer out of guilt, or held back a product launch because it wasn't quite perfect yet — this episode is for you.In Part 1 of this two-part conversation, Jan sits down with her friend and peer mastermind member Andee Hart — sales strategist, mentor, podcast host, and the woman behind She Sells Differently. What started as a candid chat between two Christian women entrepreneurs turned into one of those conversations you have to stop and take notes on.They dig into the scarcity mindset — not as a buzzword, but as something both of them are actively working through in their businesses right now. You'll hear Jan and Andee get honest about the fears that keep women playing small, what the Bible actually says about money and calling, and what it looks like to ruthlessly focus on the work you were made to do instead of the work you can just do.In this episode you'll learn:Why accepting every client at every price point is actually a scarcity decision — and what it costs you long-termHow over-delivering and underpricing can rob your clients of real investment and ownership in their growthWhat Lydia (yes, the Lydia from Acts) can teach us about building wealth as a Christian womanThe 10X vs. 2X framework from 10X Is Easier Than 2X by Dan Sullivan & Benjamin Hardy — and why doing less is often the path to growing moreHow perfectionism and distraction are two of the enemy's biggest tools against women entrepreneursWhy being in a room with people further along than you is one of the highest-leverage things you can do for your businessEpisode Highlights[00:00] Jan introduces Andee Hart and the concept of peer mastermind relationships among Christian women entrepreneurs[02:36] The conversation begins — Andee opens with a question about scarcity mindset and Jan's real experience of taking every low-ticket client just to survive[04:08] Why doing "everything you can do" isn't the same as doing what you should do — and the difference between capacity and calling[09:51] How scarcity mindset shows up specifically for Christian women: underpricing, over-giving, and the fear of charging what you're worth[10:29] Jan gets personal — losing both parents in four years and how grief clarified her sense of calling and stopped her from playing small[11:55] The theological case for not playing small: when we hide our gifts, we rob others of what God intended to serve them through us[13:07] Money, ministry, and the women who funded Jesus' work — a biblical reframe for Christian entrepreneurs who feel guilty about profit[14:13] Lydia, purple dye, and what would have happened if she'd diminished her gifts[16:24] The difference between showing up with excellence and over-delivering — and how to sit with the Lord before you build your offers[19:36] The overthinking loop, the enemy's strategy, and how perfectionism causes inaction for recovering perfectionists (relatable!)[22:04] Jan shares her own current business evolution — letting go of steady retainer income to trust God with the 20% she's truly called to[23:11] Andee recommends 10X Is Easier Than 2X and explains the core idea: ruthlessly eliminate the 80% to do the 20% that actually scales your business[27:32] Being in the room with women further along — why playing up raises your game (the tennis analogy is good)[30:19] Teaser for Part 2: next week, Andee and Jan go deep on pricing — don't miss itKey Takeaway"When we make little of the gifts that God has given us, we rob others of getting to experience a beautiful gift — a way the Lord has designed to serve them through us. You can still charge someone and serve them well."Resources Mentioned

I don't end many podcasts by inviting someone to come hang out at the farm, but this was one of those conversations. In this episode, I sit down with Patrick Engasser, a bestselling author, speaker, and coach who's built and led a seven-figure sales team…all while being blind since birth.  This isn't a “feel good” story. This is a wake-up call. We talk about what it actually takes to succeed when the odds aren't in your favor, how to stay aware of opportunities most people completely miss, and why building the right systems is the difference between being stuck and having real freedom. Patrick also shares powerful perspective shifts around adversity- how the very thing you think is holding you back might actually be your greatest advantage. And yeah…we even get into guide dogs.  If you've been waiting for the “right time” or better circumstances… this episode will challenge that hard.   Find Patrick at www.TalkwithPatrick.com  If I Can Do It, You Can Do It!: Inspiration for Eliminating Excuses, Overcoming Challenges, and Succeeding in Business and Life by Patrick Engasser https://amzn.to/4cIg1VH    Things mentioned in the show: 10X is Easier than 2X by Dan Sullivan https://amzn.to/4mErRok  4-Hour Workweek by Tim Ferriss https://amzn.to/4cB3XFF  The 7 Habits of Highly Effective People by Stephen Covey https://amzn.to/489fc72  Goals: How to Get the Most Out of Your Life by Zig Ziglar https://amzn.to/4twO7mY  The Power of Intention by Wayne Dyer https://amzn.to/4cX7rUk    --- Click here to change your life- http://eepurl.com/gy5T3T   Hit me up for a one-on-one brainstorming session- https://militaryimagesproject.com/products/brainstorming-session-1-hour    Check out my Linktree for different ways to rock your world! https://linktr.ee/ruggeddad    Check out the sweet Hyper X mic I'm using. https://amzn.to/41AF4px    Check out my best-selling books: Rapid Skill Development 101- https://amzn.to/3J0oDJ0 Streams of Income with Ryan Reger- https://amzn.to/3SDhDHg Strangest Secret Challenge- https://amzn.to/3xiJmVO This page contains affiliate links. This means that if you click a link and buy one of the products on this page, I may receive a commission (at no extra cost to you!) This doesn't affect our opinions or our reviews. Everything we do is to benefit you as the reader, so all of our reviews are as honest and unbiased as possible. #passiveincome #sidehustle #cryptocurrency #richlife

We're announcing AIEWF speakers this week! Take the AI Engineering Survey!Today's guest Ethan first joined us for the LS Paper Club as the lead on NVIDIA Cosmos World Model, but then joined xAI and built Grok Imagine in 3 months:He comes back on Latent Space with some nuclear hot takes: that Video Models primarily get their intelligence from LLMs, not from training on video data, and that the next frontier for truly interactive, realtime, long-horizon world models is to work on LLMs (perhaps Interaction Models as well…)Put it this way: In the near term, the next Sora won't be a better video model, but a video agent.Generative Media may more closely follow the evolution of AI coding which went from focusing on one-shot output performance and cost, to multiturn reasoning and planning models for agents and systems that can plan, edit, test, debug, and submit PRs.At a certain point, coding models got so good that the only significant next step to improve performance was handling the orchestration of these models.Now as the performance of video models increases significantly across realism, consistency, & prompt adherence while becoming more cost efficient, the next evolution of video generation may also be systems that can plan, generate, edit, critique, and iterate across an entire creative task. In this episode, Ethan joins swyx and Vibhu to unpack what it actually takes to build frontier image and video systems: data, VAEs, diffusion transformers, audio-video alignment, inference speedups, and the hidden cost of storing and moving massive video datasets. From building NVIDIA's Cosmos world model to joining xAI as Grok Imagine was being built from zero to one, Ethan He has been at the center of some of the most important work in video generation, multimodal models, and real-time world models.We go deep on Grok Imagine, how a small xAI team shipped its first multimodal video model in three months, why iteration speed matters more than almost anything in model development, and why many of the biggest gains come from fixing tiny bugs in data and training pipelines. Flipbook: The future of VideomaxxingVideo agents are almost a sure bet to be the trend in the coming year. We end with a glance at what's beyond video agents:Flipbook caused a minor sensation this year when it was released, but most treat it as a fun demo. Ethan takes it very seriously — with the speed and cost of inference coming down every year, the future of custom video JIT UI is closer than you think. We talked about why videogen models may become the front end of AI, how generative UI could replace traditional HTML/CSS, why world models need to be real-time, interactive, and long-horizon, and why the future of video generation may depend more on language models and agents than on diffusion alone.We discuss:* Why fast iteration mattered more than meetings* Why small training bugs can drive huge model quality gains* Why coding models may make compute the bottleneck again* How image and video models are trained with synthetic captions* The role of VAEs and latent space in frontier video models* Why image models are the foundation for video models* The tradeoff between temporal compression and real-time interactivity* Flipbook, Neural OS, and the future of generative UI* Why future interfaces may go from user intent to pixels* The hidden cost of training video models: storage, egress, and GPU hours* How step distillation and consistency models (like OpenAI sCM) makes video inference orders of magnitude faster* Grok Imagine 0.9 and large-scale audio-video generation* Why audio-video alignment is harder than text-video alignment* Ethan's definition of world models* Reference-to-video, video extension, and long-context video generation* Why xAI's research communication undersells Grok Imagine* How xAI culture shaped the speed of development* AI watermarking, SynthID, and detecting generated media* Why prompt rewriting matters for video models* Grok Imagine Agent and the rise of video agents* Why language models may unlock better video generation* Robotics, physical AI, and embodied world models* Why Ethan left xAI and shifted focus toward LLMs* Self-managed context, memory, and the next frontier for language modelsEthan He* LinkedIn: https://www.linkedin.com/in/ethanhe42* X: https://x.com/EthanHe_42Timestamps00:00:00 Introduction00:01:25 From NVIDIA Cosmos to xAI00:03:24 Building Grok Imagine from Zero to One00:10:07 How Image and Video Models Are Trained00:18:53 Video Compression, VAEs, and Real-Time Tradeoffs00:22:10 Generative UI, Flipbook, and Neural OS00:32:10 The Cost of Training Large Video Models00:37:04 Distillation, GANs, and Fast Video Inference00:41:21 Audio-Video Generation and Grok Imagine 0.900:48:34 What Makes a World Model?00:55:51 Reference Videos, Long Context, and Video Memory01:00:11 xAI Culture, Research, and First-Principles Building01:09:45 AI Safety, Watermarking, and Prompt Rewriting01:13:10 Video Agents and AI-Assisted Creation01:27:32 Why Language Models Unlock Better Video01:31:15 Robotics, Physical AI, and Embodied World Models01:32:38 Why Ethan Left xAI01:34:16 Self-Managed Context and the Future of LLMs01:38:43 Ethan's Career Path and Closing ThoughtsTranscriptIntroduction: Ethan He, Latent Space, and the Path to xAISwyx [00:00:00]: We're here in the studio with Ethan He, most recently of xAI. Welcome.Ethan [00:00:10]: Thank you. Glad being here.Swyx [00:00:11]: We're also here with Vibhu. you were first coming to us or joining the latent space world because you were working on Kosmos at NVIDIA, and you did a paper. We loved it. you presented it as well, so thank you for doing that.Ethan [00:00:23]: I've actually, I also presented the MoEs twice at latent space.Swyx [00:00:29]: How did you actually hear about us? Did we reach out to you? Is that how it worked?Ethan [00:00:33]: No, actually, I-- the community. Like I realized, oh, there is this online community that people talk about AI and also learn from each other through papers every week through the Paperclip. It's very nice.Ethan [00:00:49]: I learned a lot.Swyx [00:00:49]: I think three years stop. We haven't stopped even on Christmas and New Years. many weeks I want to stop but it keeps going.Vibhu [00:00:58]: No, that was good. I think you had posted that you worked on a paper, and I was “Oh, very cool. We have Paperclip. Present then.”Vibhu [00:01:04]: But I might have reached out to you after.Swyx [00:01:05]: you-- because it's an amateur club, right?Swyx [00:01:08]: so it's very unusual and but we have sometimes paper authors come by and actually explain the paper. Today we just did, the poolside paper, which was apparently very good.Vibhu [00:01:18]: Came out yesterday.Vibhu [00:01:19]: pretty interesting, right? Fully open. They talk about everything, systems. So it's a good one. We'll, we'll recommend people to read it.Swyx [00:01:25]: Bring us up to speed on your transition to xAI, ‘cause I actually don't even know when you joined. just like tell the, tell the story about the sort of transition.From NVIDIA Cosmos to xAI: Scaling Video and World ModelsEthan [00:01:34]: Before xAI, I was working on Kosmos world model as in-- at NVIDIA. So Kosmos is, it's a giant video foundation models that can-- that aims to simulate the world and for-- it serves as a foundation of-- for all of the roboticists to build on top of. There, once I built the Kosmos one, I realized as this thing also has a scaling law similar to language model, we need to scale up the video models further. that's, that's why I realized I need to move to somewhere with much more compute resources. That's how ISwyx [00:02:13]: Than NVIDIA?Vibhu [00:02:14]: The GPU rich came themselves.Vibhu [00:02:19]: And timeline-wise, when was Kosmo? It was pretty early, right? It was open world model, open paper, everything.Ethan [00:02:25]: It was end of twenty-four.Vibhu [00:02:28]: End of twenty-four.Ethan [00:02:30]: Then at mid twenty-five, I moved to xAI. At that time-- I joined about the time when xAI was about to build video models and in multi-model models. There were no infra, no data, and no model, and it just-- as a few engineers, we built it in three months and released the first model, Grok Imagine zero point nine.Ethan [00:02:55]: And since then, I keep working on video models and move more from training and to post-training of the video models. For example, like a reference to videos, kind of like the cameo feature and, video extensions. And, before I left, I worked on a world model, leading a small team to focus on the real-time long horizon video generation.Building Grok Imagine From Scratch in Three MonthsSwyx [00:03:24]: Can you give like a rough roadmap of okay, you're on a brand-new team. Grok previously was only text, or they partnered with BFL for their image gen stuff. What do you-- what are the building blocks, right? You have compute, data you can procure somewhere. Like just what are like the sequence of things that people should think about when you're setting up a new team?Vibhu [00:03:43]: actually even deeper, not just data you can procure. You guys had to go through getting the data too, right? So you shipped it pretty fast, but yeahSwyx [00:03:51]: three months is likeVibhu [00:03:52]: From everythingSwyx [00:03:52]: actually like very surprisingly fast.Ethan [00:03:55]: One thing I say like thanks to my experience at NVIDIA, ‘cause first time when we were building Kosmos together, we built it, for about a year. So this is like the second time I do it. Roughly have an idea, what to do. I say the most important thing is the talent. Everyone were very strong and clever, very close with each other towards a common goal. So that speed up things a lot. So you reduce the communication bandwidth among people, and everyone can work towards the same goal. It's, it's like every day there's not that much meetings on the calendar, like maybe like a, like a sync a day, and after that it's, it's just all building. It was pretty fun at that time.Ethan [00:04:47]: And another thing is that xAI has very strong foundations of like data inference, model inference, and the supporting there can help the model develop a lot. When I look at, training models, I don't so actually the top important thing is like how many, how many iterations can you do, per day? and the more iteration can you do, you can, you can train the model much faster. So if you have very strong infra and you have a lot of compute, you can, you can train these models in very short period of time. That can give you a much larger buffer to, for errors, and it also gives you the opportunity to spot more bugs.Iteration Speed, Compute, and Debugging Model PipelinesSwyx [00:05:46]: What is an iteration? Is it like a few hundred steps or what are youEthan [00:05:50]: Let's say just the train-training the model, like from acquire new data and maybe design new algorithms and train a new model, maybe at smaller scale orSwyx [00:06:01]: So cycle time for like any hyperparam that you're searching.Ethan [00:06:04]: Cycle time and tune to like eval this model. Is this model better than my previous iteration?Ethan [00:06:11]: SoSwyx [00:06:11]: So it's like before you, someone had already set this up that you can iterate very quickly.Ethan [00:06:15]: I think the foundation there is extremely good forDeveloping and research models.Ethan [00:06:23]: And often I find is it-- this is kind of boring, but like a lot of the improvements does not come from new algorithms. It comes from finding small bugs here and there in the data pipeline, in the, in the model training pipeline. Those give, those give the biggest boost to the model quality.Vibhu [00:06:46]: It's interesting, right? So you say it's like small team, less communication bandwidth, but also a lot of quality is like find little bugs. It seems counterintuitive, right? You have a lot of people, you can iron out more of those, but it's interesting to see the other side, right?Swyx [00:07:00]: I also wonder, have you-- do you try using LLMs to look for bugs? I don't know.Ethan [00:07:05]: I remember at that time it was mid two thousand and twenty-five, so it's the coding model wasn't quite there yet. I remem- I remember like December two thousand and twenty-five, it was extremely good. Yeah, I've been, I've been using it at that time. It's, it's helpful. sometimes it produce codes that are kind of difficult to maintain, even though like the first time it built something extremely fast. But it gave the, like a spaghetti code, thousands of lines that I couldn't maintain, and the LLM itself couldn't figure out what's, what's wrong and how to improve on top of it. But now I find it much better. Yeah, I want to bring up another point here is now coding models are much more efficient and can help us implement stuff much faster. Compute might become a bottleneck again because previously, like if you want to train a new model, say you want to generate new synthetic data and then or write a new algorithm, it might take a few weeks. And during that period of time, you don't-- you might not have experiments to run. But now you can build that thing within a few hours, then you can immediately train a model.Ethan [00:08:24]: Now you have to have enough compute to try all of the ideas. So compute might be the bottleneck of iterating speed again.Swyx [00:08:36]: yeah, I actually, honestly, I think it's like kind of a stressful job because you're “Well, I should be trying everything, and if I'm not, then I'm not doing my job well.”Vibhu [00:08:48]: there's also the stress of you're eating thousands of GPUs per hour, which is very expensive and, compute can go to other researchers.Swyx [00:08:56]: You got the daddy Elon toVibhu [00:08:57]: You got daddy Elon.Ethan [00:08:59]: It wasVibhu [00:09:00]: But there's still finite amount of compute, like you want to use it, you want to use it well, you want more of it.Ethan [00:09:06]: That was quite stressful indeed. Yeah, I think one thing is the-- with coding models now, like a lot of these jobs can be automated, which is much better. A second, it's a, it's a marathon, so you got to maintain good health and, a regular schedule.Vibhu [00:09:28]: It's, it's hard to hear that when you shift from zero to nothing in two months.Swyx [00:09:32]: and, I think obviously the culture at xAI is very famously, people work very hard. one thing I did want to dive into, in our-- in the notes that you, that you sent ahead of time, you had specific comments about the cost of Video Gen training. presumably this is on the Colossus-1, right? the two hundred megawatt cluster. Any whatever you want to just share on that.Vibhu [00:09:54]: I think there's, there's three things we're talking about, right? So there's Video Gen, there's also the Image Gen model that you put out. Do you want to like complete the, okay, so zero to one, you have a few months. Just what are the stages of create Image Gen model?Swyx [00:10:06]: Oh, yeah, maybe I got distracted.How Image and Video Models Are Trained: Synthetic Captions, Tokenizers, and VAEsVibhu [00:10:07]: Sorry. and then, from there's Video Gen, there's Audio Gen. Would love to get into those next. But what is that first few months like? So small team, a lot of bugs, iterations, but what does it look like? Do we take something off the shelf? Do we just get data compute? What's, what's the few months like? How do you go to state-art Image Gen model? How do you just start?Ethan [00:10:28]: I cannot comment specifically how xAI did, but it's, it's a quite standard process. I can draw some, examples from Cosmos. So mainly it's building a video model, you actually need to build a image model first. And building these two models, the data you need is a hundred percent synthetic pair of language and image or language to video. Because on the, on the internet, actually, the videos don't naturally associate with text. So you can say, oh, like on YouTube, you have the title and you have the description and the commentsSwyx [00:11:11]: TitleEthan [00:11:11]: of a video, but usually they're not relevant to the video itself. And say maybe like the video is a natural scene of mountains or something, and the title is, I'm so happy today.Ethan [00:11:26]: So they have they have no correlation at all. So the first step is to, you have to generate synthetic pair of language with the videos. So you gather videos from the internet, and you use a VLM to caption the videos. So that part, here's a question, like how do you, how do you gather VLM to begin with? So if there's noSwyx [00:11:55]: You, so you fuse the model, right? LikeEthan [00:11:57]: Say if there's no like VLM exists, like how do you generate the text to the beginning, right? It's, it's impossible.Swyx [00:12:04]: I see.Ethan [00:12:05]: In the beginning, it's like you ask human to describe the video as detailed as possible.For example, you ask them to describe everything, like all objects, all characters, and all interaction and dialogues in the, in the videos. So that's in the protocol of Cosmos labeling. We require the objective we give to the labelers was that you have to describe the video as detailed as possible, such that a blind person hears a blob of text can reconstruct what the video is like from their head.Swyx [00:12:43]: Video or image? You're talking about images.Ethan [00:12:44]: Video or image, either one of them.Vibhu [00:12:47]: This was pretty common when we went from clip and DALL-E, right?Vibhu [00:12:51]: It's all training on really detailed captioning of images. So same is applied to video, but insteadEthan [00:12:57]: same appliedVibhu [00:12:57]: of using multimodal model to pass in video images and write rich descriptions, you can alsoSwyx [00:13:04]: I think there's this traditional perspective of supervised, or, very highly human curated thing. I feel like there's a unlock with unsupervised, right? Where like you have enough to bootstrap that you can just throw common corpus on it or, whatever. like unsupervised vision and language pairing, right? Like where you just have, interspersed image and text and it just learns. To me, that is the VLM breakthrough that is different from the clip, different from the LM era.Ethan [00:13:36]: It's interesting to see that you kind of need both data.Ethan [00:13:41]: For example, for theSwyx [00:13:41]: You need it to bootstrap it up. YeahEthan [00:13:43]: for the generative model training, there's also usually like a small percentage of unlabeled data. So the model is instructed to generate a video without any text instruction. That can also help the model generalize. So after this stage of generative synthetic pair, so, one important common step is to train a compressor or a tokenizer of the image or videos. So because, if you train-- If you can technically, theoretically train image or video models on pure pixels, but the problem is that the, it's, it's a lot of tokens. So like one image, it's, a thousand by a thousand, it's like one million tokens, one million pixels. It's impossible to train transformer on that. So it's, you need to train a tokenizer, which can go from image to latent space and latent space back to image.Swyx [00:14:45]: That's why we named the podcast.Swyx [00:14:48]: But, basically, you're talking about vocabulary science.Ethan [00:14:50]: so vocab.Swyx [00:14:51]: And so, what is, what is imp-- like a million is impossible?Ethan [00:14:54]: In generative models, the vocab is continuous. It's a continuous space. We can think about like you map an image to a vector. It's a, it's a fixed length vector. It's sixteen or forty-eight, something like that. And then you map that vector back to the image space. And the mapping is, has-- The mapping is patch-based. So you say you haveEthan [00:15:22]: a sixteen by sixteen patch and you match, you map that patch of pixels into this latent space.Swyx [00:15:29]: We've covered thisVibhu [00:15:30]: This is like the vision transformersSwyx [00:15:32]: VAEs,Ethan [00:15:33]: VAEs.Vibhu [00:15:34]: You basically compress your input, you do your generation, you're reasoning all that generation in smaller dimension, and then you project back out.Swyx [00:15:43]: VAE is a form compression, but I think the for me, the patching thing is from VIT, right?Ethan [00:15:48]: You can make those.Swyx [00:15:49]: Literally the, yeah, the paper is titled like sixteen by sixteen is all you need. something like that. and then I think also, people make a lot of comparisons with this kind of patching with convolutions.Swyx [00:16:02]: Which is you're, you're kind of re- reconstructing the old paradigm with the new.Ethan [00:16:05]: Actually, in VAEs, there are, there are both convolution networks and transformers. You can actually do both.Ethan [00:16:14]: After this VAE, so what you've got is you've got latent space tokens and you've got the language tokens. So now the training of the diffusion transformer, usually generative models use diffusion transformers. It is actually quite standard. It's, it's very similar to how you train a language transformer models. It's not that much difference. It's just the tokens, the visual tokens in, visual tokens out. The only difference is there's a denoising process. So you train the model to unmask some of the noise. So you add, you add random noise to the visual tokens, and then you train the model to remove those noise to generate the clean tokens. Any inference, the model can iteratively remove noise from a hundred percent noise.Swyx [00:17:12]: And then there's also, to speed things along on the tech tree of diffusion, there's CFG, and then there's, there's also, latent diffusion that, there's, there's someone in there. I think, somewhere along the line, obviously, like stability and all these other guys, pioneered a lot of this, architecture. I don't know if you want to get into that or just, or do the video side up to you.Bootstrapping Video from Image Models and Temporal CompressionEthan [00:17:37]: After you train such model, such image model, the reason it's a, it's a foundation for video models is that image models are cheaper to train, and they have much denser connection between language and text. So, sorry, language and images. For example, you train a billion, you train on a billion images, and there's a mapping from the text to the image. And the cost to train the same, like the, a billion, a billion text to a billion videos, that's much more expensive because videosNaturally have more tokens than images. Because the diffusion models, their understanding of, language purely come from this mapping. So if you don't have enough mapping, so if you only train on like a ten million videos or something, there-- you might not see enough language tokens in your training, so your model does not understand human intention enough. So that's why you really-- you train-- you first train this image diffusion models, and then you bootstrap the video model from there.Swyx [00:18:53]: One thing I did want to ask, because I-- actually, I think you're, you're the first per-- video model person I've ever talked to, I think. we've, we've like talked to Luma and all those folks. There's all these tricks in video compression where basically frame by frame there's not that much difference, so actually you don't have to regenerate or save the whole frame, right? but I think MP4 compression or something else like that.Swyx [00:19:16]: is it tempting to use that? Or as far as I can tell, everyone just treats it as, “No, we would just generate every frame.” Is that roughly the state-art?Ethan [00:19:27]: There are a few different approaches. Let's say first, like you want to just directly use MP4 compression and use that as the tokens for the transformers to train, right? So people actually have tried that, but the main challenge is the latent space for the MP4 tokens were not, were not very comprehensible for the models. It's, it's extremely hard to train on that. And there's aEthan [00:20:01]: So that's why they created VAEs, which creates more continuous, latent space, so the models can understand that latent space and learn from it much easier. Even within the VAEs, there are different difficulties of the latent space. So you can imagine something the simplest, the most naive VAE is like you have an image, and you just shuffle all of the images into a, into a vector. So you don't need to train any VAEs, right? But that latent space is extremely hard for models to train on top of. That's why there are some debate on like how do you compress the tokens. So you mentioned like you can compress frame by frame. Also, you can compress, the temporal dimension.Ethan [00:20:52]: The difference is if you compress the temporal dimension, you get a much higher compression rate. Because there's temporal redundancy between frames, because, this frame and the last frame, likely they are mostly similar, so there's only some small difference. for example, I think in 12.1 VAE, they have like a eight by eight by four compression rate. So the four temporal tokens are compressed into one tokens. That can save a lot of, save a lot of the context length. If you do it frame by frame, you have to do maybe like eight by eight by one. Your context length will be four times larger. That being said, the benefit of the frame-- per frame compression, we might come back to this later, is, real-timeness and interactivity. ‘Cause if you, if you strain the output of the model, frame by frame, you can-- the model can respond to any user request immediately. So if you have like a temporal four compression, four times compression, thenSwyx [00:22:06]: It might be laggyEthan [00:22:07]: there's a lag there in nature.Swyx [00:22:10]: So you're very pilled on this. let's just go ahead and bring it up ‘cause we have the visual prepared anyway. There's some frontier applications of real-time video gen. So Flipbook is one of the examples that went viral recently, right? What is Flipbook?Real-Time Generative UI: Flipbook, Neural OS, and Diffusion Front EndsEthan [00:22:23]: Flipbook is kind of like a web brow- web browser. You can see like it has the web bro- browser UI on top. The difference is all of the UIs are generated by generative image model in real time, and anything here are fake. But you can, you can explore inside this wor- this imaginary world. Say like we-- here we have engineering the Great Pyramid. Like the model generates this for us to understand how it works, and if we want to navigate around and understand further, we can click on some of the, some of the description here, and the model will generate a new page, new subpage describing the details we want to know about.Swyx [00:23:14]: So it's basically kind of we're playing a video, but it's pausing for our next interaction, and then it just plays the next thing based on our interaction.Swyx [00:23:23]: Which is kind of cool.Vibhu [00:23:25]: and you kind of decide your story. So this was, how do you make a pyramid? levering technique seemed interesting, right? It shows how do you take Okay, I want to know what is thisSwyx [00:23:35]: The demo, the demo tweet had more animation between frames.Vibhu [00:23:38]: I think it's just skipping,Swyx [00:23:39]: Oh, it's just skipping a lot of frames.Ethan [00:23:40]: they also have a video modeVibhu [00:23:42]: It takes a lot. There's a lot of peopleEthan [00:23:42]: but, a lot of people are using it.Ethan [00:23:45]: So it's not available.Vibhu [00:23:46]: There's a live video stream. We can try,Swyx [00:23:50]: So this is an example of the kind of future that you see at the extreme. We don't-- we're obviously not in it today.Swyx [00:23:56]: But in a world where inference is completely free this is better than generating code and text?Ethan [00:24:02]: So this is, this is a final state of where Viva will be at for word model, I think. Imagine internet doesn't exist, and then you type in google.com. Like what should, what should, what should a model show you?the model can imagine something, and this is what the model imagine. And these web pages, they completely do not exist. So I think as the inference costs come down, we are going to have generative UI for everything. If you think about how the coding model works, so they write code for a web page, and they render the code might be con- converted into binary, and the binary render the pixels on the screen. So we in machine learning, every time we have some breakthrough, obviously it's, it's more intuit. So why don't we have like user instruction to the pixel directly? So the generative UI will be user intention to the pixels directly. And say like even if I want email, let's say everyone have the same interface, but I want, I want it slightly different. I want the email to show to me like a TikTok, so I can swipe left and right for the emails. And or maybe you want something else. We can have completely different things. Or like I have I'm looking at, Instagram stories, and I don't like the Like button. I always may click it. And, generative UI resolved it. So it's going to be a revolutionary replacement of the interface. So in the future, we might have much more powerfulEthan [00:25:50]: LLMs and coding models running behind the scene. And in the, in the front-end, the diffusion model will actually be the front-end to show stuff to you. That's how I imagine it.Swyx [00:26:02]: Diffusion front-end, deterministic back-end.Swyx [00:26:04]: Something like that. I find that very expensive, but,Vibhu [00:26:08]: I find it interesting you called LLMs writing code on the back end deterministic, but okay.Swyx [00:26:14]: you write it onceVibhu [00:26:15]: Compare it toSwyx [00:26:16]: And then you execute.Ethan [00:26:17]: If you think about the cost, say, let's say H100 costs $1 per hour, and if you use this eight hours a day and thirty days, so, every month you're paying this two forty, you'll actually not wanna pay for that. That's even more expensive than Cloud Code Max. But if you think about the compute costs come down like two times every year, and I think the future will likely arrive like within few years.Vibhu [00:26:49]: It's everything, right? compute cost comes down, compute gets faster, model gets smarterEthan [00:26:54]: More efficientVibhu [00:26:54]: model gets smaller.Swyx [00:26:55]: I don't know why you say two times, ‘cause I think it's like 100 times. In language models, it is roughly one hundred to a thousand times every twelve to eighteen months, for the same given level of LMSys, ELO.Vibhu [00:27:08]: That's a net of everything, right? That's model performance alongside compute. So different than just compute costs come down. But, a very interesting future.Swyx [00:27:19]: So the web designers will have to shout out that accessibility is an issue, right? how do you deal with screen readers or whatever. But yes, this is higher bandwidth storytelling than anything you can possibly generate with code, right? So I think that's the rough idea.Ethan [00:27:34]: And I'd like to add a little bit that so human naturally have the maximum bandwidth when we are looking at things, look at videos, and we also have maximum output bandwidth when we are talking. So in the future, it might be something like we talk to AI models, and the AI model responds back with a generative UI. So that would be the maximum input and output bandwidth to interact with AI models before neural link happens.Vibhu [00:28:06]: And it's also very custom, right? Some people are very visual, some people are not as visual, right? They prefer the text. But the best thing about generative UI, right, it can also be text.Swyx [00:28:17]: There's another project that we wanted to highlight, which is the Neural OS. Kinda similar idea, but here you're literally operating, simulating an operating system with a video model.Swyx [00:28:27]: and you can play Doom, you can do Firefox. I find this like mildly less impressive, obviously, because it's an OS that I can run.Swyx [00:28:37]: But here everything is imagined.Vibhu [00:28:40]: I was, used to the Command+W to close the Firefox tab. It didn't crash. That's why I saidSwyx [00:28:45]: It's too immersive.Vibhu [00:28:46]: It's, it's too immersive for me.Swyx [00:28:47]: Too immersive.Vibhu [00:28:48]: I wanted to close the tab.Vibhu [00:28:49]: But yes, I can play generated diffusion.Swyx [00:28:51]: this is shockingly fast.Swyx [00:28:54]: Because I remember there was a demo about like maybe one to two years ago. Someone tried to do the first-person shooter with a image model. There was no consistency. It was very slow. But here it looks like realistically it's-- this is Doom.Vibhu [00:29:07]: I think there's two sides to that, right? There's okay, what is running a game? The heavy part of it is actually the game engine, all the lighting, all that stuff, the graphics. This is just kind of video, right? Like we've solved consistency. This is still, it looks like a few years old image generation. There's some temporal consistency, but it's, it's kind of just images stitched together as frame video. But it's a good visual representation to pi- to picture the future you wanna see, right? that's, that's what I see in these more so.Ethan [00:29:38]: This reminds me of how the video models gets better and better. So Neural OS is kinda if you just look at it feels like it's just a crappy version of the, like the Windows we could have, right? And, but the difference is, so the model, this model is overfitted on the existing operating systems. It can generate nothing different than that. But it's actually also similar to video models. So when we are training these video model, image model, we train them on internet. There's no imaginary supernatural stuff on the internet. But once we train this model, you can prompt the model to generate something supernatural that have never existed in the data set. So if you train your Neural OS or neural computer on the standard screen recordings on the entire internet. The model can imagine completely new interface to interact with the computer.Swyx [00:30:43]: This is one of those things that is magical to me. usually generalizing out of distribution is bad, but somehow we have learned some kind of internal world model that you say, this plus, but it looks like rainbows and butterflies, it'll do it and it will kind of make sense.Swyx [00:31:03]: So yeah, that's kind of cool. Yeah, I don't know if there's any comment more on there. I do, I do wanted to, I did wanted to touch a little bit more on the model architecture stuff, which I think you were getting. It's, really fascinating. We don't get a chance to talk about this enough. So one of the papers that we covered, we've covered every annual, segment anything release. and I don't know if you follow-- you're a computer vision guy, so youEthan [00:31:26]: I knowSwyx [00:31:27]: . So they did memory attention, which is kind of interesting. And I always think, anything where you can, across the temporal dimension, keep some consistency, I think it's, very fascinating, and I don't know if Basically, does that-- the CV side bleeding into video gen side, I think is underexplored, right? we talk about it for labeling, but actually you can borrow the architecture itself.Ethan [00:31:50]: There's, there's also complete different approaches, right? you brought up the term world model, so we went from video model to world model. There is diffusion, but there's also other approaches that people are doing. So maybe we get into those after as well,?Swyx [00:32:03]: He has a whole definition of world models and stuff. I feel like we threw a lot at you. Whatever you want to comment on.Why Video Models Are Expensive: Storage, I/O, and Training ScaleEthan [00:32:10]: I think one thing that we should actually comment back on is okay, so we were talking about the steps to train image gen to video model. One thing we don't see as much of is okay, you brought up the delta in training data, right? SoEthan [00:32:24]: you won't have as much a video model might not generalize, but what is the cost of training a large video model? So we know for LLMs roughly, okay, even like the poolside thing that came out today, right? It's a Gemma level model trained on roughly forty trillion tokens at this many H200s over this much time, right? You can see what is the exact cost of that. So how many GPU hours over how much H200 costs? So how do we do the back-end math of, same thing for video models, image models. How do you, how do you kind of break that down? I can share some back-envelope calculation. So surprisingly, video models is-- the cost is very-- is comparable to language models and obviously the largest scale is language model, maybe like a medium scale to language models. I said just storing the videos alone, it costs a lot. You can, you can maybe look up on AWS or something.Ethan [00:33:20]: You really, say if you have a billion videos and let's say, let's just say like each video, like five megabyte, then you need five petabyte to just store those videos. And also remember we talk about you use a VAE to compress the videos, and you also need to store, typically you need to store those continuous feature, in-- also in your storage. That's also comparable size with the videos themselves. So just storing these videos and the features is tens of petabytes alone. And,Swyx [00:33:58]: I just, I just looked up the calculation. Five petabytes on S3 Standard is one hundred K per month.Ethan [00:34:05]: AndSwyx [00:34:05]: It's comparableEthan [00:34:05]: and you needSwyx [00:34:06]: AndEthan [00:34:06]: And then like tens of petabytes, two hundred K. And even more expensive is you have the ingress and egress.Swyx [00:34:13]: Oh, yeah.Ethan [00:34:14]: Like you-- through the internet. You have to just to download those videos, I believe it's, it's more expensive on AWS than just storing those videos.Swyx [00:34:25]: Storing, yeah.Ethan [00:34:25]: And each training runs, you probably need to pull them once. If you train multiple times, it's, it's even more than that. So it's like just storing the network, those costs is just, it would be a few, a few millions per month to just storing everything, not to mention the GPU cost.Ethan [00:34:45]: AndSwyx [00:34:45]: my side tangent, the compute rental, like GPU rental is very efficient. There's one side, okay, you can be XAI and build your data center. Should we not just build our, storage compute as well? LikeEthan [00:34:57]: Of courseSwyx [00:34:57]: cloud cost compared to just,Ethan [00:34:59]: You save so muchSwyx [00:35:00]: store. Yeah, exactly.Swyx [00:35:01]: Especially with like egress and stuff. So.Ethan [00:35:04]: That's a good idea, but it also comes to-- there are some of its own challenges.Swyx [00:35:09]: Of course, of course.Ethan [00:35:10]: like people who build the GPU data centers, they might not expect this much, storage. And yeah, people build storage, typically they just build it somewhere with just CPUs.Swyx [00:35:23]: I just looked it up. Five-- AWS only charges for egress, not ingress. Tier five for five petabytes is two hundred and thirty K.Ethan [00:35:32]: Even more expensive than the storage.Swyx [00:35:34]: But storing is per month, right? You check in, then you cannot check out. so it's so cool. It's okay. So there's that side.Ethan [00:35:41]: So the TLDR, my backhand mathSwyx [00:35:42]: Data is larger than you think. Yes.Ethan [00:35:44]: my backhand math of GPU hours times GPU cost is also very much, I'm missing some storage.Swyx [00:35:49]: You're also-- you're basically like also more IO bound than normal training.Swyx [00:35:55]: Yes. ‘Cause like data loading, so caching everything, it becomes super important.Ethan [00:36:00]: So in Cosmos, we did a lot of optimizations to make it not IO bound. So, speaking of the training, actually training the model, the GPU cost, if you look up like the open source model, how big these video models are, I think like LTX has nineteen B parameters. That's a dense model. And people are also exploring, MoEs, so it might be twenty B active and, like a hun- hundreds B, total. So that's, that's even-- that's similar size as medium-sized LLM models. And if you, if you look at number of tokens-Uh, we disclose that in Cosmos. It's also like tens of trillions of tokens on the visual tokens. So putting this together, the cost of, training these video models, it's actually comparable with LLMs. Not to mention, the infra is slightly different from LLM, so it might be less efficient to train these models.Inference Speedups: Step Distillation, Consistency Models, and GANsSwyx [00:37:04]: Do you get the benefits of traditional diffusion speed-up? So for, images, there's LCM, LoRAs for, fine-tuning. There's, there's a lot of stuff that's beenEthan [00:37:15]: Flow matching.Swyx [00:37:16]: there's flow matching. There's a lot of stuff that's been done. there's some overlap that applies to diffusion on the inference side and stuff or?Ethan [00:37:23]: so the difference-- the inference side is a completely different story.Ethan [00:37:28]: I think for the training side, it might be a little bit hard to reduce that cost. And for the inference side, the biggest gain is from the distillation of these models. You can-- It's called step distillation, slightly different from knowledge distillation in LLMs. So you-- Typically, for flow matching models, you need like 100 steps or something. Like a distortion model even need even more, like 1,000 steps to generate a good image or video. A step distillation is try to learn to generate fewer step from the model itself. It's kind of like now we-- you use the full model to generate in 100 steps, and then you take a model that only generate 10 steps and let that model to learn from the perfect one.Ethan [00:38:25]: why this workSwyx [00:38:27]: Strong to weak seemingly.Ethan [00:38:28]: It is. It's kind ofSwyx [00:38:29]: DistillationEthan [00:38:29]: kind of like strong to weak. the-- from the modeling perspective, the strong model, the teacher model is trying to model the image and videos of inter-internet, and that distribution is extremely complex. But the step distilled model is just trying to learn from the teacher. The teacher is a model, and the size is fixed, as the distribution is much simpler than the whole internet. That's the intuition I have why step distillation can work. So usually these models serve in productions, they only run in a few steps. In Cosmos, I believe we have, we have like four step and eight steps. If you do some simpler task, image-image translation, it can even run in fewer step, like one step in Cosmos Transfer.Swyx [00:39:22]: I think this is the same intuition that guides a lot of the consistency model work. I sent you a link for, SCM. I don't know if you covered that. To me, that was actually one of, the most impressive papers I've ever seen from OpenAI.Swyx [00:39:34]: That this is the unifying grand concept of consistency models. I don't know if you have any comments on this.Ethan [00:39:41]: So there are, there are a few different approaches,Swyx [00:39:46]: Oh, yeah. Here it is.Swyx [00:39:47]: Two steps versus twenty or 100 steps, whatever. It's already done.Ethan [00:39:52]: So there are, there are a few different approaches, for example, consistency model, and there are also Actually, we shouldn't forget GAN. So GAN, actually, that was, that was the OG ofSwyx [00:40:05]: OGEthan [00:40:05]: step distillation ‘cause it trained just one step to begin with. So actually, a lot of, uh-- For example, there's a distribution matching distillation which use, which uses GAN, as one of the laws for distillation. It-- GAN just tells you, “Hey, generate an image,” and thenEthan [00:40:31]: it has a discriminator to tell, is this image real or not? So the model, the model just need to learn one of the distribution, not the full distribution. Because in training, the model is asked to reconstruct the ground truth image from the internet, which is extremely hard. And in-- When you're training GAN, it's a step process. It's just a, “Hey, you generate image. Does this image look as real as the image from the internet?” Which is a much simpler task. And, yeah, combining a lot of these approaches together, people typically do that, like consistency model and distribution matching and GAN, and we can get these few step models.Audio-Video Generation and Time AlignmentSwyx [00:41:21]: Then there's one step I wanted to add, which is audio and video.Ethan [00:41:26]: So, Grok Imagine zero point nine, I believe it's, it's a first audio video transmodel deployed at a large scale. SoSwyx [00:41:39]: And that was your first model?Ethan [00:41:40]: that was, Grok Imagine's first model. It's, it's audio video, joint generation. I think the hard part is, the modality alignment, ‘cause before this transmodel, we have, we have text to video alignment. We have this, correspondence between text and video. Typically, most of the VLMs, they understand images and videos. Video's very rare, and they don't understand audio mostly. And if you look at the audio generation on the LLM side, you can talk to them perfectly fine, but if you ask them to sing a song or something, it typically is not very good. Also, they don't have, they don't have music either. The hard part is thatUh, actually audio has two component. It has like a discrete component, a continuous component. The discrete component is like the language.Ethan [00:42:44]: So when we speak, it's just, someSwyx [00:42:47]: It's an ASR issue, yeah.Ethan [00:42:49]: It's, it's text token with some characteristics, I would say.Ethan [00:42:54]: But musicSwyx [00:42:56]: I think the speech guys would disagree with this.Swyx [00:42:57]: Like disfluencies and then,Vibhu [00:43:00]: There's tones you can get angry.Ethan [00:43:01]: Well, I say largely.Ethan [00:43:03]: the mu- but the music is completely different. It's, it's very continuous, and you cannot model them like discrete tokens in language models. this is like the hard part for models is, not to mention we have to align text, video, and audio together.Ethan [00:43:26]: SoVibhu [00:43:26]: How?Ethan [00:43:28]: So significant-- some significant challenges are like-- So first, like we talk about as the VLMs, they cannot understand most of them cannot understand audio.Ethan [00:43:39]: So you have to have some way to do the synthetic data generation for audio. You have to caption the model, and that involve, that involve synthetic data and human data effort a lot. And not just surprisingly, most of the LLMs are very bad at recognizing, like the beat, tone, and the details of the of music. They can, they can give some general prediction of which song is this, but it's very hard to describe the details of the music. like we mentioned in image generation, like you have to describe image as detailed as possible so that someone blind can reconstruct that. So here is like someoneVibhu [00:44:32]: DeafEthan [00:44:32]: someone deaf can reconstruct how the music sounds like without actually listening to it. Maybe you can think of it need to have the-- or they call the script.Vibhu [00:44:49]: Subtitles, yeah.Ethan [00:44:49]: You gotta have all the details of the music, and the dialogue.Vibhu [00:44:55]: So is the challenge there typically stuff like music and audio, or is it just Like is there a baseline? Okay, there's enough data where we can understand, narration, conversation, but there's nuances in audio that's where you hit all the data issues or is it just from stage zero, you just do it all right?Ethan [00:45:15]: So one important thing is like the alignment. So the model, the model has to know like the video and audio, the, uh-- it has to have a time-based alignment, like at which time step the video and the audio token correspond to each other. But we actually don't have this kind of alignment for most of the other modalities. If you think about like text and image, text and video, they are loosely aligned. So you can, you can have a description of what's going on in the video, but you don't have to exactly, You typically don't have exact description, oh, at, time step one second like what happened?Vibhu [00:46:02]: It's veryEthan [00:46:03]: At time step two second what happenedVibhu [00:46:03]: coarse. Yeah.Swyx [00:46:05]: So what was the ideal time step? You have to oblate it, and then it's like four seconds or something.Ethan [00:46:09]: So that comes down to how you design the model to, for the model to be aware of as a time, as a time modality. So the model is like a time aware. And that's something pretty unique if you think about LLMs. So if you ask LLM to complete a task, say they, uh-- you ask them and they will say, “Oh, this task will probably take twelve hours to complete,” and they come back in one hour. Say “I've already spent two days on this and I've exhausted everything.”Ethan [00:46:47]: So the LLMs them-themselves, they don't have a sense of time there.Vibhu [00:46:53]: I actually don't think that's just them not having a sense of time. I think it's somewhat based, right?Vibhu [00:46:58]: Like you tell someone, “Okay, go work on this feature. Go implement this,” there's a general understanding you would have of how long that would take without LLMs working at LLM speed, right? So you think back like two years ago, if I tell you to like build me like a new front end for latent space, have a search bar, have all this, you'll estimate that it'll take a few days, right?Vibhu [00:47:19]: So you tell an LLM, “Go build this.” It'll take me a few days. But I think it's somewhat grounded as opposed to them not having the best-- Not saying that they have a great understanding, but I think that example is like you can see where it comes from, right? You're trained on all over the text.Swyx [00:47:35]: They're, they're trying to estimate what a human would say.Vibhu [00:47:37]: because that's what the, that's what the data kind of represents. It's not themEthan [00:47:41]: It came from the corpus on the internet. People have a estimate of how much time.Vibhu [00:47:45]: And not even just in direct like training samples, right? Just your world understanding of tokens of how long stuff takes, right? Go read a book. It'll take you a while, right?Vibhu [00:47:56]: Even if you do nothing but read a book, it takes a few days. So yeah, LLM, I read it took me a few hours.Vibhu [00:48:01]: It'll take me a few hours to go through this research. But this is a tangent.Swyx [00:48:05]: Somewhat, yeah.Swyx [00:48:06]: This is a train of thought I haven't really expressed until now is, which is basically like a full world model must also be recursive, meaning that the participant in the world model must also be aware that they have a world model. which is like this whole recursive thing down the, down the line. but yes, and that the world model can be wrong and that they need to update it and blah. Yeah. We've, argued this on the, newsletter as well, that there needs to be sort of recursive or adversarial world models.World Models: Real-Time, Long-Horizon, Interactive VideoVibhu [00:48:34]: just, to ask, how do you define world model?Swyx [00:48:38]: Oh, yeah, let's go there.Ethan [00:48:40]: SoVibhu [00:48:40]: So just for context, we talked about, video generation, and then there's a-- if you say there's a distinction between world models, what's your, what's your definition? How do you see the two?Ethan [00:48:53]: So disclaimer, I'm not going to debate, what is world model. Yeah. there are many definitions, so I'll just talk about my definition. Since I came from the multi-model, multi-model domain, so mainly talking from video. So world model is like real-time interactive long horizon videos. So there are three parts. so we-- let's talk about them one by one. So the so interaction, so we just, we just look at Facebook and neural computer. So the interaction part of it, so you, world model can allow you to interact with them through keyboard, mouse, and maybe also voice. So these all is-- all is a modality. You can, you can interact with the model, and the model should respond reasonably. Second part is real time. So once you, once, say, you move your mouse, if, say, the world model generate a game, how fast can the game respond? So if you're like professional CS: GO players- -my say, oh, you have to respond- He's beginner within sub ten milliseconds or- Yeah even less. So that's not most of the- No, sixty FPS. Let's go. Oh, three hundred FPS. Oh, five hundred FPS. Wait. okay, yeah. I didn't do the math, but yeah, okay. Uh- Yeah, three hundred FPS, that's a three millisecond. So you have to respond- Oh, s**t. Okay. YeahEthan [00:50:29]: within a millisecond. Most of the video models cannot do that. Yeah. And, but if you, say, if you have a video model that is, say, like a digital human, the response time might be more generous. Maybe typically, for real-time voice interaction, it's like two hundred millisecond. So that's, that's much more generous. But even two hundred millisecond is pretty, it is pretty tricky, ‘cause remember we mentionedEthan [00:51:01]: you have this, temporal compression coming from the VAE. So if you, if you don't compress the temporal dimension, your sequence length is going to explode. So if you want to have this real-time, real-timeness in your model, you have to do is one context problem. And the third part is long horizon, ‘cause we-- if you're not going to just play with, video games just, a few seconds, most video models only a few seconds. We're going to play with minutes, hours. The model have to be able to generate long-form content.Ethan [00:51:42]: So putting these three together, it's, real-time, long horizon interactive videos. I think the final state will be, for example, like a video, a video version of Playbook, where you can, you can interact with, a neural computer. You move your mouse, and you click on the generative interface, and it will reply to you through pixels- generating in real time. But getting there, it's, it's a very long way to get there. So one of the first step, at Grok Imagine, where I led a small world model team there, was to build video extension. So, video extension- it's the first step of interactivity. Yeah. It's, it's the first step. Yeah. So it's the first step- You have it here, video editing, yeah. Yeah. Yeah. So the first step is because, this unlocks long horizon videos. Typically, for most of the video generation models, you give it a prompt or an image as an initial frame. You generate video, that's it. That's just, one time, done. And some creators would try to, use the last frame as a first frame for the second video. It can-- sometimes it works, but if you do it a few times, it says the quality would decrease. And- It doesn't have that context- Yeah over the full video, so the temporal- Yeah, exactly. Yeah, ‘cause you only gave it the last frame, of course, right? Yeah. Exactly. And- it's actually a pretty fun hack. if you've seen like- Oh, no, he's saying something better. Yeah. And for example, like Vue, I remember Vue 3 has like a second context of the last video. It is slightly better than using the last frame, but it has the same problem-- similar problem that it, the quality would decrease. if you extend a few times to, one minute, the video quality would look much worse than the first video. Second, another problem is that the model doesn't have long-range knowledge of, what's happening before. Say, if they generate some dialogue, some, two people speaking, and their voice might change, over some time, especially if the second conditioning, it does not cover the previous context. So these are the core challenges. So the Grok Imagine video extension, it has historical context of all of the previous generated videos. It can, It has, it has the context of, who is speaking and what objects have appeared and everything, having that to generate the next video. So if we naively do this, you can imagine, just, put all of the previous history video tokens into the context. The context lens will easily explode. Especially for video models, that can be like a few, a few million context, I would imagine- context lens. Yes.Yeah.Swyx [00:54:58]: Let's run with that.Ethan [00:54:59]: for example, like in Cosmos, I think just five seconds of video is like a fifty K or sixty K number of tokens. So like if you do, if you do fifty second, that's a five hundred K tokens. If you do longer than that, easily explode. This long horizon, problem was the first step we're trying to solve world model. It turns out people, yeah, people love video extension. Like a lot, a lot of the creators love using video extension to create longer form videos. This is the part I liked that you have a, you have an intermediate step toward the final goal instead of just a straight shot to the final version very much.Swyx [00:55:48]: But I can see you have a strong vision of where we want to end up.Long Context, Redundancy, and Efficient Interactive VideoVibhu [00:55:51]: Does it seem like it's an efficiency issue? okay, we're at a few million tokens context,. If you draw the parallel to language models, we had very short context, two thousand, eight thousand, then, you scale it up one million, ten million. sure, there's effective context, but at the end of the day, it's just what's it worth? sure, there's a whole training data side. In video, it might be slightly easier ‘cause we have a hundred million token video, right? Just take a movie with the full context there. Like is this efficiency from an inference standpoint that like it's expensive, but we know how to solve it? Or like why is this not the approach? So like my broader point was on your second point of world models, you say it needs to be interactive and live, right? You should be able to play a game and see the interaction live. So one thing I see with research is a lot of what you actually serve is different than what you build, right? So we talked about distillation. You train big model, you distill it, you do quantization, speculative decoding. We do all this stuff to serve it efficiently. Should we not just have a solution, like a world model that can interact well, do inference optimization, serve it, distill it secondary, so make it real time after you solve it? So like a-- another parallel is say, continual learning, right? What we need is someone to solve it and show it works inefficiently. Give it a few years, people will make it efficient. Same thing with regular attention, right? It worked. Over a few years, people have different forms of attention, and we've scaled it to be efficient at log context,? So kind of two things there, right? One is it seems like it works. You've scaled it. Can we not just scale it a lot more efficiently over time? Do we need a separate approach if this works? And same thing with interaction, right? if we can get it done, like if we can solve some way that it works, we can solve making it more efficient from an inference standpoint later.Ethan [00:57:53]: that's actually a very good point. So in videos, there's actually a lot of redundancies. So we solve a lot of the pixel redundancy from VE, but there's more redundancy in long range and long horizon videos. Say, if a character appear in the first clip and then it disappeared, it only reappear at the end of the video, you probably don't need the-- the context, like in the middle of the generation. So you only need that character, where you need. So that's why, I helped build another feature. It's a reference video.Vibhu [00:58:36]: Is it here?Swyx [00:58:36]: is it the same model release or different one?Ethan [00:58:39]: It's a different one.Ethan [00:58:41]: You probably need to search onSwyx [00:58:43]: I'll find itEthan [00:58:43]: X reference to video.Ethan [00:58:46]: So reference video allow you to like upload up to seven images as condition and generate the video. Say, if like I want-- it can, it can be characters or objects or even scenes. Say like I want, I want condition on, Sean's selfie and holding a bladeSwyx [00:59:07]: We have a dogEthan [00:59:08]: or whatever.Swyx [00:59:08]: We put the dog in the thing.Ethan [00:59:09]: you can put them there and the video models will generate the video from and copies the context over. So that can solve a lot of the problems there, like the long context problem. It doesn't need to have a very long context, but it's-- I feel like it's an intermediate solution. The modelSwyx [00:59:29]: It's cheating.Ethan [00:59:30]: the model should be able to like selectively know, where should I draw the references. So say if I want to generate a movie, I generate it autoregressive, like a ten second at a time or something. And now this character appear, I can look back to where it first appear and, bring that back. Yeah, this one, I put the references. Yeah, that's, Optimus, Einstein myself, Annie.Vibhu [01:00:02]: Oddly enough, I used Grok Search to find it, and it pulled your LinkedIn post. But yeah we found it.Ethan [01:00:08]: Interesting.Vibhu [01:00:10]: ButxAI's Underrated Work, Culture, and WatermarkingSwyx [01:00:11]: this is a problem. This is not your fault, but like XAI doesn't communicate all this work that you do very well because they just have the model release and then that's it. But actually, these details are very good.Swyx [01:00:22]: As far as I understand, everything you just described is state-art, like no one else has done it.Vibhu [01:00:30]: A lot of-- yeah, I have a lot moreSwyx [01:00:32]: And then, and then you just put this blog post with the cookies. I'm this is not enough,?Swyx [01:00:37]: but I, obviously this is like the high level numbers that people want to know. But no, okay, soVibhu [01:00:42]: And I wonder, like part of that is also some labs don't share research into what happens. And ifSwyx [01:00:50]: No, but this is literally bragging about how good they are, right?Swyx [01:00:54]: Like, why would you not say that you are capable of extending with full context? this is not a secret sauce. This is like we did the work. yeah, I don't know.Ethan [01:01:02]: different labs have slightly different communication styles.Swyx [01:01:07]: Anyway, if anyone from XAI is listening we are always happy to help you tell your story. Yeah, okay, so you did references, and I think, I think kind of the point you're, you're making is it is sort of like a kludge, right? this is-- you can do seven, but what about 100?Swyx [01:01:23]: Right? Then you need a completely different thing.Ethan [01:01:26]: So I think it's-- this is, a mechanism to, select the context from the history, and you might not put the entire history into the context. for example, there's a paper called Frame Pack, which haveEthan [01:01:41]: a heuristic that the latest history, the last one second, I put the entire history, and the history before that, I would, compress it and makes the video smaller. So they follow this pattern, this build overall pattern that the maximum sequence length is fixed. So the further you are from the current frame, you have a smaller image. So this is just a heuristic. I think it can be more automatic. The model is aware like which history part of it can be select. So this part of the research is actually being actively, worked on by a lot of people. It's also quite interesting. I feel this is actually, this part of long context is a little bit ahead of the LLM part.Ethan [01:02:31]: So for example, like in LLMs, if you-- so contexts keep growing. Let's say if you call tool and the tool call history is extremely long, that's still in context, and keep growing, keep growing. Even if you switch the topic to something else, the whole context was there. There are some agentic harnesses that help you to, say, prune the tool results and, prune Like when you, when you query a file, only show like the top 200 lines or something. Those were very heuristic-driven.Swyx [01:03:08]: For listeners, we did a write-up on the cloud code, leak where there are eight different kinds of pruning, including like you prune the tool results and all that. So you can, you can read up on that kind of thing.Ethan [01:03:17]: I think, one breakthrough in continual learning might be like a way to automatically, manage its own context.Swyx [01:03:27]: These are all heuristics, and they will be replaced by machine learning.Ethan [01:03:30]: InterestinglyVibhu [01:03:32]: TheEthan [01:03:32]: the same thing is being researched in both LLMs and video models.Vibhu [01:03:36]: The interesting thing is also like in the paper you showed, it's actually happening at the model level, right? Compared to like language models, sure, we have base attention, but we'll do our own compression, we'll do our own pruning, which is separate from model error.Vibhu [01:03:49]: Eventually, it all just boils in, hopefully.Swyx [01:03:52]: I think this is a form of like attention, but like also know sort of reasoning attention. I feel like that's different than normal attention.Swyx [01:04:03]: Does that, does that make sense?Ethan [01:04:04]: It's, it's different in the sense that attention, not to mention, set sparse attention aside,

Trump just signed his biggest crypto executive order yet, giving the Fed 120 days to evaluate opening payment rails to Coinbase, Kraken, Circle, Ripple, and Anchorage. The move could break the bank monopoly on Fed plumbing right as Bitcoin sits at $76K after $1B in ETF outflows and $660M in liquidations. Add Warren's attack on OCC crypto charters, JPMorgan saying Bitcoin ETFs are recovering 2X faster than Ethereum, Goldman dumping XRP and Solana for Hyperliquid, and the SEC readying tokenized stocks, and you have one of the most pivotal weeks of the cycle. Is this the bottom, or the setup for Bitcoin's next leg higher? Learn more about your ad choices. Visit megaphone.fm/adchoices

Sir Ben McDougal is a 2X author, 5X founder, podcaster, teacher, content creator, technologist, and traveler. Check out his latest book, BrewedFromWithin.com Welcome to the Conscious Millionaire Show - Become an Ultra-Performer. Now 3X week M / W / F   Are you an Entrepreneur, Founder, or CEO? Revenues $250K to $5M? Sign up for your Breakout Session...get custom steps to build a fast-growing, highly profitable business that makes an impact.     BOOK Your Breakout Session Now     Join Host JV Crum III, with 2 exits and over 75M revenues in his companies, he is the Ultra-Performer Advisor for Founders, Entrepreneurs and CEOs ready to achieve at your the top 1%.     SUBSCRIBE to Conscious Millionaire Show     Season 12 of the award-winning Conscious Millionaire Show. The World's #1 Ultra-Performance podcast. Millions of Listeners. 190 countries -- Inc Magazine "Top 13 Business Podcasts" with 12 seasons and 3,200+ episodes.

Sir Ben McDougal is a 2X author, 5X founder, podcaster, teacher, content creator, technologist, and traveler. Check out his latest book, BrewedFromWithin.com Welcome to the Conscious Millionaire Show - Become an Ultra-Performer. Now 3X week M / W / F   Are you an Entrepreneur, Founder, or CEO? Revenues $250K to $5M? Sign up for your Breakout Session...get custom steps to build a fast-growing, highly profitable business that makes an impact.     BOOK Your Breakout Session Now     Join Host JV Crum III, with 2 exits and over 75M revenues in his companies, he is the Ultra-Performer Advisor for Founders, Entrepreneurs and CEOs ready to achieve at your the top 1%.     SUBSCRIBE to Conscious Millionaire Show     Season 12 of the award-winning Conscious Millionaire Show. The World's #1 Ultra-Performance podcast. Millions of Listeners. 190 countries -- Inc Magazine "Top 13 Business Podcasts" with 12 seasons and 3,200+ episodes.

It's Microdose Monday on Fitness Stuff for Normal People which means Marianna and Tony are rapid fire breaking down multiple topics you guys have been asking about instead of spending the whole episode on just one. This week, they're diving into reverse dieting, why creatine might actually be one of the most underrated tools for fat loss, how coffee can impact your gut health, and the sneaky things nutrition labels hope you don't notice. If you've ever felt overwhelmed by conflicting health advice online, this episode is packed with quick, practical breakdowns that actually make sense. Shorter topics, real explanations, and plenty of little “wait… WHAT?” moments along the way.⁠⁠⁠⁠⁠⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps:(2:02) Coffee and Gut Health
(15:08) Reverse Dieting
(26:27) How to Accurately Read a Nutrition Label
(39:31) Creatine for Fat Loss

On this episode of Fitness Stuff for Normal People, Marianna and Tony break down the 5 biggest dieting mistakes that keep people stuck when trying to lose fat or build muscle. They talk through why making sudden, extreme changes usually backfires, how not tracking your food can leave you guessing, and why chasing fast results slows you down long term. They also get into the trap of focusing on small details that don't actually move the needle and how black and white thinking around food can hurt your consistency. If you've ever felt frustrated with your progress, this episode will help you simplify the process and focus on what actually works.⁠⁠⁠⁠⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(3:49) Mistake 1 Making sudden or massive changes
(9:02) Mistake 2 Not tracking food
(17:36) Mistake 3 Thinking faster not longer
(23:43) Mistake 4 Little details
(32:31) Mistake 5 Thinking black and white

Learning a new language as a busy executive sounds great in theory. In practice, the process for how to learn Spanish, for most people, is downloading five apps, getting overwhelmed, and quitting within a month. Sound familiar?In this episode, Omar Newman—Spanish educator, coach, and founder of Carolina Language Solutions—joins Julian Hayes II to break down exactly how busy operators and professionals can learn Spanish without the noise, the overwhelm, or the wasted time.Omar brings 34 years of experience, 22 of which were spent teaching in the school system, before building his own adult language coaching practice. He explains why most people are learning Spanish the wrong way, how to build real conversational fluency faster than you think, and why the biggest barrier isn't grammar, but instead, confidence.He also unpacks why Spanish sounds so fast to English speakers, how to use high-frequency phrases to build momentum fast, and the mindset shift that separates people who stall from those who actually speak. This is a practical, no-fluff conversation about one of the highest-leverage skills an executive or entrepreneur can develop, and a natural extension of what Executive Health is all about: optimizing your full performance, mind included.— Episode Chapter Big Ideas (timing may not be exact) —0:00 - Introduction and why 2026 is the right time to learn Spanish0:45 - Omar's origin story: from Fort Bragg to 34 years in the language3:00 - Why culture and music pull people into a language—not the language itself7:36 - The language noise problem: too many apps, too many voices10:26 - High-frequency phrases and the ICE method for building sentences fast17:41 - Repetition, reps, and treating language like a training program21:44 - Confidence as the real barrier and how to overcome it26:30 - Why Spanish sounds so fast to English speakers28:47 - Syllable-based vs. stress-timed language: the real reason for the speed gap32:26 - The 2X speed hack: train your ear in English to hear Spanish better37:27 - Staying consistent: the SOTT plan and three activities a day44:15 - The four levels of fluency (and why the European framework overcomplicates it)49:02 - How conversational fluency holds up in real business settings56:42 - The easiest forms of Spanish to learn first1:01:43 - Learning two languages simultaneously: how Omar manages six1:06:01 - Three things to do right now if you're serious about Spanish1:11:09 - Where to connect with Omar— Key Quotes from Omar Newman — "It's not that they're speaking too fast. It's that you all listen slow.""You don't go to Baskin-Robbins and try every single ice cream they have. Stay in your lane.""You're already creating a false scenario of how it's going to be received when you use the language in public. Once you do it, that's it.""You can't start building the roof on the house when you don't have a foundation."— Connect With Omar Newman —Website: https://carolinalanguagesolutions.com/ YouTube: http://www.youtube.com/@SpeakSpanishNowInstagram: http://www.instagram.com/carolina_language_solutionsLinkedIn: https://www.linkedin.com/in/omar-newman-26877b4b/ — Connect with Julian and Executive Health —LinkedIn — https://www.linkedin.com/in/julianhayesii/X — https://x.com/thejulianhayesDon't let your biology become the bottleneck to the enterprise you're building. Book a private call —https://www.executivehealth.io/contactWebsite — https://www.executivehealth.io/***DISCLAIMER: The information shared is not meant to treat or diagnose any condition. This is for educational, informational, and entertainment purposes. The content here is not intended to replace your relationship with your doctor and/or medical practitioner. Consult your provider before making any decisions.

What if change didn't take years, but minutes?In this episode, Sagi Chekroun shares how Neuro-Linguistic Programming (NLP) can help you create rapid, lasting transformation by working directly with the unconscious mind—the true driver of your habits, emotions, and results.Most people try to change through willpower, discipline, or information. But real change doesn't happen at the surface—it happens beneath it. When you understand how your mind actually stores patterns, you can begin to shift them faster than you thought possible.We break down why change often feels difficult, and challenge the belief that transformation has to be slow. You've already created habits, fears, and emotional patterns instantly. The same mechanism can be used to rewire them.One of the most powerful ideas discussed is secondary gain—the hidden benefit behind every habit. Whether it's stress relief, comfort, or a sense of control, your mind holds onto patterns because they serve you in some way. If you try to eliminate a habit without replacing what it gives you, it often comes back in another form. Real change comes from upgrading the benefit, not fighting the behavior.We also explore the mindset shift of “10X is easier than 2X”—why aiming for bigger transformation can actually be simpler than incremental change, because it forces you to think differently and step into a new identity.From there, we get practical.You'll learn simple NLP tools you can start using immediately:How to “scramble” negative memories by changing how they're stored (visual, auditory, and physical components)How to use dissociation to safely process intense experiences without reliving themHow to create a mental “control panel” to interact with your inner worldHow to build anchors—physical triggers that instantly shift your emotional stateHe also guides me through a powerful 3-minute future pacing exercise. This practice helps you step into your future self—seeing through their eyes, feeling their emotions, and allowing your mind to begin wiring in the path to get there. Instead of chasing a goal, you begin to embody it.A key theme throughout the episode is that less is more. You don't need long, complicated routines. Short, consistent practices—just a few minutes a day—can create real, lasting change when done correctly.This episode is for anyone who feels stuck, wants to break patterns, or is looking for a more aligned, effective way to grow—mentally, emotionally, and spiritually.Resources mentioned include Richard Bandler, Paul McKenna, and Jose Silva, along with additional NLP tools and teachings.Take 3 minutes today to try it yourself:Visualize your life one year from now. Step into that version of you. Feel it fully.Then ask: Who do I need to become to live this?Because real change doesn't come from forcing yourself forward. It comes from becoming someone new.

Reserve your spot for the next Passive Income Flywheel Masterclass: 

In this episode of Fitness Stuff for Normal People, Marianna and Tony break down what gut health actually means and why it impacts way more than just digestion. From metabolism to mood, your gut is tied into everything, and this episode focuses on the habits that truly make a difference. They walk through the science behind fiber and prebiotics, hydration, stress and the brain gut connection, daily movement, and how your overall eating patterns shape your microbiome. They also get into the real effects of antibiotics, probiotics, alcohol, under eating, and sleep so you can understand what is helping your gut and what might be holding you back. This is a practical, no fluff look at gut health built around realistic habits you can actually stick to.⁠⁠⁠⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps3:42) What is Gut Health?
(8:24) Fiber / Prebiotics
(16:11) Drinking enough water
(19:32) Limit stress / Brain and Gut connection
(25:26) Daily movement
(27:28) Limited processed foods / eating patterns
(34:17) Antibiotics
(38:23) Probiotics
(42:57) Alcohol
(46:28) Under eating
(58:56) Sleep

You're showing up to the gym, getting your steps in, being mindful with your food, and still not seeing the progress you expect. In this episode of Fitness Stuff for Normal People, Marianna and Tony break down the five most common reasons fat loss stalls, even when you feel like you're doing everything right. They cover where calories tend to slip through the cracks, how people misjudge what they're actually burning, why daily movement matters more than you think, and how stress, sleep, and overall lifestyle can quietly impact results. They also talk through the role of expectations and consistency, and why progress often takes longer than people anticipate. If you've been feeling stuck or unsure what to adjust, this episode will help you refocus on what actually moves the needle.⁠⁠⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(6:21) Underestimating calories you're consuming(13:10) Overestimating how many calories burned(19:02) Not moving enough(26:41) Rest of your life sucks(39:32) Tie..Unrealistic expectations / Not consistent

SEASON: 6 EPISODE: 23Episode Overview:Welcome back to Becoming Preferred, the podcast where we interview the brightest minds in business today to help you level up your game and become the best version of you.If you are an entrepreneur or a marketing leader, you've repeatedly heard the mantra 'do more with less'. But our guest today says that's the wrong goal. She believes we shouldn't just be doing more; we should be redesigning the work itself.Lisa Cole is the Chief Product, Marketing, and AI Officer at 2X, where she heads their AI Center of Excellence. With over 20 years of experience and three major B2B transformations under her belt, Lisa has become the 'secret weapon' for CMOs looking to bridge the gap between human creativity and machine efficiency. She is the author of Brand Gravity, The Revenue RAMP, and her highly anticipated new book, The Limitless CMO.Today, we're going to talk about how to move from being a 'cost center' to a 'revenue engine,' the reality of AI adoption , and how to build a brand that has enough 'gravity' to pull customers toward you without you having to chase them. Join me for my conversation with Lisa Cole.Guest Bio: Lisa Cole is Chief Product, Marketing & AI Officer and Head of the AI Center of Excellence at 2X. For more than 20 years, Lisa has helped B2B marketing leaders achieve more with less by redesigning how marketing work gets done through strategic outsourcing, AI adoption, and performance management.Lisa has led three major B2B marketing transformations, including award-recognized work, and she's been a trusted advisor to CMOs for nearly 16 years. She's also the author of Brand Gravity, The Revenue RAMP, and her newest book, The Limitless CMO.Resource Links:Website: www.lisacole.aiCompany Website: https://2x.marketing/Insight Gold Timestamps:03:31 I shifted and pivoted into marketing06:13 I remember when I started the very first question a company would ask is, should I have a website?08:36 The underlying principle there is still true, you actually have to ask them11:07 By the time they've actually made that decision to engage with your brand directly...12:43 Marketing is the one that makes sure that you are both findable, understood, and chosen16:21 Could marketing do a better job of being proactive with that, with the sales department?20:24 The premise behind brand gravity is...23:33 No better time to be a founder or entrepreneur, if you're comfortable with just putting your point of views, and your secret sauce out in the world25:52 Your superpower is going to be wherever is uniquely human28:02 Are there some must haves that small businesses should have?31:53 The most common piece of feedback I get is...32:22 My measure of success here is engagement34:27 You have to be okay with change36:10 And you argue for redesigning how work gets done instead of just working harder37:37 My favorite question to ask is, how long does it take for you to get a campaign from an idea into market?40:02 You head the AI Center of Excellence at 2X42:04 The new book's called The Limitless CMOConnect Socially:LinkedIn: www.linkedin.com/in/lisacole01Twitter (X): https://x.com/CMOInnovatorCompany: https://www.youtube.com/@2xmarketingTikTok: https://www.tiktok.com/@thelimitlesscmoEmail: lisa.cole@2X.marketingSponsors: Rainmaker LeadGen Platform Demo: https://calendar.summit-learning.com/widget/booking/JKItVP7WErmCBjU2cCIxRainmaker Digital Solutions: https://www.rainmakerdigitalsolutions.com/

Most people think they need more discipline or motivation or that they just need to lock in a little harder, and sometimes that is true, but a lot of the time the people actually making progress are not better than you, they have just built an environment where the right decisions happen more automatically. In this episode of Fitness Stuff for Normal People, Marianna and Tony are breaking down the tools they actually use to make losing fat, gaining muscle, and sticking to your diet easier to execute day to day, from gym add ons like straps, belts, and apps, to simple kitchen tools that take the guesswork out of eating, to a few things outside of both that help you stay consistent without having to rely on willpower every single day.⁠⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(4:43) Lifting Straps or Hooks
(9:37) Wrist Wraps
(10:56) Liquid Chalk
(14:42) Lifting Belt
(19:38) Lifting Apps
(22:01) AI tools(27:05) Food Scale
(29:00) Meal Prep Containers
(31:00) Veggie Chopper
(32:00) Air Fryer and Misc Kitchen Tools
(38:00) Grocery List
(41:00) Cronometer(42:46) Scale
(48:24) Our Free Calculators
(49:47) Sunrise Alarm Clock
(52:04) Sleep tracker(54:08) Books

I don't know about you, but I find the cereal aisle to be a pretty wondrous place.  We're all probably pretty used to it, but just imagine the cereal aisle from the perspective of a person who has never seen it before. It's a lot. I mean, how do you pick a cereal when you don't know what each of them has to offer.  Of course, this is where marketing comes in.  Marketing is that thing that tries to connect us with products, services, organizations, and every other matter of thing that is trying to get our attention. In the Attention Economy, that's a tall ask. The Attention Economy refers to how human attention is a scarce resource and can be traded for profit. The key point here is it is a ‘scarce resource,' and one that is only getting more scarce. As an educator, I know this very well. We only have so much attention to give, and there is more and more that is competing for it.  Marketers have a much harder job, but luckily for them they have more tools at their disposal to help. The trick is how to best use them. Today's guest on Experience by Design is Lisa Cole, Chief Marketing, Product, and AI Officer (or CMPAIO) of the company 2X. She also is the author of the just released book The Limitless CMO). She also is the author of the books “Brand Gravity” and “The Revenue Ramp.”  We talk about how growing up she was the kid that always tried to take things apart to see how they worked. That translates very well to exploring how AI can be leveraged for marketing today, digging into it and seeing how it works in order to capitalize on its abilities in the Attention Economy.  We discuss how we need to spend more time learning than just doing, finding ways to do things differently than before if we want to stand out from the noise. Lisa highlights the importance of reimagining workflows and tasks before applying AI to optimize outcomes and prevent inaccuracies. Afterall, the product produced through the use of a tool is always dependent on the skill of the person using it. Lisa's new book shows how marketing leaders can overcome constraints through three key levers: running marketing like a business, leveraging outsource models, and adopting AI. We also discuss the relationship between sales and marketing, and how many companies continue to make it harder to buy when they finally get your attention.  Finally, Lisa shares how she used an AI-based tool to interview her on long drives to help her generate ideas for her book.  In this new era of attention competition, there is a lot to learn to make people aware of our presence, and this conversation is a great place to start no matter what you are marketing. Even cereal. Lisa Cole: https://www.linkedin.com/in/lisacole01/ 2X Marketing: https://2x.marketing/ “The Limitless CMO”: https://www.amazon.com/Limitless-CMO-Transform-Order-Taking-Market-Making-ebook/dp/B0GSCXR8HM

In today's episode of Fitness Stuff for Normal People, Marianna and Tony get real about something a lot of people struggle with but don't always say out loud… having a bad relationship with food. When you start taking your fitness seriously, tracking, and trying to make progress, it can quietly shift the way you think about eating in ways that aren't always healthy. This episode breaks down how that slippery slope happens and gives you five evidence-based ways to rebuild a healthier, more sustainable relationship with food. If you've ever felt stuck between “doing it right” and actually enjoying your life, this one's for you.⁠⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(2:27) Current relationship with food(11:57) How to improve your relationship to food(12:03) Pay attention to hunger cues(22:37) Stop labeling as good or bad(33:03) Eat mindfully(40:53) Ditch the trendy diets and food rules(45:15) Stop letting other people's opinions influence your food choices

From High Revenue to High Profit: The Missing Piece in Your Business with Chris Hallberg, EOS   Find Rocky Lalvani @ www.ProfitComesFirst.com  or email him at rocky@profitcomesfirst.com  Make more, work less video: https://youtu.be/    Hire a Green Beret: Why Veterans Transform Your Business    In this episode, Rocky Lalvani sits down with Chris Hallberg, ranked #9 on Inc. Magazine's Top 50 Leadership & Management Experts, to discuss why hiring the right people and implementing disciplined systems are the real keys to building a profitable business. Chris shares insights from his military background, his veteran-powered recruiting company Business Sergeant, and his work implementing the Entrepreneurial Operating System (EOS) with hundreds of companies. Learn why Green Berets might be your secret weapon, how to stop bleeding money through bad hiring decisions, and why your profit problem might not be a revenue problem at all.    Learning Insights  The true cost of bad hiring: A single bad hire in a $100,000 role costs approximately $500,000 when accounting for turnover and lost productivity. A-players cost only 1.2X to 1.6X more but deliver 2 to 10 times the value.  Veterans are exceptionally rare and valuable: Only half of 1% of the US population has special operations training. They don't cost more to hire than regular candidates but deliver exponentially more value through proven leadership under pressure.  High revenue does not equal high profit: The biggest pattern Chris sees is companies saying yes to every opportunity. Without a strong number two person (COO/integrator) to say no and protect margins, you get high sales but low profit.  Your yes person needs a no person: Visionary CEOs naturally seek opportunities. They need a strong integrator to say no and protect profit margins. Without this balance, money disappears and profit suffers.  Accountability is natural with the right people: When you hire aligned, quality people who share your values, accountability happens without friction. If you can't hold someone accountable, you have the wrong person in that seat.  Use math, not gut feeling, to make decisions: Create a go/no-go matrix based on realistic data. Input assumptions about revenue, time, and resources. Let the numbers tell you yes or no instead of relying on passion or intuition.  Discipline beats opportunity every single time: The road to business failure is paved with companies that couldn't decide what to say no to. Clear, disciplined decisions about strategy and fit matter more than saying yes to everything.    The Big Takeaway  The difference between businesses that struggle and businesses that thrive isn't complicated. It's not about working harder, better marketing, or a superior product. It's about two things: the right people in the right seats, and the discipline to say no to opportunities that don't fit your strategy and profit model. Most visionary founders and CEOs are wired to say yes. They're opportunity seekers. That's their strength. But without a strong integrator, COO, or number two person who protects profit margins by saying no, companies end up with high revenue and low profit. They're exhausted, understaffed, and serving too many customers at too thin a margin. Additionally, most business owners are flying blind when it comes to hiring and decision making. They rely on gut feeling instead of math. Veterans, particularly those from special operations backgrounds, bring a rare combination of perseverance, problem solving, accountability, and calm under pressure that most candidates can't match. They've been selected and tested in environments where failure isn't an option. They understand what real adversity looks like, which makes business challenges feel manageable by comparison. The math is simple: invest more upfront in the right person, hold them accountable, create systems for evaluation and improvement, and say no to opportunities that don't fit. Do this, and your business transforms.    Conclusion  Building a profitable, scalable business requires more than good ideas and hard work. It requires the right people in the right seats, clear systems for making decisions, and the discipline to say no. Chris Hallberg's work with hundreds of leadership teams and his experience as a veteran demonstrate that these principles work regardless of industry or company size. Whether you hire a Green Beret through Business Sergeant or simply apply the framework Chris and Rocky outlined, the message is the same: your people and your discipline are what create profit. Everything else is a distraction.    About Chris Hallberg  Chris Hallberg—known as the "Business Sergeant"—is a top-ranked leadership expert, military veteran, and serial entrepreneur who transforms good companies into great ones, fast. Ranked #9 on Inc. Magazine's Top 50 Leadership & Management Experts—ahead of Simon Sinek—Chris blends battlefield-tested leadership with the Entrepreneurial Operating System (EOS) to deliver proven results.  He scaled and sold a startup during the Great Recession at an 8× multiple, built royalty-generating sales systems, and became Colorado's first EOS Implementer, guiding 100+ teams to achieve 90%+ employee engagement rates and 100+ Best Places to Work awards. Today, he co-builds a $5M AI-driven EOS platform while coaching billion-dollar contractors, national chains, and franchises with a remarkable 85% success rate.  With his no-nonsense, high-energy style, Chris simplifies strategy, strengthens culture, and shows leaders how to drive 30%+ EBIT on predictable systems—making him a powerhouse guest for any podcast.    Links  Website: https://goexpand.com/   LinkedIn: https://www.linkedin.com/in/chris-hallberg-01516315/  Facebook: https://www.facebook.com/people/GoExpand/61577326657347/#  Instagram: https://www.instagram.com/goexpandplatform?igsh=MXV5N2I1Mml0MXF4aw%3D%3D  YouTube: https://www.youtube.com/@GoExpand    Profit Blueprint Calculator I Profit Comes First https://lp.profitcomesfirst.com/profitblueprintcalc-page    Watch the full episode on YouTube: https://www.youtube.com/@profitanswerman  Sign up to be notified when the next cohort of the Profit First Experience Course is available!  Free Copy of the Profit Blueprint Book: https://lp.profitcomesfirst.com/landing-page-page   Monthly Newsletter signup: https://lp.profitcomesfirst.com/newsletter-signup  Relay Bank (affiliate link): https://relayfi.com/?referralcode=profitcomesfirst  Profit Answer Man Facebook group: https://www.facebook.com/groups/profitanswerman/  My podcast about living a richer more meaningful life: http://richersoul.com/  Music provided by Junan from Junan Podcast  Any financial advice is for educational purposes only and you should consult with an expert for your specific needs. 

In today's episode of Fitness Stuff for Normal People, Marianna and Tony are back with Micro-Dose Monday, breaking down a mix of the most interesting topics you've been asking about lately with quick, practical, evidence-based takes. They get into the science behind microneedling and whether it actually delivers for skin and hair health, unpack the $85 million David Protein Bar situation, and look at what really happens when you take a much higher 25g dose of creatine compared to the standard 5g. They also cover a new study that challenges the idea that cutting out sweet foods reduces cravings or improves health outcomes. If you've been curious about any of these, this episode gives you a clear, no-nonsense breakdown of what actually matters.⁠⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(2:33) Science of Microneedling(25:25) David Protein Bars(36:58) 25g of Creatine(43:17) Cutting Sweet Foods Study

In this episode of Fitness Stuff for Normal People, Marianna and Tony take an objective look at alcohol and its impact on overall health and performance. Rather than taking an all-or-nothing approach, they break down how alcohol affects key areas including fat loss, muscle building, gut health, hormones, sleep, and mental health. The goal is to provide clear, practical insight so you can better understand the tradeoffs and make informed decisions that align with your goals.⁠All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(2:38) Levels of drinking(5:44) Weight loss & Building muscle(14:22) Gut health
(21:07) Hormones
(27:42) Mental health
(37:54) Sleep

Spring hit, the clocks changed, and suddenly summer doesn't feel far away. In this episode of Fitness Stuff for Normal People, Marianna and Tony break down exactly what to do (and what not to do) when you realize beach season is basically around the corner and you need results fast. No crash diets, no influencer nonsense. Just the real steps that actually move the needle. If you're two months out and ready to lock in, this episode is your game plan.All Free Calculators (Calorie, Protein, etc.)⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Sign up for Fitness Stuff PREMIUM here!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ALL of our complete 12-week training programsBonus episodes every FridayJust $5 /month⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Legion Athletics⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠BOGO 50% off for your first order + 2X points on every order after thatuse code “FSPOD” at checkoutTimestamps(7:45) What not to do
(19:29) Water retention/bloating(19:45) Step 1 - Calories (25:15) Step 2 - Maximize Steps(27:11) Step 3 - Water(28:24) Step 4 - Water retention
(30:05) Step 5 - Strength Training

IN THIS EPISODE: Every organization begins by serving customers. Not all of them get it right — or keep it that way. So where does the customer  relationship break down? In this episode, Denise Silber speaks with HBS MBA Caroline Evans de Gantes, a transformation leader who has spent more than two decades helping companies design strategy, culture, and operations around customer love as an organizing principle. Caroline recalls arriving at Harvard Business School in 2008 just as the Lehman Brothers collapse unfolded — an experience that shaped her leadership perspective: progress comes from experimentation, learning quickly, and working closely with frontline teams. Drawing on transformations across industries known for difficult customer relationships, she shares what it takes to reconnect organizations with the people they serve. From mobile phone insurance and broadband installation to the transformation of leading real-estate marketplace SeLoger within the Aviv Group, the conversation explores how leadership culture, incentives, and technology can be realigned around customers — and how doing so improves business performance. Throughout the discussion, Caroline shows that making customer love an organizing principle starts on Day 1. GUEST BIO: Caroline Evans de Gantès, a 2010 MBA graduate of the Harvard Business School, has over 20 years experience transforming offline industries by placing the client front and center. From insurance to telecommunications  to real estate, Caroline has deployed this approach to transform culture, product, and strategy, and turn profits around. Most recently, Caroline led a transformation at SeLoger (a 30 year old French real estate marketplace) to 2X the growth rate and a great NPS (Net Promoter Score), through redesigning culture, organisation, and business model. She merged SeLoger, and other real estate companies in France, launched new services to take back market share from rivals and championed a customer-centric culture in France, Belgium, and Germany. The resulting business, Aviv, is the leading European real estate marketplace with 50M monthly active users.  Caroline is originally from the State of Mississippi and has fashioned a career as the bridge between people and technology and the US and Europe. 

There is a storm coming with the challenges of navigating the TRUSTEE CRISIS. It is one of the biggest blind spots in the “GREAT WEALTH TRANSFER” and will be the source of mountains of litigation for the unwary, https://youtu.be/hwQev88A03M Summary In this conversation, Frazer Rice and Jennifer Zelvin McCloskey discuss the current crisis in trusteeship, highlighting the shortage of qualified trustees amidst a significant wealth transfer. They explore the importance of modern trust planning, the challenges faced by individual trustees, and the need for better education and training in the field. The discussion also covers the emotional and interpersonal aspects of trusteeship, the functions and responsibilities of trustees, and the necessity of managing risk effectively. They emphasize the importance of building a pipeline for future trustees and improving the perception of the profession, while also identifying opportunities within the trust industry. https://open.spotify.com/episode/4qpkrVdaUa2AfDxgl7j3yN?si=XVgG3jE_Qpqq2JTqi8XLXQ Editing and post-production work for this episode was provided by The Podcast Consultant (⁠https://thepodcastconsultant.com⁠) Takeaways The coming crisis in trusteeship is already here. There is a significant shortage of qualified trustees. Trusteeship requires strong interpersonal skills and emotional intelligence. Managing risk is a fundamental aspect of trusteeship. Trustees critically need education and training. The role of a trustee is evolving with increasing complexity. Beneficiaries need to understand their rights and the trustee’s role. Custodial responsibilities are essential for asset protection. There are many opportunities for growth in the trust industry. Trust law and investment management are distinct fields. This Episode is for . . . Anyone that has an estate plan with a trust in it and doesn't know what a trustee does Any advisor who works w/ multi-generational situations (that’s everybody in wealth management) Any RIA looking to sell Financial types worried about compliance world Fiduciary litigators Chapters of “THE TRUSTEE CRISIS: Navigating the Challenges” 00:00 The Coming Crisis in Trusteeship 02:06 Importance of Modern Trust Planning 04:11 Challenges with Individual Trustees 08:03 The Dwindling Pool of Qualified Trustees 10:06 Functions and Responsibilities of a Trustee 12:20 The Emotional and Interpersonal Aspects of Trusteeship 16:05 Managing Risk in Trusteeship 19:07 Building a Pipeline for Future Trustees 22:10 The Role of Education in Trusteeship 25:07 Improving the Perception of Trusteeship 28:19 The Need for Better Trust Education 30:39 Bifurcation of Trustee Functions 33:26 Distribution Functions and Beneficiary Relations 36:52 Custodial Responsibilities in Trusteeship 40:19 Consequences of Poor Asset Management 46:41 Curriculum for Trustee Education 52:13 Opportunities in the Trust Industry Transcript of “THE TRUSTEE CRISIS: Navigating the Challenges” Frazer Rice (00:01.068)Welcome aboard, Jennifer. Jennifer Zelvin McCloskey (00:02.723)Thanks Frazer, how are you today? Frazer Rice (00:04.782)I am doing great. We’re going to dive into a topic that is near and dear to both of our hearts. And that is what I’m describing as the coming crisis in trusteeship, but I think it’s already here. Which is the concept of qualified trustees being in short supply, right in the face of a gigantic wealth transfer. And first of all, before we get into that, just describe what you do on a day to day basis first. Jennifer Zelvin McCloskey (00:33.445)Sure, I actually wear a bunch of hats. Day to day, right now, I’m a full-time practicing trust and estate attorney. I’m also an individual trustee for a variety of trusts that need either somebody here physically located in Delaware for a short period of time or even a successor trustee. But I’ve also spent many, many years building programs in trust management and trust administration. Because there is this crisis of human capital that just does not exist. I built multiple programs. They’re housed out of the University of Delaware. So I act as a trust and estate attorney, do planning, administration, I teach in the area, I build programs in the area, and I serve as a trustee. PEAK TRUST MANAGEMENT CERTIFICATE Frazer Rice (01:23.182)A full plate to be sure. To me, I came out of Wilmington Trust and another trust company served an individual trustee too. I’ve seen all these different flavors of trusteeship. My general sort of bon mot around that is that the individual trustees. I’d say 95 % or higher don’t really have an appreciation of the risk and responsibility that they’re taking on. And then the corporates have their own issues, which we’ll get into in a little bit. If we pull back even further, modern trust planning in wealth management, why is this so important? Jennifer Zelvin McCloskey (02:06.275)That’s massively important. It’s not just for the mass affluent or the ultra high net worth. It’s for everybody. We have all of these assets that we have this hyperfocus on building and increasing our wealth. Making sure that we have the ability to sustain ourselves throughout our entire lives. But if we don’t do this type of planning, if we don’t have structures and implementation for when we die, then our assets that we’ve planned so diligently for will fall off of a cliff. We lose the ability to control ultimately what happens to those assets. Layered on top of that, of course, is the tax component for ultra high net worth folks who are trying to really focus and direct their assets to make and create generational wealth transfers. Without this type of functionality and wealth planning and estate planning long-term, people lose control of what they’ve spent so much time building. Frazer Rice (03:13.338)One of the things I tell people as far as trusts are concerned is that, you know, we’re putting these structures together. They’re durable enough to withstand taxation or creditors or other asset protection features, create some guidelines around distributing the assets to the next generation or other constituencies. But also have some flexibility to be able to deal with the things we can’t look into the crystal ball and figure out over time. And that those three things just putting a document together that tries to do all that is hard enough, but then to put it in the hands of somebody or something to administer and to exercise discretion around it. That’s where the real art and science kind of stitched together and create this issue. You know, as we think about that too, the idea, the history of these types of scenarios kind of goes back to, you know, you’d put a structure in place and then you’d go hire a bank and they’d take care of everything. How do you look at that and say, all right, we’ve gone well past banks to individuals and then to dedicated institutions. What is the problem there? Jennifer Zelvin McCloskey (04:22.956)Now the problem, there’s two problems. In my opinion, what I see is that, you know, your individual trustee by and large is Uncle Joe, right? He’s the guy that everybody goes to in the family. The responsible one. He’s the smart one. The wealthy one who, great, doesn’t know what the fiduciary duties are. He doesn’t know that he has a duty of impartiality. He doesn’t know that… Frazer Rice (04:32.419)Right. Jennifer Zelvin McCloskey (04:48.475)He can’t self deal unless the instrument says so. Doesn’t understand how the instrument works. He doesn’t understand the nuance and the legalese written into the instrument. But he’s flying by the seat of his pants and everybody looks to him as the respected one in the family. No one knows that they have the ability to challenge him. So with your individual run of the mill trustee named in the instrument, they just don’t have the expertise, they don’t have the technical knowledge. Don’t know what they don’t know. They can get into trouble in that way. The other problem that you have with professional individual trustees oftentimes is that they are not formally trained. They may be an attorney who is working in that area, who’s doing plans for people who may or may not know what the full scope of being a trustee is. They may not realize, I have to get a special insurance policy because my malpractice insurance policy doesn’t actually cover this type of fiduciary engagement. There’s a lot of landmines that individuals can run into when they’re doing this type of work. On the corporate side, the problems that we run into is that there’s just a complete and utter lack. Frazer Rice (05:50.061)Hmm. Jennifer Zelvin McCloskey (06:12.059)Of available educational programs to teach people the proper way to be able to understand trusteeship. It has always been, and it just has developed over time through, you know, oh, we’ll give it to the bank, the bank will do it. This apprenticeship model, and that just does not scale well because if you learn improperly at the edge of a desk from somebody that learned improperly at the edge of the desk. Then the person that you’re teaching now at the edge of the desk is learning what you learned improperly. So anecdotally, I did karate for a long, long time. And the man who taught me karate, I’m almost a secondary black belt to like, was serious in karate. And the man who taught me karate said, you practice, it makes permanent. Don’t practice wrong. Because when you’re practicing wrong, you’re making permanent wrong things. And that’s what the apprenticeship model has the risk of lending itself to. It’s not that every trustee that learns at the edge of the desk learns wrong, but the risk is too high because the fiduciary responsibilities and the duties are too high to run that risk. The other problem is that we have a dwindling pool of really qualified senior trust officers because of just the nature of the job. You’re a human being, you’re an individual, you age, you retire. And it’s not something that people go to school and say, when I grow up, I want to be a trustee. They fall into it sideways. And unless there are academic programs that are out there that people are aware of and that they can get some formal training, some formal education to enter into the field. Frazer Rice (07:49.742)Yeah Jennifer Zelvin McCloskey (08:03.82)Separate and distinct from, I’m in the field and now I want to get a CTFA. I want to earn my certification to really show that I have the chops in this area. We have this shrinking pool of expertise. We have a lack of knowledge, a lack of formal education, and an apprenticeship model that doesn’t scale. On top of, with the individual side and the corporate side, this massive wealth transfer and an explosion of trust complexity that’s all taking place at the same time. Frazer Rice (08:31.918)One of the issues at the corporate level too is that as you say that the impregnance model is not necessarily the best way to do it. They’re cutting back on training programs. The business model around being a trustee or even a specific trustee does not make the big money. And so the ability for those types of institutions to develop the people.who ultimately are now in a very sort of pro-employee environment where there’s such a demand for trustees that they can kind of switch around and get a 10 or 20 % bump each time they go because people are desperate to have them. There’s a real cavern there to try to create the permanence that you’re looking for in a structure that really rewards consistency over time, especially as it relates to discretion and process of decision-making. Jennifer Zelvin McCloskey (09:23.15)Yeah, that’s exactly right. And that leads to this revolving door in the industry, because people are just trying to make more money and they’re going and bouncing to different trust companies. And there isn’t that backfill. Just because it’s a trust company and there’s policies and procedures, trusteeship is about relationships that you make with your beneficiaries, the relationships that you develop with multiple generations in a family. And when you have somebody that’s acting and serving in that and they move, they leave, they’re no longer acting and serving in that capacity, a new personality comes into the mix and it can really be disruptive. So having that consistency and minimizing the attrition is so valuable. Frazer Rice (10:06.766)The other thing I try to bring up, especially to individual trustees, is that the thing that you’re signing up for is probably going to look a lot different in five or 10 or 15 years when people are aged on, they remarry, they have kids, etc. That the conditions are a lot different than what they were before. And it’s going to be difficult to take on a structure that has eight people when before there were two. Jennifer Zelvin McCloskey (10:37.517)Yes, and that’s that complexity, that increased sophistication and complexity of trust structures that are available now to people. With the increase in the exemption, these trust structures, they’re not necessarily changed. For example, qualified personal residence trust, if people really need that anymore, but there’s a ton of them sitting around there. Are trustees properly administering it? Did you actually transfer the real estate into the trust at the time? So there’s all kinds of sophisticated structures that the trustees may or may not have the right skills. But they’re saddled with having to do it. Frazer Rice (11:19.47)Let’s take a step back and just talk about the functions of a trustee for a second. I break them down basically into three. Which is the first one. You have to administer the trust, meaning you have to dot the I’s, cross the T’s, make sure things get executed, tax returns are filed, statements get sent out to the extent that that happens, and that the administration of a structure like that occurs. Then I talk about the concept that the investments have to be made monitored moved around decided and that they’re appropriate for all classes of beneficiary that are in there and then the distribution function which is The assets have to be distributed according to the law. First the trust then maybe the intent or the law if everything is silent and that those three things are very different components and that it’s tough to find somebody who’s great at all three housed within one brain. Jennifer Zelvin McCloskey (12:20.217)Yeah, I agree with that 100%. It is a three legged stool. It’s the investments, the administration and the distributions. And in that administration umbrella in and of itself, there’s a tremendous amount of work that sort of goes unsung. know, it’s not the sexy stuff where you’re investing and making a bunch of money for your income beneficiaries and managing to preserve the corpus for your principal or your remainder beneficiaries. And it’s certainly not the personal interaction that you’re doing with your beneficiary day to day. Making distributions, helping them, seeing the product of that help. It’s the making sure you file ax returns are properly. Understanding how to read that tax return. Even if you’re not preparing it, making a proper selection on the accountant that you’re using to prepare those tax returns if you’re not preparing it. Make sure to set up statements properly, make sure that in this world of silent trust documents that you’re not sending a statement to somebody who’s not supposed to have it. Communicating with beneficiaries on an even keel. Making sure that you’re not inadvertently violating your duty of impartiality because it’s more than just a substantive duty, there’s a procedural duty as well. That’s really, really challenging to find within one human being, let alone add on top of it somebody who’s financially savvy enough to understand investments and all of the different complex investment tools that are out there, as well as having the personality and the interpersonal skills to keep beneficiaries engaged and happy. Frazer Rice (13:56.426)Just on top of that, the EQ, the bedside manner, and the ability to simplify the complex, et cetera. At the same time, that dedicated note taker that is able to document everything that happens within a decision. Whether distribution or investment or otherwise, that it’s just two different people most times. I find that something falls apart as time goes on. Ultimately if things aren’t laid out correctly, that’s when conflict starts to simmer. Then you know if there is something that’s wrong. That’s allowed to compound that’s where you get into a huge problem later on. Jennifer Zelvin McCloskey (14:36.922)It’s all that feeling. People are behaving in ways that they may or may not be able to articulate their emotional proximity to. When you’re talking with beneficiaries. There’s something simmering under the surface that you inherited because you’re a trustee. You may not even be aware of it because the beneficiaries may not even be able to articulate it. You have to have a certain sense. A gut check of feelings of rntuitively being able to read what’s going on under the surface. To pull it out of people in a very balanced and even keel way. It’s not an easy job by any stretch of the imagination. On top of financial literacy and personal liability and executive functioning skills, being detail oriented, making sure your documentation is not overly explicit. isn’t, you know, scarce. You’re now wondering how and why did you make those decisions? People don’t think about the decisions that they make on a day to day basis. We don’t think in a way to articulate why I made this decision. Why I exercised this type of judgment. And that’s what we’re being asked to do as trustees is to document what is my decision making process? Why am I making the decision? What are my factors involved in making that decision in a way that’s defensible. If we ever need to defend it. Frazer Rice (16:05.292)Well, in favoring one class of people over another is usually where the rubber hits the road on this. People who are used to seeing the income from a trust and don’t want that touched come hell or high water. Then future beneficiaries who’d like to see the trust go from X to 2X to 5X. So that they have something larger to enjoy. You have a natural tension that you have to manage. It’s just not easy. If you don’t document the hows and whys of what you’re doing, you set yourself up for a problem. From one class or another looking at you saying, you you should have done it differently. To go back to that liability component. You’re the only one who sits in the chair of having made that decision. You’re the one with the bullseye on your back when it’s called to account. Jennifer Zelvin McCloskey (16:53.093)That’s right, that is exactly right. And now add on top of it, you’re just named because you’re Uncle Joe and everybody goes to Uncle Joe. You have no technical background and you just don’t know the landmines that are there. You don’t know what you don’t know. Wouldn’t it be wonderful if we were able to create a pipeline of really sophisticated entry level employees or folks that are, you know sophisticated in financial literacy that now want to take the job to become trustees, that we were able to give them this technical roadmap for what the job actually is and then have them get the ability to apprentice on all of those policies and procedures. What does this corporation do? How do we document things? When you’re trying to learn it all at one time, it’s like drinking from a fire hose. Let’s give people the ability to really have a chance at doing it successfully. Frazer Rice (17:53.048)So let’s dive into that pipeline issue for a second. We already diagnosed that the, let’s call it the trust companies or the banks are, they’re just not resourced enough. They can’t run people through an internal school to do it quote unquote correctly. The apprentice model really kicks in. Which means you’re at the sort of mercy of what people are good at, not good at, et cetera. People turn over quickly so that apprenticeship doesn’t even work anymore. The RIAs I think are the worst place to learn about this type of thing. They have a completely different modus operandi as far as keeping clients happy. The word fiduciary means something so different to them than it does to an actual trustee. I wouldn’t feel good about the training on that front to sort of create trustees And then so law schools. They’re they’re just trying to create people the trust in the states vertical as a general matter. Let alone trying to delineate into a trustee situation. You’re putting the pipeline together and you put these programs together. How do you stitch together the needs and what does that manifest itself into? Jennifer Zelvin McCloskey (19:07.642)So that’s a really, really good question. I think that the very first place that we start with answering that question is advising on a trust as an attorney. It’s different from the administration of a trust and the skills that you need for that. So when you create a program like this where you’re trying to teach about trust management. You have to start with the technical skill. The legal side of what is it that we’re even doing? What is a trust? What are the fiduciary duties? Where do they come from? Then we have to, after we teach or create a structure or foundation on what the legality is. Now we go into how does this translate into administration? So when I created the programs, I looked at what’s the law they need to know? What is the level of sophistication of the student? And what do I need to, from a foundational perspective, teach first? What are the building blocks? And then how do I translate that into administration? The one thing that I have found is trust law does not equal investment management. So if people are coming along… Frazer Rice (20:26.254)No question. I’m nodding audibly at that comment. I like that. Jennifer Zelvin McCloskey (20:31.226)Your fiduciary duties as a trustee are fundamentally different than those of an RIA, where some RIAs are not even fiduciaries by law. They’re not. So being able to delineate and explain where that line is, what makes you a fiduciary, what are those duties, after you know the legal basics. And taught to you at a level that you can understand. I don’t expect everybody to be a lawyer. And people have asked me time and time again, do I need to be a lawyer to know this? No, you don’t need to be a lawyer because you’re not advising on the law. You’re advising on the administration of a legal structure and how that administration affects the fiduciary duties that are inherent in the relationship. Then how those fiduciary duties are translated out to the beneficiary. That’s the way that I’ve always built these programs. Where do I start? Start with the law. Where do I go from there? Start with how the administration translates the law. And then how does that administration get heard by the beneficiary? Where does the RIA come into the mix? The RIA should not be dabbling in advising on trusts. They should know that they need to bring in somebody who has this particular skill. And if they’re not doing that, they’re doing the client a disservice by trying to give one-stop shop advice. Frazer Rice (22:06.85)Yep, no question about it. One of the things that…we delve into the world of trusts and their function, et cetera, is that you’re dealing with an ecosystem from client to outside advisor, whether RIA or even accountant, et cetera, that they’re looking for certainty and airtight. quality to these structures that you put them in place and then everything runs like a clock going forward. When in actuality, I think there is a bandwidth of risk around everything. And so it’s the poor trust officer or individual trustee who sometimes has to be the bearer of bad news to say, yeah, you know, I think this is going to work 98 % of the time, but there’s a 2 % problem here or we’ve got this to fix or something like that and everybody else sort of sighs with disappointment and gets mad at the administrative function when in actuality they’re really doing their job and trying to, you know, keep a lot of things that are spinning out of control kind of within view. How do you get a trust officer or that administrative function or even the full trustee function to be comfortable with that risk and everything that’s involved with that? Jennifer Zelvin McCloskey (23:20.504)You have to start with explaining that there is risk and we’re not our job is not as a trustee to eliminate risk. Our job is to manage and identify risk. It is inherent in the job. There is going to be risk. No matter what you do, you cannot divorce risk from trusteeship. It’s a matter of identifying perceived risk and actual risk. And if you can teach that, if you can teach These are the things that are going to trigger a likely outcome. They’re gonna trigger a likely risk. Then you can essentially, you can’t foresee everything. I mean, there are things that are just gonna happen. But in a trust instrument, you’ve got contingency plan upon contingency plan upon contingency plan. That’s what the flexibility of those structures are building. We need to, as trustees, be able to recognize What is the risk with contingency plan A? The risk with B? What is the risk with C? How can we minimize the risk? And how can we make sure that we’re managing perception of risk versus actual risk? Frazer Rice (24:29.31)as someone who’s been in trust companies, advised trust companies, advised trustees, and advised clients, the lack of appreciation for the management of that risk and that that as the intersection of the business model of trusteeship and risk management and use of discretion and making hard decisions and even kind of an insurance quality around these structures, how do you fix that, where people place a level of respect on the job that I think is completely lacking in the wealth management ecosystem? Jennifer Zelvin McCloskey (25:09.089)Absolutely. It’s a tough one to answer. How do you fix it? First and foremost, I think that it’s a top-down fix, especially at a corporate trust company, a bank, and even an independent trust company that’s not affiliated with a bank. The management has to… really understand the function of the trust company. For so long, it’s been just an extra service that we provide and and we’ll do this, the back office trust company. It’s really, really important that the management recognizes what the functionality of the trust company is and stops treating it as sort of a back office stepchild. From the corporate level, I think that’s the very first place we start. Frazer Rice (25:38.478)Mm-hmm. Jennifer Zelvin McCloskey (25:57.818)The second place we start is investing in our trust officers, investing in the team, giving them the education that they need, continuing to give them education, providing training programs, whether they be in-house, external, bring in trainers. None of this is set it and forget it. At the individual level, I think it’s really, really important to have functions like the Individual Trustee Alliance, groups like that, where you have an ability to talk to other professionals that are doing what you’re doing. That’s another way to impress upon people that we have to manage the risk and we can’t do it all alone. Nobody knows everything. You really have to, you have to talk to other people. You have to engage. have to, what is it called when we were practicing law and we’re a little bit outside of our comfort zone, we have to consult with other people who know more than we do. It’s our obligation as lawyers. It’s the same thing with a trust company, with a trustee, whether you’re an individual or you’re not. Widen that circle. Frazer Rice (27:08.474)I think this is my idea for the day that there’s got to be a bit of a public relations campaign sort of describing what’s going on here because I think especially when we go into the family members that sort of occupy these roles, they have no earthly idea what they’re doing. They’re usually doing it for free. Everything’s hunky dory up until a point and everyone hopes that everyone is not going to sue each other if something goes wrong. But the level of wealth that’s being transferred now is now so significant that everyone sort of talks about, AI is going to get rid of lawyers. Nope, not in fiduciary litigation. I think that’s a medium term growth industry, especially around insurance, around ILITs, around revocable trusts, around elder care. But this is my advertisement for people who are in law school looking for a productive way to go. I think that one is going to be, I think that one’s recession proof, at least for a while until I retire anyway. So my thought is that awareness over these things, and it’s probably going to take a very difficult case or a class action suit, something like that, where somebody really gets hurt in order for that awareness to come up. Jennifer Zelvin McCloskey (28:24.922)Yeah, I would agree. think that some of the solutions would include better trust education, you know, whether it be for RIAs, lawyers. Trust in the states is a throwaway class in law school. And there are so many law schools that are essentially rolling it back because bar exams aren’t testing it anymore in a variety of states. And ACTEC is definitely working with the law schools to try and increase trust in the states being taught and certainly being tested. So education for lawyers coming out of law school, education for RIAs that are advising on trusts, education for trust officers, for trust administrators, trust professionals in general, clear role delineation. What is the role of the RIA? The role of the trust officer? What is the role of the trustee if they’re an individual trustee? And then creating a culture of collaboration on what we’re doing as a team for the beneficiary, not substitution, but collaboration with the advisors and the trustees. Frazer Rice (29:32.59)Let’s go into the role delineation for a second. About 20 or 30 years ago, the concept of bifurcating or sort of cordoning off the different functions I described before the investment, the administration and the distribution has come into vogue. I think that came out of frustration with bank trust companies where you got one set of advice for every trust that they had as far as investments and distributions and administration and a lot of modern larger families wanted something a little bit more specific to their needs. And that’s really turned, it’s exploded as an industry for increasing sophistication and size of wealth. Along those different functions, where maybe the administration goes to a professional trust company or a trust officer in the state that you want, Then there’s some intersection maybe in the distribution committee. And then the investment side of it is a bit of a free for all, think, depending on what you’re, dealing with. How do you educate the, that continued the delineation, but the coordination within those types of structures. Jennifer Zelvin McCloskey (30:41.275)Yeah, I think it’s really important. And I’m a Delaware lawyer. I’m licensed in multiple states, but Delaware is my home. It’s where I learned how to be a lawyer. It’s where I grew up as a lawyer. So this directed trust model that you’re describing, where you’re bifurcating, truly bifurcating these particular functionalities of a trustee, it originated in Delaware. sort of, we didn’t, I mean, we invented it, right? We codified it. It was being done, but we codified it. The idea of making sure that everybody understands what their function is and knowing that there’s a limit of liability that’s built into the instrument and communicating what that means to the RIA that is named in the document. I can’t tell you how many times I have heard companies, heard trust companies say, we’re advisor friendly. And I’m like, not unless you’re directed, you’re not. Frazer Rice (31:37.528) “THE TRUSTEE CRISIS: Navigating the Challenges”Yeah. Jennifer Zelvin McCloskey (31:40.439)If you are directed, you are 100 % advisor friendly because there’s no chance that that trustee is going to try and take the investment management. They’re not a portfolio manager. Not a clerical administrator. They’re not a passive rule follower. We need to identify what does that trustee actually do when they are an administrative or directed trustee. Clarify that role so that people who are engaged in this bifurcation, this structure where we’ve got a distribution committee, maybe it’s individuals who are close to the family, close to the beneficiaries, where you don’t have somebody who’s objectively uninvolved with the family members making decisions as to whether or not there’s a distribution that should be made. But also advising those rolls those advisors that your administrative trustee is not just a pencil put a paper pusher. Not just checking boxes. They really do add value to the role that they provide and making sure that everybody understands what each other are doing, having regular meetings amongst the team instead of operating in a vacuum or operating in a silo. And taking the approach of it’s not my job, misunderstanding trustee powers and the advisor’s authority. So when that’s delineated, when that’s really understood, not just by the advisors, but also by the beneficiaries, there are so many beneficiaries out there, Frazer, that have absolutely no idea that they actually hold all the cards. They don’t know. Frazer Rice (33:25.87)Along that line, so in the administrative, we just walked through pretty nicely. The distribution function is one that, let’s talk a little bit for a second about what it means to ask a trustee for a distribution and maybe the difference between income and principal and why having a steady hand at the wheel within that function, whether it’s a corporate trust company of qualified individual or family input in that function, why real good thought needs to go into how that’s staffed. Jennifer Zelvin McCloskey (34:04.73)Yeah, absolutely. 100%. In a corporate trustee ship or a corporate trust company structure, there’s always going to be distribution committees, right? So if you are the trustee, you’re going to have to go through a committee that’s looking at what your reasoning is for making that distribution. They’re asking questions about what have been the prior distributions? Have they come from principal? Have they come from income? What is the spend rate on that trust? How is this going to affect long-term spend rate? Is this an aberration? Is this something that’s gonna become a habit? Really understanding what the distribution, the guidelines are in the trust. What is the distribution standard? Making that decision? What are our factors? And how many people are at the table? Who’s communicating that to the beneficiary? Does the beneficiary know that the trust officer alone does not have the ability to say yes or no? That when they’re in this ecosystem of a corporate trust company, they have their checks and balances to make sure that that risk is being managed. So when you’re looking at corporate trust companies, are a lot of layers behind understanding what the distribution standard is, whether it’s hems or if it’s purely discretionary. The other thing that you need to look at when it’s not a corporate trustee and it’s an individual trustee is, how is that individual trustee making that decision? Are they doing it in a vacuum? Alone? Are they favoring one beneficiary over another because they like them more, you need to have some communication to the beneficiaries so that they understand what they are, what their interest is, what they are entitled to, if anything, and why the trustee stands in that position as the gatekeeper. And I really think in my heart of hearts, we need to make a shift from a gatekeeper trustee Jennifer Zelvin McCloskey (36:16.708)to a beneficiary enhancement trustee, where the beneficiary is really taking on the understanding that the trustee is there to facilitate enhancing the beneficiary’s life. That even though the trust may have started at the outset as a tax strategy or something that the grantor decided they needed to do with the advice of counsel. At the end of the day, you wouldn’t have been named as the beneficiary if there wasn’t some sense of love or obligation even, that it’s for your benefit. It’s in the name. Beneficiary. Trustees need to understand that and beneficiaries need to be taught. Frazer Rice (36:54.958)Right. Frazer Rice (37:00.646)And it goes to the circle back to the notion of making sure that you write down the whys of the decision because ultimately if the concepts of favoritism or you didn’t communicate this or anything, the idea of having the beneficiary submit a budget but having them understand why they are submitting a budget and then if there is some discretion that’s happening around that decision that the data points that are informing that discretion, that’s gonna keep everybody safe a lot later on. Jennifer Zelvin McCloskey (37:32.666)Absolutely. I break it down into a couple of different factors. It’s fiduciary decision making. How is that fiduciary making the decisions they’re making? Why are they making those decisions? And who is being affected by the decisions? Document interpretation. Do you understand the document that you’re administering? If you don’t understand the document you’re administering, hopefully best case scenario, you know what you don’t know and you ask. But if you don’t understand the document and you don’t even have the wherewithal to say, hey, I need help to understand the document, it’s really problematic. The third part, balancing beneficiary interests. Really taking on board this idea of the principal income problem that all the assets in the trust are not the same. That some of it doesn’t at all in any way affect a certain class of beneficiaries. And at the same time, it’s inextricably intertwined in the way that it affects another class of beneficiaries. And then risk management and governance. How is this being governed? How are we managing perceived and actual risk as a trustee? Frazer Rice (38:40.13)The investment function, which I alluded to before, I see storm clouds on that horizon, not really at the RIA level, because I think there’s sort of a default mode that investment policy statements are in place. Diversification is a true commodity at this point. And I never really worry about an RIA sort of understanding how to invest to get to a certain expected return and deal with the risks and drawdown and all that stuff. The storm cloud I see is when individuals sit in that role and they are being tasked with, let’s call it quote unquote, overseeing concentration, meaning that trust is holding a building, farmland, a nuclear reactor, crypto, all of these different things that sometimes can be, A, they have their own different maintenance responsibilities that are not just looking at a fidelity statement, but that they also have their own volatility And, you know, in the case of a building, you got to make sure it’s managed correctly. are they going to get sued or the windows kept up, all of that stuff, and that there’s a whole different component there. And I’m waiting for the shoe to drop on some fact pattern there where somebody is sitting in the role of an investment advisor. It doesn’t say trustee in the document, so they don’t really think that they have trustee liability. But. they sit in that role and all of a sudden somebody finds 10 55 gallon drums of green fluid in the basement of a building and all of a sudden the trust has a big set of red brackets that say minus $100 million that you owe to the federal government and the EPA. How do you think about that? Jennifer Zelvin McCloskey (40:21.454)Hmm. Jennifer Zelvin McCloskey (40:25.242)That’s a heavy question. so the Delaware stock answer, obviously, direct it, right? It’s just to get the trust, cut off the liability. At the first, at the inception of your hypothetical is bad drafting, right? So if there’s no statement as to whether or not your investment advisor is acting as a fiduciary or not, Frazer Rice (40:35.042)Right. Jennifer Zelvin McCloskey (40:52.836)What does your statute say? Does your statute impose that they are as a default a fiduciary or not? So that’s the very first step. That’s bad drafting. We need to know. But if it’s silent, let’s say it’s just a lousy document, there’s, God knows. Anybody who’s seen trust documents knows that, you’ve seen them all, right? And everything in between. Some are good, some are bad. If this is a bad one. Frazer Rice (41:13.08)Seen good and you’ve seen bad. Jennifer Zelvin McCloskey (41:20.079)Then we need to document the statute. If we can correct it, modify the document, let’s modify it. But if all of that can’t happen, then I would say the best way to handle it, make sure you have adequate insurance. mean, over-insure that, over-insure it. Make sure that there’s regular checks on the actual… Assets that are in the trust, if you have a concentration and that concentration is real estate, get the advice of counsel, put that bad boy into an LLC, get yourself some distance from the actual asset itself being held in the trust, hold an interest, hold a financial interest, push it down to the corporate level. But if you can’t do all of that and you’ve got those 500 gallon drums of green fluid and now you’re… Frazer Rice (42:14.286)You Jennifer Zelvin McCloskey (42:15.371)You you’ve got a super fun site. What do you do? You don’t shy away from it. Have to address it head on. You got to take the accountability. You got to communicate and document, communicate and document some more. Talk to your beneficiaries. Make sure that they’re aware of where it went wrong, why it went wrong. Because I have found in my exposure in the industry over time and in reading case law, it’s when you’re trying to cover stuff up. Frazer Rice (42:43.913)Jennifer Zelvin McCloskey (42:44.027)You’re just making more problems. Bad news doesn’t age well. It doesn’t get better over time. You have to approach it head on and make sure that there’s communication and documentation. Meet with your beneficiaries. If there’s a trusteeship where you are appointed as a trustee individually and you’re not having at least quarterly meetings with your beneficiaries, If you’re not going out and seeing the asset, if you’re not going out and making sure that the asset is properly custodyed, you’re not, you’re violating your fiduciary duty. You are not doing what you’re supposed to do. Frazer Rice (43:21.804)You brought up an interesting word there, custody, which is the administrative function, whether held corporately or individually, one of the major things you have to do is to safeguard the assets. And that’s a big two syllable word that carries a lot of weight with it. That custodial function, how do you teach the trust officers or the individual trustees where that starts and stops? Jennifer Zelvin McCloskey (43:48.579)Yeah, mean, custody is super, it’s a really touchy, touchy subject, especially with the dynamic way that trusts have developed in the current climate from tangibles. You know, I’ve got artwork and my beneficiary wants to hang the artwork in their house. Well, do you have custody? Has it been assigned to the trustee and how do you maintain that asset? Make sure nothing’s happening to it. Do make an appointment, go over to the, visit your artwork? What if it’s prize horses, you know? What if it’s, you know, a stud that, you know, we’re gonna need to breed and it’s gonna be the next Triple Crown winner? How do you make sure that the barn is properly safeguarded? It’s a really touchy subject, especially with things like tangibles and things like assets held away when you technically custody the asset, but you don’t have control over the asset. I think in the education part for custodying, what I do in my programs and when I teach this is I make sure that we talk about different types of asset classes. And what the risks, again, what are the risks that you run with these asset classes? How can we manage the actual and the perceived risk of holding that asset? Even if you have custody and name only, but you don’t have physical custody, how do you maintain your control over that asset? Because it’s really the C’s, right? The custody and control. Just because you don’t have custody doesn’t mean you don’t have control. So we have to make sure that there’s an education that’s provided about the different asset classes, whether it’s tangibles, intangibles, assets held away, if it’s a concentration of stock, if it’s crypto, and most trust companies are not taking crypto. I think that there’s like a circuitous way that they’re getting in right now, but it all boils down to education, isolating what the issue is and educating people on it. Frazer Rice (45:59.586)I’ll give you a third C, it’s consequences, which is what happens when you don’t understand these functions. on the crypto side of things, Jennifer Zelvin McCloskey (46:01.786)Uhhh Frazer Rice (46:11.544)Holds the key to get to the crypto. What happens if that trust officer quits and walks away with the key and they’re like, well, multi-sigil figure this out. I’m like, okay, that’s not that. That doesn’t make me feel great at the moment. And now there have been some advances, which is good, but traps for the unwary to be sure. the good news too for crypto is for people who want exposure, the spot ETFs take away 90 % of the problems with that. But as we start to think about winding down here, because I have a feeling we could probably talk for four or five hours on this subject, when putting your programs together, what does a curriculum look like? And we don’t have to go through it bit by bit, but how does that work when someone comes to your program? How much time does it take? What’s the commitment? Jennifer Zelvin McCloskey (46:47.172)Yeah, I think so. Frazer Rice (46:54.851)Mm-hmm. Jennifer Zelvin McCloskey (47:06.33)So the program that I created that’s really available anywhere across the country is called the Peak Trust Management Certificate Program. Peak Trust Company, may be familiar with it. They have name rights because they gave the donation to the University of Delaware for me to build the program. So it’s housed at the Lerner College at the University of Delaware, but bears the name of Peak Trust Company. I look at five different things. The first thing is trust law and administration. So like I said previously when we were talking, you lay that foundation of what is the legal component of this? What is the baseline that people have to know? And then what is the administration? The second component is, and it’s inextricably intertwined as taxation. What is the income tax? What are the deductions? And now let’s take all of that income tax knowledge, individual income tax knowledge, and build on it with fiduciary income tax. What is DNI? What is FAI? How does it go out to the beneficiary? What’s the character of the distribution? How do we manage that? What are we deducting in the trust? So teaching taxation and not because trustees necessarily are tax preparers, but because the trustees obligation is to be able to understand and read that tax return, they need to know how to spot problems. So from my perspective, teaching fiduciary income tax is a critical component. It also helps. Yeah. Frazer Rice (48:38.828)No, no, I was gonna say no question about that. And there are elections to make, just because it doesn’t just go on autopilot, there are choices to be made so that if you’re the trustee, you may not have to prepare the tax return, but you may have to make a choice on the tax return and you’ve got to be informed because that can be an issue. Jennifer Zelvin McCloskey (48:58.651)65 day elections, perfect example, right? You just, you need to understand what your role is and how it overlaps with that of the CPA. The third part, of course, investments. Investments are inextricably intertwined, whether you’re doing it yourself as the trustee or you’re directed or even delegated, which is like the hairy scaries of every trusteeship known to man, because you’re not actually in control, but you’re responsible. So it’s the gray. When I build a program, because of the, you know, the directed trusteeship being so popular in today’s day and age, we have to talk about not just investments of, you know, marketable securities, not just the custody of tangibles, but also subscription documents, because so many alternatives are held in trust right now. unique assets, need to know how the trustee is actually carrying out their fiduciary duty when it comes to engaging in an investment that is an alternative investment. The fourth component is of course compliance. We cannot ever get away from compliance and I think we could do a whole nother podcast on compliance in trusteeship but. You know, it’s a regulated entity. And even if you’re an individual trustee and you’re not using what those compliance frameworks are, what the guidelines are by OCC, Reg 9, FDIC, if you’re not looking at that and using that as a guideline, don’t do the job. understanding KYC, BSA, AML, all of those compliance components that have tentacles. That’s the fourth part. And then for the fifth part of this program, because it’s specifically geared toward trustee education in trust companies, although it can be applicable, very applicable to individuals, is operations. I was very fortunate that I was able to partner with SCI on building the operations component. So we license their platform called Plato. It’s essentially their training platform. Jennifer Zelvin McCloskey (51:12.888)so that trustees can see how fees are set up, fees, that’s a whole other podcast, fees, statements, distributions, how are we doing this? How are we documenting everything? What are the logistics of the day-to-day operations? So that’s how I built the program and it’s available anywhere in the country. It’s 10 weeks, how long does it take? I would say from three to five hours a week of an investment that you’re making at a bare minimum. Obviously there’s a whole lot more of depth that you can go into. The resources are built in. But I would say 10 weeks, about 50 hours of time where you’re actually engaging with the material. And then I bring in guest lecturers on each different area of expertise for lack of a better description. And they get a certificate at the end, they get a digital badge, and now they really have something where they can add value day one in a trust company or as a trustee. Frazer Rice (52:17.902)With Delaware being, you one of the real gold standards as far as trust jurisdiction, I assume that everything that comes out of this program is pretty transportable to the other useful jurisdictions, let’s call it, within the country. know, the Tennessee’s, the South Dakota’s, the Nevada’s, the Alaska’s, Wyoming’s, New Hampshire’s, et cetera. Obviously, there are hairs to split with different foibles in their law, but everything that you’re describing sounds like works everywhere else. Jennifer Zelvin McCloskey (52:47.928)And I’ve always taken the approach, you’re 100 % correct, I’ve always taken the approach of UTC. I base everything off of UTC and if there’s something different or unique based upon the jurisdiction that you’re in, I always encourage people you have to look at your statute, you have to look at the jurisdiction that you’re actually practicing this in and administering in. I use Delaware, South Dakota, Alaska as examples quite often when we’re talking about the directed stuff, but By and large, it’s UTC. Frazer Rice (53:20.966)It just a weird subset. So special needs trusts and islets, which are two types of trusts, very specific. One holds life insurance. The other is designed to really take care of people who can’t take care of themselves. And they are types of trusts that a lot of trust companies don’t like to take on because the liability is harder or the profit margin is less. For those individuals who get the opportunity to participate in those and I put that in air quotes. How would you advise people to get ready for those types of situations? Jennifer Zelvin McCloskey (53:58.308)People who are in need of those types of trusts. Frazer Rice (54:02.122)Well, maybe both. The people who need those trusts, you know, they’re going to, they, you know, it’s almost like they get set up and then the staffing gets kind of figured out later, barely. And then, you know, the, for the people who end up taking on that role, they really have no idea of what they’re in for in a sense. Is there sort of like a mini, I’m not going to say a full course like you’re describing, but a crash course in, in what’s going on here and what can I do to keep myself safe? Jennifer Zelvin McCloskey (54:30.271)Unfortunately, no, I don’t know of one. and there isn’t much built in. there’s, we talk about a little bit in the program that I built, but, those are specialized and eyelets we talk about a little bit more there, you eyelets had their day and sort of they has done ish. but special needs trust. It’s a whole other ball game because It really incorporates state law and social security and Medicaid, all of those government benefits that I think you would need something more specialized than my program that I developed. And I don’t have a great answer for that, I’m sorry. Frazer Rice (55:12.482)No, there’s not a great answer for it because it’s tough. it’s a, all of which is to say for someone who’s involved with those things and feels confused by what’s going on, that’s one where it’s worth it to spend the money to lean on a dedicated Medicaid elder care, special needs type of lawyer on that front because there are traps for the unwary. Okay, now we’re starting to butt up against an hour here of. Jennifer Zelvin McCloskey (55:29.764)Yes . . . Frazer Rice (55:38.827)Four hours. No, I’m kidding listeners. We’re not going to talk for four hours, but How do people find your program and and then I’ll ask a bonus question at the end Jennifer Zelvin McCloskey (55:49.339)So the program is on the University of Delaware’s website. You just type in peak trust management certificate and it’ll pop up. My name will be there. I think my picture might be there. It’s all over my LinkedIn. So if you look me up, you’re going to see the peak trust management certificate program. You can always email me, jennifer at zeldenlaw.com. Happy to push people into it. start, I’m in the new cohort right now. We’re two weeks into a 10 week program. But we have a new cohort starting in May. I think it’s May 4th. So may the fourth be with you. Frazer Rice (56:24.622)Terrific. So the final question here is really more of a crystal ball question. In this trust industry, trustee industry, what are the real, I’m going to say opportunities out there, and we’ve sort of painted a picture of doom and gloom and its low profit margin and things like that. Where can someone who is thinking from a business perspective about this find something? Once they’re properly educated about it and being able to participate in it. Jennifer Zelvin McCloskey (56:57.582)There are so many opportunities. There is an absolute need for good trustees everywhere. Trust companies from coast to coast, individual trustee alliance. People really, really need trustees. There’s tremendous opportunity with Heritage Institute, not the Heritage Foundation, but the Heritage Institute. There’s opportunities with…various family offices and various trust companies for education, for beneficiary education. So many opportunities out there. Trust companies are just clamoring for people. So if people are interested in becoming a trustee, getting that education, you will not have a hard time finding a job. Like you said, it’s basically recession proof. This wealth is going to transfer. We need sophisticated, knowledgeable trustees. on the receiving end of that transfer so that it happens correctly. Frazer Rice (57:56.578)I’d go so far as to say financial advisors. I just gotta say, a CFP is useful, CFA is on your investment side, but something like this, you know so much more about how intergenerational wealth works than what’s happening in those particular situations that I think it helps people stand out when I see something like that on a resume. Jennifer Zelvin McCloskey (58:00.302) “THE TRUSTEE CRISIS: Navigating the Challenges”That’s all the podcast. I hear you. I hear you. Frazer Rice (58:24.386) “THE TRUSTEE CRISIS: Navigating the Challenges”All right, with that, Jennifer, it’s great to catch up and I will have all of your information on the show notes and I will either see you at the ITA conference in Dallas or what I’m down in Delaware next. More Around “THE TRUSTEE CRISIS: Navigating the Challenges” BUILDING A TRUST COMPANY TENNESSEE AS A JURISDICTION DIRECTED TRUSTEES DELAWARE WELL BEING TRUST THE TRUSTEE CRISIS: Navigating the Challenges https://www.amazon.com/Wealth-Actually-Intelligent-Decision-Making-1-ebook/dp/B07FPQJJQT/ Keywords for THE TRUSTEE CRISIS: Navigating the Challenges trusteeship, wealth transfer, trust management, fiduciary duties, trust education, estate planning, risk management, trust administration, individual trustees, trust companies, the trustee crisis, navigating the challenges, the great wealth transfer,

From the archive - This episode was originally recorded and published in 2022. Our interviews on Entrepreneurs On Fire are meant to be evergreen, and we do our best to confirm that all offers and URL's in these archive episodes are still relevant. Austin Netzley is the founder of 2X and author of the updated book, From 6 to 7 Figures. Top 3 Value Bombs 1. A big key to success is not as sexy as a lot of gurus out there talk about. Think about operational excellence, turning your business into a machine that can thrive without you. If you take that perspective, it's going to lead to a very healthy business. 2. Make the strategic decisions on how you're going to be different, how you're going to solve real problems, and who your customer is. It will make your entrepreneur life much easier. 3. Systems make things much more repeatable, consistent, and not reliant on you. Get a FREE copy of the updated & expanded book for a limited-time only - From 6 to 7 Figures Sponsors HighLevel - The ultimate all-in-one platform for entrepreneurs, marketers, coaches, and agencies. Learn more at HighLevelFire.com. Freedom Circle - A powerful community of entrepreneurs led by JLD. Are you ready to go from idea to income in 90-days? Visit Freedom-Circle.com to learn more.

Want to get away from the crowds? Want a high mountain lake or stream all to yourself?  The best way to do this is to take a backpacking trip, but you need to prepare more than you would for a car trip or a trip to a lodge. What exactly should you take and what should you leave behind?  What kinds of flies and accessories should you bring? How can you save weight and still have enough gear for a fun fishing trip? Derek Bargaehr [37:36], an experienced fly fisher and backpacker, gives us tips on how to make the most of your next backpacking trip. In the Fly Box this week, we have some questions. A couple of which could only be answered by my co-workers at Orvis so we have responses from both Pete Kutzer, our casting guru and Shawn Brillon, our bamboo rod craftsman. How can I easily estimate how much backing is on my unlabeled reels? A listener relates how some podcast advice on emergers helped him and his son have a successful trip I took a lesson on two-handed casting and it was all done on grass.  Was this wrong? What advice do you have on cleaning the ferrules on bamboo fly rods? Are Orvis bamboo fly rods impregnated? On a tarpon trip, the fish were in deep water so I used a sinking poly leader on my floating line.  Should I have used a full-sinking fly line instead? Is the Albright knot a better knot than the nail knot for attaching a leader to a fly line or backing to a fly line? When connecting pieces of tippet I will normally go up two X sizes, like from 2X to 4X.  Is this wrong? Is it OK to clear a casting lane on a trout stream? What can I do to find bigger trout during the dog days of summer?

What happens when a medical device executive walks away from a stable 17-year career to build a digital marketing agency in one of the most complex industries on Earth? In this compelling episode of The Proven Entrepreneur Show, host Don Williams sits down with Saul Marquez, founder and CEO of Outcomes Rocket, to uncover the untold story of pivoting, perseverance, and the power of trusting your gut in healthcare entrepreneurship.Before he founded Outcomes Rocket, Saul Marquez spent 17 years in medical devices, living what many would call a "dream career." In this candid conversation with host Don Williams on The Proven Entrepreneur Show, Saul shares how he walked away from that security to build a specialist healthcare marketing agency serving health tech, medical device companies, and ambulatory provider groups. If you've ever felt torn between a safe path and a bigger vision, this episode will feel uncomfortably familiar in the best way.You'll hear Saul break down his simple decision framework of "dreams and data"—how he used both gut instinct and real-world evidence from over 2,000 podcast interviews with healthcare leaders to validate his leap into healthcare entrepreneurship. The discussion dives deep into health tech marketing, medical device marketing, and why "you sell like you buy" is one of the most practical mindset checks for any founder, CEO, or marketing leader. They unpack why "cheap marketing" is so expensive, when to stop DIYing and hire real experts, and how to think about podcasts and webinars for lead generation in today's healthcare landscape.Saul also shares a raw story about a webinar gone wrong—a painful early failure that turned into a long-term client relationship and bulletproof SOPs (Standard Operating Procedures) through radical ownership. Along the way, Don and Saul talk about comfort zones, "burn the boats" moments, choosing mentors wisely, and what big thinkers in healthcare are really struggling with behind the scenes. This conversation covers essential topics for anyone navigating the intersection of healthcare business growth and digital transformation.What You'll Learn:How to validate a business idea using the "dreams and data" framework before taking the leapWhy healthcare commercialization requires strategic marketing partnerships, not budget solutionsThe psychological barriers keeping healthcare professionals from becoming successful entrepreneursHow B2B healthcare marketing differs from other industries and why it requires specialized expertiseThe role of content marketing, webinars, and podcasts in healthcare lead generationWhy 10X growth is easier than 2X growth—and what that means for your strategyHow to recover from catastrophic failures and turn them into competitive advantagesThe importance of surrounding yourself with mentors who've already achieved your goalsPractical strategies for health tech and medical device company growthThe "three M's of leadership": Mindset, Mission, and MethodologyWhy This Episode Matters:This isn't generic entrepreneurship advice—it's a masterclass in healthcare business strategy from someone who's built a thriving agency while serving some of the most complex organizations in the industry. If you're a health tech founder, medical device company leader, healthcare CEO, ambulatory provider, or anyone in the healthcare sales space wondering why your marketing isn't working, this conversation will rewire how you think about growth.Subscribe to The Proven Entrepreneur Show for more unfiltered conversations with leaders building billion-dollar healthcare businesses and navigating healthcare commercialization.

Try Activations free for 2 weeks + get a huge discount at https://www.activations.com/podcast/sahara   In this episode, I'm sitting down with the incredible Mimi Bouchard, founder of Activations, to talk about abundance, manifestation, career pivots, and money mindset.   One of the biggest questions I get is how I created a business where I'm literally paid to be me, making seven figures while talking about past lives, sacred feminine energy, and rose serpentine between Bali, Egypt, and India.    Growing up, no woman in my family even worked. I was told you get a job you hate, and that's life. I thought it was either do what you love and be broke, or sell your soul to make money.   Mimi's story is so similar. She didn't come from wealth. She grew up watching her parents fight about money, surrounded by scarcity and fear. But she made these massive mindset shifts that completely transformed her reality.    Now she's living in the Bahamas, married to her dream partner, and running a seven-figure business with Activations, a revolutionary audio app that's not meditation, not affirmations, it's activation. Over 700 guided visualizations paired with cinematic music designed to rewire your brain while you move through your day.   We dive into: ✨ How to rewire limiting beliefs around money and abundance

What if perfectionism is actually killing your success? In this Habits and Hustle episode, Kim Perell, a nine-time serial entrepreneur, joins me to share why waiting to feel ready is the biggest mistake aspiring business owners make.  We dive into Kim's morning routine, her partnership with Jay Shetty on Junie, and why exercise is her number one productivity tool. We also discuss why iteration beats innovation, and how she balances building multiple companies while raising four kids. Kim Perell is a 9X founder, 2X bestselling author, and investor in 100+ companies. Kim is a dynamic TV personality on Entrepreneur Magazine's Elevator Pitch and regularly appears on Good Morning America, The Today Show, CNBC, Fox, and in Forbes, Inc., and The New York Times. Her book "Mistakes That Made Me A Millionaire" shares the unfiltered truth about the journey to success, proving that every mistake holds the potential for million-dollar lessons. What We Discuss: 04:20 - The number one mistake: waiting to feel 100% ready before starting 05:09 - The Marine Corps 70% solution and how to apply it 06:56 - Why iteration beats innovation (and saves time) 07:21 - Co-founding Junie with Jay Shetty in a crowded beverage market 58:10 - Daily routine: waking at 6 AM, red light therapy, and meditation 59:28 - Running a household with four kids like a company 01:00:29 - Workout routine: HIIT, Peloton, and personal training at home 01:01:08 - Supplement stack: Momentous protein and creatine 01:03:00 - Why exercise is about mental health and focus, not competition 01:03:27 - Exercise as the number one longevity hack above all supplements …and more! Thank you to our sponsors: Therasage: Head over to therasage.com and use code Be Bold for 15% off  Air Doctor: Go to airdoctorpro.com and use promo code HUSTLE for up to $300 off and a 3-year warranty on air purifiers. Magic Mind: Head over to www.magicmind.com/jen and use code Jen at checkout. Momentous: Shop this link and use code Jen for 20% off  Manna Vitality: Visit mannavitality.com and use code JENNIFER20 for 20% off your order  Prolon: Get 30% off sitewide plus a $40 bonus gift when you subscribe to their 5-Day Program! Just visit https://prolonlife.com/JENNIFERCOHEN and use code JENNIFERCOHEN to claim your discount and your bonus gift. Amp fits is the perfect balance of tech and training, designed for people who do it all and still want to feel strong doing it. Check it out at joinamp.com/jen    Find more from Jen:  Website: https://www.jennifercohen.com/ Instagram: @therealjencohen   Books: https://www.jennifercohen.com/books Speaking: https://www.jennifercohen.com/speaking-engagement Find more from Kim Perell:  Instagram: @kimperell Website: https://kimperell.com/