Podcasts about Kling

BMW spielt in Kooperation mit Sony Pictures in über 70 Ländern einen Animationsclip zum Spider-Man-Filmstart auf dem Bildschirm im Auto aus, freigeschaltet per Banner-Klick in bestimmten Modellen. BMW nennt es eine besondere Überraschung und ausdrücklich keine Werbung, in Foren beschweren sich Fahrerinnen und Fahrer trotzdem. BMW-Manager Durach hatte das Auto 2023 noch als letzten privaten Rückzugsort bezeichnet. Ich ordne ein, was das für Werbung im Auto und für das Vertrauen der eigenen Community bedeutet.Außerdem in den Marken- und Marketingnews der KW 33:⚽ DFB-Pokal Rebranding 2026: Kurz vor der ersten Hauptrunde am 21. August 2026 baut der DFB die Pokalmarke mit Strichpunkt und Manera um. Der Wettbewerb wird künftig über die gesamte Saison erzählt statt nur über den Finaltag, digital in kräftigem Violett, in den Stadien weiter in Pokalgrün.

Hey! Andy's back! He's joined by Ant and Dun and they have the most QPR of QPR cup exits to discuss, and predict the Pompey away game- Millwall. A tough draw ends with a tough draw- A first half of zero consequence- A livelier second half with two early season goals- Mbengue fails to clear but...- Charles fails to Kling on- Own goal to the rescue again, new set piece coach earns his corn- Saito tripped in the box, no? - Not long to wait for penalties though- Dun's prophesy comes true. QPR do not slot from the spot- Chair, Saito, Vale and Poku all see their efforts saved by Behemoth Jensen.- 2-0 is the most dangerous scorelines in penalty shoot-outs- QPR can concentrate on the league, before the league has even started- The positives, and the negatives (Pierce Charles may appear in both)- Highly rated but oft-injured left back Dennis Cirkin signs on a free- Tylon Smith, part of our youth revolution, off on loan to New England Revolution?- Where do we finish this season? A hat-trick of 15th's?- This week in Stockholming with a bit of New Yorking.- PSA, Don't eat the lettuce or the Guac- Ant's Kit Korner- Predictions for Pompey- Lovely Stuff - Sweden, Ultraworld Pod and Pets. (With love to the ones we lost, all the time we spent with them is the loveliest stuff)- The return of Jacob's Stanza's Rate, review, comment, stream, follow, download etc, etcSee you at the Factory on Saturday at 10am for the big kick off season opening party!

War früher alles besser. Oder fanden wir uns einfach nur besser? Zum Beispiel jünger, fitter, unbedarfter. Das zumindest ist die These des Kängurus in der "Känguru Rebellion" von Marc-Uwe Kling. Die Philosophin Dr. Rita Molzberger und die Journalistin Nora Hespers fragen sich, vor diesem Hintergrund, wie wir Kindheiten erleben und vergleichen. Warum wir manches früher besser fanden - und anderes aber auch nicht. Warum Kinder Erwachsene bemitleiden und was das daüber aussagt, wie unsere Welt eingerichtet ist. Und was ist eigentlich mit einem ordentlichen Onboarding fürs Erwachsenenleben?Ritas Literaturliste:El-Mafaalani, Aladin (2025): „Superdiverse Kindheiten“. In: El-Mafaalani, Aladin/ Kurtenbach, Sebastian/ Strohmeier, Klaus Peter: Kinder. Minderheit ohne Schutz. Köln: Kiepenheuer & Witsch, S. 65–86. Hasse, Jürgen; Schreiber, Verena (Hrsg.) (2019): Räume der Kindheit. Ein Glossar. Bielefeld: transcript. Hedderich, Ingeborg; Reppin, Jeanne; Butschi, Corinne (Hrsg.) (2021): Perspektiven auf Vielfalt in der frühen Kindheit. Mit Kindern Diversität erforschen. Bad Heilbrunn: Verlag Julius Klinkhardt. Kling, Marc-Uwe (2026): Die Känguru-Rebellion. Berlin: ullstein.LVR (Hrsg.) (2025): Superdiversität: Aufwachsen und teilhaben in Vielfalt. Jugendhilfereport 3/2025. Online abrufbar unter https://www.lvr.de/media/wwwlvrde/jugend/service/publikationen/dokumente_97/25.03_JHR_WEB.pdf (Datum des letzten Abrufs: 30.07.2026)Vertovec, Steven (2024): Superdiversität. Migration und soziale Komplexität. Berlin: Suhrkamp.Viernickel, Susanne/ Fuchs-Rechlin, Kirsten (2025): Grundlagen der frühkindlichen Bildung. Bad Heilbrunn: Verlag Julius Klinkhardt. "Arm sollst du bleiben" - Dokumentation von Monitor (WDR) in der ARD Mediathek (Veröffentlichungsdatum: 18.06.2026, Datum des letzten Abrufs: 01.08.2026)Noras Linkliste:Schönborn, Lea (25.04.2025): Kinder und Bildung - Was, wenn vieles falsch wäre, was wir über Kindheiten wissen? Krautreporter.de (Datum des letzten Abrufs: 01.08.2026) Hosted on Acast. See acast.com/privacy for more information.

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

If someone handed your competitor every prompt, workflow and AI system you use today, would you still have an advantage?   That is the question that opens this live roundtable, recorded in a sealed room at the first Founder AI Studio Live in Toronto. Ten women who are not just using AI but building with it: intellectual property strategists, AI educators, investors, community builders, and operators who have built and exited companies.   No keynote. No panel. Just founders comparing notes from the front line.   IP lawyer Andrea Bolden takes on the part most founders get wrong: whether your prompts, workflows and custom GPTs count as intellectual property, why switching off the training toggle is not the protection you think it is, why people trademark a name while giving away the entire body of work underneath it, and which AI tools will actually indemnify you if you get sued.   The room also gets into what happens to knowledge businesses when knowledge stops being scarce, why AI-native builds beat AI retrofits, the environmental cost nobody wants to raise, the risk of putting your child's face into these systems, and the six companies capturing most of the value from all of it.   It ends with a lightning round on the one tool each founder actually uses.   CHAPTERS 00:00 Context: what this room was 01:35 Not a keynote, not a panel 02:16 The question: would you still have an advantage? 03:09 Forty-five products that all do the same thing 05:46 Being the tastemaker 07:54 AI as a democratizing force 10:03 AI-native builds vs. AI retrofits 11:00 The contractor line item that changed everything 14:24 From "done for you" to "done for you-ish" 16:01 Building instead of talking about building 18:35 The real unlock is the speed of learning 23:00 If you sell knowledge, what are you selling now? 24:42 Dyslexia, ADHD and AI as a translation layer 28:22 Are your prompts and GPTs your IP? 29:28 The training toggle myth 30:30 Trademarking the name, losing the body of work 32:15 Tim Ferriss and what is already public 34:08 Can AI legally be an author? 35:32 Which AI tools will indemnify you 37:03 Writing a book with AI 39:27 Just because AI can, should it? 43:14 Share the why, not the whole how 44:00 The one thing to do next week 45:22 Canada's AI strategy and where the money is going 47:32 Environmental impact and who absorbs it 49:55 Your child's face inside an LLM 50:52 Sycophancy and outsourced thinking 51:58 What work do we reserve for humans? 54:16 Six companies and who holds the stake 56:14 Lightning round: the one tool 1:01:27 The advantage was never the tool   TOOLS MENTIONED Claude / Claude Code / Claude Cowork, ChatGPT, Microsoft Copilot, Lovable, Gamma, Codex, DeepSeek, Kling, Go High Level   LINKS Next Founder AI Studio Live   IN THE ROOM Host: Monique Bryan, Brand Authority Strategist Featuring: Andrea Bolden, IP and business lawyer   PARTNERS Captured by Perspective Studio Productions Presented in partnership with BDC Capital, Inclusive Entrepreneurship   Founder AI Studio Live is an invite-only working session for established women founders already building with AI. The room is the asset.   #AIforFounders #IntellectualProperty #WomenInAI Who Knows You is hosted by Monique Bryan, brand authority strategist and built for founders, operators, and experts who are doing real work and ready to be picked for it. Take the AI Visibility Audit to find out where your positioning is breaking down and what to fix: [RUN YOUR AUDIT] Connect with Monique:Before we build, let us talk. https://moniquebryan.com/book/ - Website: moniquebryan.com LinkedIn: Monique Bryan Instagram: @moniquebryan

In this episode of Film3 OG, Jordan sits down with British AI artist, filmmaker, music producer, and DJ Diane Laidlaw, known as Afro Futcha, for an inspiring and deeply personal conversation about creativity, representation, music, grief, and the power of finally allowing yourself to be seen.Diane shares how discovering Midjourney during a difficult period in her life helped reconnect her with imagination, purpose, and a creative identity she had been suppressing for years. What began as experimentation with images became filmmaking, music videos, international recognition, and an entirely new chapter in her life and career.They talk about Diane's journey from web design and music production into AI filmmaking; the personal loss behind her award-winning video What Do I Know; and the extraordinary year that followed, including winning Project Odyssey, speaking at TEDAI in Vienna, appearing on the BBC, and taking her work to audiences around the world.At the heart of the conversation is the importance of representation: women seeing women create, Black artists reclaiming and telling their own stories, and creators from historically excluded communities gaining access to tools that allow them to make visible what has always existed inside their imaginations.Diane also reflects on the difference between being a creator and becoming a content machine, why she protects her artistic freedom from the demands of social media, and how AI has allowed her to transform lived experience, emotion, memory, philosophy, and cultural heritage into work that makes people feel.The conversation moves through music, analog recording, Suno, Midjourney, ElevenLabs, Kling, Seedance, PixVerse, creative workflows, and the impossibility of mastering tools that change almost every week. Diane's advice to emerging artists is simple and powerful: take off the pressure, begin small, and play.This is a conversation about creative rebirth. About the courage to put your work into the world. About reaching back to help others see what may also be possible for them.And above all, it asks: How can AI help us manifest the stories, sounds, dreams, and ideas that have lived inside us all along?If you care about the future of cinema, music, representation, human imagination, and the evolving relationship between artist and machine, I think you'll find a great deal in this one.AFRO FUTCHA: https://x.com/misslaidlawClick SUBSCRIBE so you never miss an episode.⁠⁠⁠Film3™ ⁠⁠⁠ presents the Film3 OG Podcast, brought to you by ⁠⁠⁠⁠⁠⁠⁠⁠The Squad⁠⁠⁠⁠⁠⁠⁠⁠.-------------------------------About Film3™ OGFilm3™ OG is the flagship podcast of Film3™, the flagship brand of The Squad, founded by Jordan Bayne. Host Jordan Bayne sits down with the founders, filmmakers, artists, builders, and investors shaping creator-led entertainment. Episodes explore GenAI, IP ownership, decentralized financing, marketing, and the future of distribution — the next wave of cinema.Film3™Website: ⁠⁠https://film3.io⁠⁠The Squad: ⁠⁠https://filmsquad.io⁠⁠The Squad on X: ⁠⁠https://twitter.com/Film3Squad⁠⁠The Squad on Instagram: ⁠⁠https://www.instagram.com/film3squad⁠⁠The Squad on YouTube: ⁠⁠https://www.youtube.com/channel/UCF-bQ4_cNRth0XE-C60wzvw⁠⁠The Squad on Discord: ⁠⁠https://t.co/tvUph8QCZm⁠⁠Jordan BayneWebsite: ⁠⁠https://www.jordanbayne.com⁠⁠X: ⁠⁠https://twitter.com/jordanbayne⁠⁠Instagram: ⁠⁠https://www.instagram.com/jordan_bayne⁠⁠LinkedIn: ⁠⁠https://www.linkedin.com/in/jordan-bayne⁠⁠Substack: ⁠⁠https://jordanbayne.substack.com⁠⁠Film3™ is a trademark of Jordan Bayne and the flagship brand of The Film Squad LLC, operating publicly as The Squad. Film3™ OG is a program of The Film Squad LLC under the Film3™ brand. All rights reserved. © 2026 The Film Squad LLC.

AGENDA: 05:00 Washington Just Put Frontier AI on a Leash 06:30 Sam Altman's Wild 5% Government Stake Idea 19:00 The AI Funding Bubble: Why Founders No Longer Fear Dilution 28:00 Alex Karp's Brutal Warning: Enterprises Don't Trust Frontier AI 33:00 Meta's Shock Pivot: Has Zuck Accidentally Built the Next CoreWeave? 41:00 Nvidia's Dangerous New Game: "Compute Now, Pay Later" 45:00 Anthropic & DeepSeek Go After Nvidia's Crown 48:00 Kling vs Sora: Did China Just Win AI Video? 52:00 Is China Secretly Winning the Open Source AI War? 01:02:00 Microsoft & Amazon's $6B Bet: AI Still Needs Humans 01:11:00 Ashton Kutcher Walks Away From Sound Ventures 01:16:00 The New Startup Talent War: No Liquidity, No Chance 01:20:00 Final Thoughts: Who Wins the AI Endgame?

AI can feel overwhelming, overhyped, and impossible to keep up with — especially when every conversation seems to swing between "this changes everything" and "this is all just noise." In this episode of 20/20 Money, I'm joined once again by Dr. Mick Kling for a practical conversation about where AI is actually useful for practice owners right now — without the doom-and-gloom, hype, or technical jargon.   Mick shares how he's been thoughtfully experimenting with AI inside his own workflows to improve efficiency, analyze practice financials, model business decisions, and simplify strategic planning. More importantly, we discuss how practice owners can start using these tools in approachable, low-risk ways without needing to become programmers, coders, or "tech people." We also discuss: Why AI should be viewed more as an assistant than a replacement The biggest mistakes owners make when using AI tools How Mick built AI-powered financial analysis tools without coding experience The "LENS" framework for creating better prompts and better outputs Practical ways to use AI for business planning, associate hiring decisions, and cash flow analysis Why dashboards and reporting may evolve dramatically over the next few years The importance of human connection in a profession increasingly influenced by technology How to begin experimenting with AI without getting completely overwhelmed by it As always, we close the conversation with practical NBSs (Next Best Steps) for listeners who want to begin thoughtfully integrating AI into their practice and business workflows.     Have a podcast-related question? Contact our team here!   Resources: Previous episode with Mick about adding an associate OD in the practice (I think we did one on this topic??) Check out Wispr Flow (voice dictation software) Book a Triage call with Adam Download the Practice Owner's Financial Toolkit 20/20 Money Ultimate Financial Success Masterclass OD Mastermind Interest Form Check out Adam's new book: How to Buy an Optometry Practice   ————————————————————————————— Please rate and subscribe to 20/20 Money on these platforms Apple Podcasts Spotify ————————————————————————————— For past episodes of 20/20 Money with full companion show notes, please check out our episode archive here!   Check out Adam's other podcast!   The Optometry Success Podcast  Subscribe on Apple Podcasts: https://bit.ly/4tttng6 Subscribe on Spotify: https://bit.ly/4tuf0YM 

In this episode of the "Origins" series, Stephen L. Kling, Jr. joins host Sean Rost to talk about his research into the American Revolutionary War in the West, the exhibition Gálvez and Louisiana in the American Revolution at the Louisiana State Cabildo Museum in New Orleans, and his October 2026 presentation for SHSMO's History on Elm series. Episode Image: Drawing of Laclede's House [Gloria Simpson Collection (S1074), SHSMO] About the Guest: Stephen L. Kling, Jr. is an independent researcher focusing on the western theater of the American Revolutionary War. His books include The Battle of St. Louis, the Attack on Cahokia, and the American Revolutionary War in the West; Cavalry in the Wilderness: Cavalry in the Western Theater of the American Revolutionary War and the French and Indian War; James Colbert and His Chickasaw Legacy; The American Revolutionary War in the West; and An Underappreciated Victory: Bernardo de Gálvez's Mississippi River Campaign Against the British in 1779. He was the primary historical consultant for the 2021 Mid-American Emmy award-winning House of Thunder documentary on the Battle of St. Louis. He is the past curator of the American Revolutionary War in the West museum exhibition at the St. Charles, Missouri Heritage Museum. Currently, he is the curator for the newly opened Gálvez and Louisiana in the American Revolution exhibition at the Louisiana State Cabildo Museum in New Orleans based on his latest book. He will be the featured speaker at the State Historical Society of Missouri's History on Elm series on Tuesday, October 13, 2026.  Kling is a frequent speaker on the western theater of the American Revolutionary War being featured on C-SPAN, The American Revolution Institute in Washington, D. C. lecture series, as well as making presentations at the Missouri History Museum, the George Rogers Clark National Historical Park, the Friends of the Cabildo, and at numerous history conference and historical society meetings. Mr. Kling is a descendant of Joseph Robidoux III who was a soldier in the Spanish militia that defended St. Louis when it was attacked by British-led forces in 1780 during the American Revolutionary War.

In this episode of the Free Radical Podcast, we are joined by Dr. Sheri Kling—Christian author, theologian, artist, musician, public speaker, and spiritual guide—for a deeply penetrating exploration of the intersections of faith, theology, psychology, science, and the arts. Drawing from her work as the author of A Processed Spirituality and her passion for helping others move from disconnection and doubt into deeper purpose, passion, and spiritual vitality, Dr. Kling offers profound insights into some of life's most challenging questions. Through her writing, teaching, speaking, and retreat ministry, she helps others engage transformative pathways toward a more vibrant and meaningful faith. Together, we explore the nature of faith and belief, the reality of crisis, evil and suffering, and what it means to understand these experiences in the context of an all-alluring God. The conversation also touches on music, creativity, spiritual transformation, and the possibility of finding deeper meaning amid life's uncertainties. This conversation is an abundant feast of wisdom, reflection, and hope—inviting listeners into a deeper and more transformative vision of spirituality. Join us for a conversation that will challenge, inspire, and nourish both heart and mind.

Join our mastermind community: https://www.skool.com/apparel-success-mastermindTry the best Ai design platform: https://www.design.com/rob88AI video generators are getting insanely realistic, and in this video I show how clothing brand owners can use tools like Kling AI, Seedance, Google Veo, Sora, Runway, Adobe Firefly, Poyo.ai, Higgsfield AI, ChatGPT, ElevenLabs, CapCut, and Premiere Pro to create realistic ads, TikTok videos, Instagram Reels, product videos, and social media content without spending thousands on photoshoots, models, or videographers.I tested the biggest AI video tools to see which ones worked best for clothing brands using real product reference images. I break down why Google Veo, Sora, and Runway struggled, why Kling AI created some of the most realistic results, and why Seedance 2.0 might be one of the best AI video generators for accurate clothing brand content, fabric, logos, product details, and lifestyle scenes.If you run a streetwear brand, gymwear brand, hoodie brand, outdoor brand, or apparel business, this video shows how AI can help you create better ads, Meta ads, TikTok ads, Instagram content, UGC-style videos, and product marketing faster than ever.

Another week, another episode of FIFA World Cup madness! This week, Tobin and Christen start by taking a detour from the World Cup to talk about what's going on at Angel City: Kennedy Fuller AND Coach Alex Straus are gone, so what's next? Then, friend of the pod and official USMNT Expert™ Meghan Klingenberg joins to talk about how the US team is faring and what is possible for them this year (winning?!?) Plus, another edition of Tobin and Christen's travel adventures in Vancouver. Where in the world will they go next? New episodes every week. Watch the video version of the show on YouTube. Sign up for our newsletter, The RE—SET:  https://re-website.com/pages/newsletter Follow RE: https://www.instagram.com/re__inc/ https://www.tiktok.com/@re__inc https://twitter.com/re__inc https://www.threads.net/@re__inc   Follow Tobin: https://www.instagram.com/tobinheath https://twitter.com/TobinHeath   Follow Christen: https://www.instagram.com/christenpress https://twitter.com/ChristenPress To learn more about listener data and our privacy practices visit: https://www.audacyinc.com/privacy-policy Learn more about your ad choices. Visit https://podcastchoices.com/adchoices

Worf finds out that being Klingon nobility isn't all it's cracked up to be as we review "Sons of Mogh"! All that Kling-ron backstory comes back to haunt our newest Deep Sheetz Nine officer as his brother Kurn shows up with a dire request. We dig deep on Worf realizing how human he has become, the uncounted consequences of his actions, and just how many serious felonies Sisko is going to let slide on this station. 

Last week we talked about how many denominations are reporting that there are hardly any young leaders stepping into congregational ministry. We're talking with pastors Goodie Bell and David Kling about the Pastoral Leadership Crisis in the U.S. and what a group of Presbyterians are doing about it.

Enterprise XR hasn't disappeared, it has quietly moved into places where it saves time, reduces errors and changes how people work every day. On this episode of the AI XR Podcast, Charlie Fink and Rony Abovitz talk with Boston Consulting Group partner Kristi Woolsey, who leads BCG's immersive practice, about how XR plus AI is already being used for training, maintenance, onboarding, retail and architecture inside some of the world's most conservative organizations.Kristi shares a Swiss Rail project where field technicians wear lightweight AR glasses that recognize who they are and which train car they are standing in front of, pull the correct procedures from internal systems and use AI to turn thick manuals into simple task checklists.She explains how this leads to double-digit efficiency gains for both experienced and new workers, and how a small behavior design choice – automatic logging for headset users versus manual end-of-shift paperwork for everyone else – helped overcome skepticism on the front line. Drawing on her background as a physical-space architect, she also describes how VR and rapidly improving 3D tools are changing the way companies design stores, offices and buildings before anything physical is built.AI XR News you should know, Charlie and Rony cover Anthropic's massive new funding round and ethics turbulence, Chinese generative video tools like Seed Dance 2 and Kling that put TV-quality visuals in reach of “garage Spielbergs,” and Meta's reported seven million Ray-Ban and Oakley AI smart glasses sold – early signals of where wearable AI and XR are really headed.Key Moments01:03 – Anthropic's huge raise and what the ethics departure might signal05:08 – Seed Dance 2 and Kling showcase a new level of generative video08:35 – Meta's seven million smart glasses and the reality behind that number12:10 – Why wearable AI may be the real “last mile” of turning us into cyborgs15:28 – Inside the early metaverse tours Kristi and Rony built for enterprises20:27 – How BCG's VR onboarding keeps new hires engaged months before day one23:30 – Swiss Rail's AR and AI maintenance assistant and what it actually does on site27:05 – Designing XR systems that give value to both the business and frontline workers30:29 – Using VR as a lab for retail and workplace behavioral strategy33:06 – How AI-generated 3D models point toward “build every space digitally first”This episode shows how “metaverse” ideas have turned into practical tools: XR plus AI is cutting training times, improving maintenance quality and letting companies experiment with spaces before they exist. Kristi's examples make it clear that the real action is in careful workflow design, not flashy avatars.This episode is brought to you by Zappar, creators of Mattercraft, the leading visual development environment for building immersive 3D web experiences for mobile, headsets and desktop. https://mattercraft.io/Mattercraft combines the power of a game engine with the flexibility of the web and now includes an AI assistant that helps you design, code and debug in real time, right in your browser. To explore what's possible with AI-powered XR on the web, start building smarter with Mattercraft from Zappar.Listen to “Enterprise XR Meets AI: How Smart Glasses, Digital Twins and Holodecks Are Quietly Changing Work – Kristi Woolsey” on the AI XR Podcast and follow the show for new episodes every week. Hosted on Acast. See acast.com/privacy for more information.

Lucas Rizzotto is one of the most distinctive artists working at the intersection of technology and human experience. He built Where Thoughts Go, a VR piece that proved genuine connection was possible inside a headset when everyone said it wasn't. He followed it with Pillow, a mixed reality app designed around the bedroom. He then spent months letting an AI algorithm run his life — wearing Mantra smart glasses, building a surveillance and memory system on himself, and documenting it as an ongoing series on Instagram and TikTok. Now he's making a live cinematic experience called Escape the Internet, which he calls Broadway crossed with a video game crossed with standup comedy. It premiered as a ghost debut at SXSW this year.Mike Boland, analyst and founder of AR Insider, sits in for Rony Abovitz in this episode. The conversation opens on the Rec Room shutdown — $250 million raised, a $3.5 billion valuation, and now a wind-down. The panel connects the collapse to a pattern: VR has always been an exotic pursuit sold as a mainstream one, and the unit economics of concurrent immersive social spaces are nearly impossible. The discussion moves to OpenAI shutting down Sora, the AI video generation race between Google VO3 and Kling, the rise of AI slop in social feeds, and Lucas confirming he quit LinkedIn because it's unreadable.AI XR News: Rec Room is shutting down after raising $250M at a $3.5B peak valuation. Snapchat is acquiring its remaining assets. OpenAI closed down Sora, overwhelmed by competition from Google VO3 and Kling. AI-only social feeds from Meta and Grok are not gaining traction — users are tuning them out.Key Moments:[05:37] – Ted's thesis: VR is an exotic pursuit that was never going to be mainstream, and Rec Room would have been healthier if it accepted that early[07:33] – Lucas: Ready Player One was the worst thing to happen to XR — it gave executives a fictional roadmap to fund[18:38] – Ted asks whether Apple can do for mixed reality what it did for the smartphone — and the panel is skeptical[27:42] – Mike on physics as the hard ceiling: Moore's Law doesn't apply to waveguides and optics the way it applies to chips[29:02] – Lucas explains why he dropped display glasses for his wearable AI experiment — they increase engineering complexity by 50x[32:17] – Lucas's AI-controlled life series: a complex algorithm watches him, mines personal data, and tells him what to do to find happiness — including an unplanned trip to Lithuania[34:12] – Ted asks if the experiment is a net positive or negative. Lucas: neutral if you're in control, net negative if Meta or OpenAI are running the system[37:52] – Lucas on convenience as a death by a thousand cuts: he optimized his life in Berlin to have everything within three minutes and became miserable[41:00] – Charlie on Where Thoughts Go: assigned it to students every semester; it only works if you surrender to it[47:15] – Escape the Internet: hundreds of people in a movie theater, all on their phones, playing a shared cinematic narrative. Lucas calls it a modern version of church[53:40] – The standup model applied to software: Lucas tested Escape the Internet at SXSW and cut 50% of the material that didn't get a reactionThis conversation sits at the intersection that the AI XR Podcast lives for: technology as creative material, not just commercial tool. Lucas's view that we've been building things people use all the time when we should be building things that blow their minds for two hours and then get out of the way is one of the sharper critiques of the attention economy you'll hear this year.This episode is brought to you by Zappar and Mattercraft — the leading visual development environment for building immersive 3D web experiences on mobile, headsets, and desktop. Mattercraft now includes an AI assistant that helps you design, code, and debug in real time, right in your browser. Start building at mattercraft.io.Subscribe to the AI XR Podcast so you never miss a conversation. Hosted on Acast. See acast.com/privacy for more information.

Subscribe to DTC Newsletter - https://dtcnews.link/signupBraydon from Pilothouse joins for a fast AI check-in on what the team is actually shipping right now. Not theory. The exact tools, prompts, and workflows behind their current creative output.If you run growth or creative at a DTC brand or agency, this is a look at how one team is collapsing production time on landing pages, video ads, and founder content using AI.The sub-agent "council" prompt: rewrite a landing page eight ways, with each sub-agent playing a role (copywriter, CEO, customer, CRO expert), then have the council rate the versions and a final decision maker pick the winner.How to keep Claude fast on long projects: ask the chat to summarize itself into a markdown file, then carry that into a fresh chat instead of letting one thread balloon.Higgsfield as a model aggregator: one place to run VO3, Kling, ElevenLabs, and image models, with one-click image-to-video and multiple aspect ratios.The founder avatar workflow: build a 30-second explainer with B-roll and slow zooms, clone the founder's voice in ElevenLabs, and ship it the same afternoon.Why Braydon leans into obviously-AI creative (claymation, Pixar-style) instead of trying to pass synthetic people as real.How Meta's Andromeda rewards ads that improve the scroll, and why social boosting organic winners is finding new scale.Who this is for:DTC operators, growth marketers, and agency creative leads who want a current, practical AI workflow rather than a hype reel.What to steal:The sub-agent council prompt, the Higgsfield image-to-video and voice-clone pipeline, and the social boosting approach to finding ad winners.Timestamps:00:29 Fable AI and One-Shot Development06:54 Higgsfield for AI Creative Production10:33 New AI Advertising Disclosure Rules13:15 AI Search, SEO, and Answer Engines14:33 How Andromeda Rewards Better Ad ExperiencesSubscribe to DTC Newsletter - https://dtcnews.link/signupAdvertise on DTC - https://dtcnews.link/advertiseWork with Pilothouse - https://www.pilothouse.co/?utm_source=AKNF621Follow us on Instagram & Twitter - @dtcnewsletterWatch this interview on YouTube - https://dtcnews.link/video

In this exciting episode of the Brave New Bookshelf, hosts Steph Pajonas and Danica Favorite welcome back author, designer, and AI filmmaker Novae Caelum to discuss the evolution of AI-empowered storytelling. Novae shares behind-the-scenes insights into his creative process, demonstrating how he uses a patchwork of cutting-edge video and audio tools like Kling, Suno, and Seedance to direct stunning, cinematic book trailers and episodic shows that bring novels to life. The conversation also dives into how authors can leverage AI to conquer the blank page through rapid "zero drafting," while addressing industry pushback and highlighting the vital role of human artistry in steering these technologies. Whether you are curious about AI filmmaking or looking to streamline your own writing and marketing workflows through Novae's Substack, AI Marketing for Storytellers, this episode is packed with inspiration and practical advice for the modern creator. Visit our website https://bravenewbookshelf.com to view the full episode notes, links and apps mentioned in the episode, and the full transcript.

Í Hvelfingu Norræna hússins stendur yfir sýningin Himna; Goðsagnir & skáldskapur úr jörðu. Sýningin er samsýning átta listamanna sem koma víðsvegar að en verk þeirra skoða endurnýjun jarðvegs sem grundvallarferli sem gerir jörðina byggilega og lífi kleift að kvikna, dafna og endurnýja sig. Í þætti dagsins gerum við okkur ferð í Norræan húsið og tökum sýningarstjórann Thomas Pausz tali um Himnu. Haukur Þorgeirsson, rannsóknarprófessor, segir okkur frá skáldunum sem sömdu Íslendingasögurnar og endar á Njáls sögu. Var höfundurinn einhver frændi Snorra Sturlusonar - eða kannski frænka hans? Gallerí Kichin var stofnað af þeim Skúla Thayer og Silju Rún Högnadóttur en upphafi miðaðist galleríið við eina óræða mublu sem þau fundu í Góða hirðinum. Mublan fékk heimili í vinnustofu þeirra í Listaháskólanum og með tímanum umbreyttist hún í sýningarrými. Í fyrra fór Kichin svo í samstarf við listahátíð í Reykjavík og úr urðu sex sýningar sem hófu göngu sína við upphaf listahátiðar og opna með viku millibili fram í júlí. Í þættinum lítum við inn í Kling og Bang og fáum að heyra af verkefninu.

Part 2 - Host Neville James transitions to a discussion with Catherine Kling, General Manager of Liberty VI, who outlines the company's ongoing improvements, including expanded fiber internet services, new hiring initiatives, and infrastructure projects such as additional cell sites and a planned retail location in St. John. Kling also addresses customer concerns about service issues when traveling, explaining technical limitations with U.S. Virgin Islands area codes and emphasizing Liberty's efforts to advocate for better recognition and connectivity.

The Worst It'll Ever Be: AI Apps in 20 Minutes, SpaceX's $1.8T IPO & Saylor's Head Fake — Bad Crypto Podcast #810 It's a bear market, so the bad boys of crypto are doing what builders do: SHIPPING. Bitcoin sits at $61,873, the altcoins are in the crapper, and Joel has officially divorced his bags. Travis explains why the 4-year cycle is alive and well — mapping this pullback exactly to previous cycles, with a projected bottom around mid-October. Then it goes full mad-scientist. Travis builds a viral-worthy "Culture Shock" site of World Cup visitors reviewing America in 20 minutes flat with Claude's new Fable model, then ships Viddl — a desktop app that downloads video from YouTube, X, TikTok, Instagram or LinkedIn with FFmpeg baked in. Joel premieres his AI-generated origin story film (1978, a food court paycheck, and a TRS-80 in a Radio Shack window) and announces his Acumen daily puzzle games are headed to the App Store. Plus: SpaceX IPOs as $SPCX at a $1.8 TRILLION valuation with ~$250B in demand, OpenAI and Anthropic file to go public, Michael Saylor's 32-BTC head fake, a trader who built his own exchange from a 42-page prompt, and the AI video tool stack the guys actually use (Kling, PAI, Higgsfield, Seedance & more). "The technology that we're using now to build stuff is the worst that it's going to be." — Joel ⏱ CHAPTERS0:00 Cold open & liftoff1:04 Episode 810 kicks off — semi-retired no more3:48 Bitcoin's 4-year cycle is mapping exactly4:45 Saylor's head fake: sells 32 BTC, buys 1,500 more6:40 Market check: BTC $61,873 & Joel divorces his altcoins7:49 The AI trading edge: OKX & the 42-page prompt exchange10:24 SpaceX IPO ($SPCX): $250B demand, $1.8T valuation11:27 Trillion-dollar AI: Anthropic & OpenAI file to go public15:48 Culture Shock: World Cup visitors review America19:09 Viddl: download any video, built in a morning23:06 Joel's AI origin story: 1978 & a TRS-8026:30 The AI video stack: Kling, PAI, Higgsfield, Seedance28:08 Acumen: 9 daily puzzle games headed to the App Store31:56 Travis's Pixar-style get-well video for his brother35:03 "The worst it's ever going to be" — why the opportunity is NOW37:18 The fine print

Marcia Kling (aka Miss Marcia) helped raise many of today's leaders with her shows on WTVC: Romper Room School and Funtime. In this episode, Marcia shares how faith has guided her through every major decision of her life, how she went from working in churches in Chattanooga and New York to children's television host, and what it took to come back to the show she loved after being told she might never speak again. If you like this episode, we think you'll also like: Jed Mescon's Morning Cup (E66) Tom Henderson's Morning Cup (E82) Bob Culkeen's Morning Cup (E116) Subscribe to the weekly newsletter and be the first to know who upcoming guests are: http://eepurl.com/iGJzII  My Morning Cup is hosted by Mike Costa of Costa Media Advisors and produced by SpeakEasy Productions.

In Episode 70, Drew Brucker and Rory Flynn are joined by Tyler Bernabe, better known as jboogxcreative, a full-time generative AI creator, strategist, and social menace responsible for some of the wildest AI videos your algorithm has probably shoved into your face at 1:13 a.m.They get into how Tyler has gone viral across multiple generations of AI tools, why copying trends is creative quicksand, how shock value actually works when it is paired with taste, and why the best AI creators are building formats instead of chasing them.The conversation also goes deep into the unglamorous machinery behind creative internet magic: Instagram to Patreon funnels, ManyChat, six-hour livestreams, creator burnout, client work, taste, consistency, and why “just post more” is advice usually given by people who should post less.Then things get properly nerdy.Tyler breaks down his current AI creative stack, including Midjourney 8.1, Seedance, Claude, YAML-style video prompting, Nano Banana, GPT image editing, Kling, Artcraft, Venice AI, Magnific/Freepik Spaces, Weavy, Reeve 2.0, and the never-ending wait for a proper Midjourney editor.Along the way, they cover Chinese prompt translation for Seedance, 10,000-character prompt workflows, reference image construction, anime style development, mood board blending, why AI should sometimes pull you away from your own creative bias, and why the smallest edit in AI video still feels like defusing a tiny cursed bomb.⏱️ Fast Hour00:00 Fast Hours welcomes Tyler aka jboogxcreative01:51 Going viral through every AI era02:35 The anatomy of scroll-stopping AI03:07 The seductive food video origin story05:32 Running opposite the AI meta08:24 Is AI art? Tyler's best answer10:35 Why clients pay for your thing12:10 Stop copying other creators17:11 Viral views vs real conversion22:03 Building a creator business solo26:28 Burnout, longevity, and going through it33:05 Tyler's current AI tool stack38:15 Artcraft, Seedance, and Chinese prompts42:05 Claude, YAML, and 10K prompts50:44 Tyler's reference image hack58:15 Dream sketches and creative prototyping01:04:10 Reve 2.0 and layered editing01:10:38 Waiting for Midjourney's editor01:16:02 Closing thoughts#FastHours #jboogxcreative #AIVideo #AIArt #GenerativeAI#Midjourney #Seedance #ClaudeAI #AICreator #AIWorkflow #ContentCreation #AIPrompting #CreatorEconomy #AIAnimation#CreativeAI

The HBS co-hosts are diligently at work prepping for Season 16 so, in the meantime, enjoy this "Minibar" episode from Jennifer Kling explaining the merits and demerits of employing Hanlon's Razor in our everyday lives! Full episode notes available at this link:https://hotelbarpodcast.com/podcast/hanlons-razor---------------------SUBSCRIBE to the podcast now to automatically download new episodes!SUPPORT Hotel Bar Sessions podcast on Patreon here! (Or by contributing one-time donations here!)BOOKMARK the Hotel Bar Sessions website here for detailed show notes and reading lists, and contact any of our co-hosts here.Hotel Bar Sessions is also on Facebook, YouTube, BlueSky, Instagram, and TikTok. Like, follow, share, duet, whatever... just make sure your friends know about us! ★ Support this podcast on Patreon ★

When the Klingons are on the rise it's up to Kirk and team to thwart an upcoming invasion, but when the Organians are a little too chill, it's time to burst into a beam of light and put some Qo'nos and Earth perspectives aside. Will this renaissance world survive? How breezy are those outfits? Will the Kling'ns be back? It's the episode that introduces stage-one of a loafed evolution.Support the production of our shows Members get benefits including bonus episodes and an ad-free experienceSign up for our mailing listGet a thing at podshop.bizGreatest Trek is hosted by Adam Pranica and Benjamin Ahr Harrison The show is produced by Wynde PriddySocial media is managed by Rob Adler and Bill TilleyMusic by Adam RaguseaDiscuss the show using the hashtag #GreatestTrek and find us on social media:YouTube | Instagram | BlueskyAnd check out these online communities run by FODs: Reddit | USS Hood Discord | Facebook group | Wikia | FriendsOfDeSoto.socialSupport the production of Greatest Trek Hosted on Acast. See acast.com/privacy for more information.

Hoy hablamos de Meta vendiendo IA con consultoría onsite, del despegue comercial de Kling AI, de Amazon empaquetando la tecnología de Alexa para otros retailers, del Pentágono financiando fabricantes de drones y de una línea prometedora de compuestos basados en vitamina K para regenerar neuronas.Puedes seguirnos en YouTube en https://youtube.com/olivernabani y puedes unirte al Discord Mashain en https://olivernabani.com/discord

Drew and Rory are back for episode 69, which is legally required to begin with at least one immature joke before immediately collapsing under the weight of Google's latest AI product avalanche.This week, they dig into Google Omni, Gemini 3.5 Flash, Google Flow, Google Pics, Nano Banana, Veo, and whatever else Google launched before anyone had time to make coffee. The big question: are these actually meaningful creative upgrades, or did Google just throw 19 AI names into a blender and call it innovation?They break down early Omni and Flow tests, why video physics still feel weird, where Seedance and Kling may still be ahead, and why Runway Aleph 2.0 feels promising but imperfect. Rory shares hands-on examples with character swaps, driving videos, golf swings, agent mode, and Flow's new tool-building features. Drew tries to keep the conversation coherent while quietly wondering if every AI product now needs a map, glossary, and mild sedative.The episode also gets into Gemini as a search replacement, creepy context awareness, privacy tradeoffs, AI tools connecting to personal data, the fuzzy definition of “agentic,” the limits of auto-clipping tools, GPT Image 2's SynthID watermarking, metadata headaches for client work, and the universal pain of wasting $15 trying to make an image model spell “stump.”If you're trying to understand what Google's AI updates actually mean for creators, marketers, AI video workflows, image generation, creative direction, and the future of agentic media tools, this episode is half useful breakdown, half group therapy for people with too many tabs open.---⏱️ Fast Hour00:00 Cold open00:32 Google's AI naming avalanche01:39 AI hype vs actual workflow value02:34 Why AI launches feel like iPhone upgrades06:12 Google's “throw everything” strategy07:08 Omni vs Veo 4 expectations07:43 Video physics and speed problems09:03 Google Pics, Flow, Omni, and Flash10:04 How Rory actually uses Gemini11:51 Gemini 3.5 Flash breakdown12:38 AI benchmarks feel like marketing13:42 Gemini as a better search layer15:18 Creepy Gemini context awareness17:35 Why AI data connections feel too early19:15 The privacy tradeoff gets darker21:19 Google Omni vs Runway Aleph 2.022:12 Google Omni testing starts rough23:39 Google Veo 3.1 feels forgettable25:21 Why Omni feels early26:19 Higgsfield clipper test fails27:59 Why auto-clipping still misses31:30 Rory tests Flow and Omni live32:41 Omni character swap struggles33:33 Runway Aleph panda test34:07 Flow's new interface and tools35:02 Building custom tools inside Flow36:10 The joy of making tools from nothing37:39 Agent mode for still-image workflows39:05 Batch creative directions in Flow40:03 Omni turns six images into video40:47 Driving physics still feel off41:55 Why consistency matters for adoption43:03 Kling, Seedance, and the update race43:59 Seedance handles complex camera motion45:42 GPT Image setup for golf video46:53 Testing the same prompt in Flow49:25 Why agentic platforms can feel thin51:10 The need for visual design systems52:21 Flow's golf swing result53:56 Everyone is racing toward agentic54:18 What “agentic” actually means56:03 Claude feels more genuinely agentic57:04 Josh Hart quote analysis detour58:44 Reverse-engineering creative patterns59:53 Pizza, calzones, and prompt structure01:00:26 SynthID and GPT Image 2 watermarking01:01:47 Metadata problems for client work01:02:51 Google Pics enters the chat01:04:03 Too many image models to track01:04:52 Midjourney color still hits different01:06:01 GPT Image 2 quality frustration01:06:59 Image models still struggle with scale01:08:26 Bad AI weeks happen too01:09:20 Midjourney 8.2 speculation01:10:01 Tell your florist

Today's guest is Becky's former teammate and best friend forever, Meghan Klingenberg. The Portland Thorns legends reminisce on their playing days and get into all the details of "Kling goes to referee camp."See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

In this week's episode of Uncrewed Views, Matt Collins speaks with Matt Kling, VP and general manager of AI Systems at MatrixSpace. The two discuss how the Safer Skies Act has shaped demand for counter-UAS technology, why gaps still exist between what agencies need and what's being deployed, and how MatrixSpace is approaching the counter-UAS challenge as a radar-first, hardware-and-software hybrid company.

See omnystudio.com/listener for privacy information.

Episode 114 of Focus Check is our full NAB 2026 Best of Show breakdown — 11 categories, 12 winners — covering cameras, lenses, lights, monitors, intercoms, power, storage, and post-production tools. Plus non-NAB releases from DJI and ARRI.   This episode is sponsored by Nanlite. Check it out at 27:58   (00:00) Intro & overview   (03:31) Kinefinity Vista Teased at NAB 2026 – Sub-$3,000 Full-Frame 6K Open Gate in a Palm-Sized Body https://www.cined.com/kinefinity-vista-teased-at-nab-2026-sub-3000-full-frame-6k-open-gate-in-a-palm-sized-body/     (14:07) GoPro MISSION 1 Series at NAB 2026 – On the 1-Inch Sensor, GP3 Processor, and an MFT Mount https://www.cined.com/gopro-mission-1-series-at-nab-2026-on-the-1-inch-sensor-gp3-processor-and-an-mft-mount/     (28:59) Radiant Images Combines 24 iPhones with Gaussian Splatting for Next-Gen Bullet Time https://www.cined.com/radiant-images-combines-24-iphones-with-gaussian-splatting-for-next-gen-bullet-time/     (35:06) FUJINON GF 32-90mm T3.5 PZ OIS WR Lens Announced – Purpose-Built for GFX ETERNA 55 https://www.cined.com/fujinon-gf-32-90mm-t3-5-pz-ois-wr-lens-announced-purpose-built-for-gfx-eterna-55/     (39:19) DZOFILM Arcana Anamorphic Prime Lenses – Hands-On With the 1.5x Hybrid Trio https://www.cined.com/dzofilm-arcana-anamorphic-prime-lenses-hands-on-with-the-1-5x-hybrid-trio/     (40:47) NANLUX Evoke 5C Detailed – Pocket Point Source with IP67 and a New HSIW Mode https://www.cined.com/nanlux-evoke-5c-detailed-pocket-point-source-with-ip67-and-a-new-hsiw-mode/     (44:45) Atomos CEO Peter Barber on the Sumo PRO-19, Rebuilding the Brand, and Acquiring Flanders Scientific https://www.cined.com/atomos-ceo-peter-barber-on-the-sumo-pro-19-rebuilding-the-brand-and-acquiring-flanders-scientific/     (52:49) Hollyland SolidCom H1 – Enterprise Wireless Intercom https://www.cined.com/hollyland-pyro-ultra-wireless-transmission-system-launches-with-4k60-and-1-5km-range/     (55:50) Core SWX x Mondobytes PowerVault Introduced – Modular, Multi-Format Charging in One Case https://www.cined.com/core-swx-x-mondobytes-powervault-introduced-modular-multi-format-charging-in-one-case/     (59:27) iodyne Pro Mini Starts Shipping – Digital Label, Passkey Unlock, and RAID-6 in a Pocket SSD https://www.cined.com/iodyne-pro-mini-starts-shipping-digital-label-passkey-unlock-and-raid-6-in-a-pocket-ssd/     (01:05:15) XLCS Killdozer Cage – Weatherproof CNC Cage System   (01:06:37) Adobe Reinvents Color Grading in Premiere, Expands Firefly with Kling 3.0, and Debuts Frame.io Drive https://www.cined.com/adobe-reinvents-color-grading-in-premiere-expands-firefly-with-kling-3-0-and-debuts-frame-io-drive/     (01:09:51) Pixboom Spark Ships in May: 12-Bit RAW and ProRes RAW Coming Next https://www.cined.com/pixboom-spark-ships-in-may-12-bit-raw-and-prores-raw-coming-next/     (01:15:33) NiSi 50mm T1 Cine Prime Unveiled – LPL Mount and Large-Format Coverage https://www.cined.com/nisi-50mm-t1-cine-prime-unveiled-lpl-mount-and-large-format-coverage/     (01:20:47) OWC Express 4M2 Ultra Announced – DIY Thunderbolt 5 NVMe RAID up to 6,622 MB/s https://www.cined.com/owc-express-4m2-ultra-announced-diy-thunderbolt-5-nvme-raid-up-to-6622-mb-s/     (01:22:53) DJI Mic Mini 2 Released – 11g Transmitter, Three Voice Tone Presets, and Mic 3 Cross-Compatibility https://www.cined.com/dji-mic-mini-2-released-11g-transmitter-three-voice-tone-presets-and-mic-3-cross-compatibility/     (01:25:17) DJI Lito X1 and Lito 1 Launched – Aimed at First-Time Aerial Creators https://www.cined.com/dji-lito-x1-and-lito-1-launched-aimed-at-first-time-aerial-creators/     (01:27:18) DJI Power 1000 Mini Released – Compact 1kWh Power Station with Retractable USB-C and 58-Minute Recharge https://www.cined.com/dji-power-1000-mini-released-compact-1kwh-power-station-with-retractable-usb-c-and-58-minute-recharge/     (01:28:41) ARRI cforce MAX Replaces the cforce plus – Twice as Fast, 15% Smaller, and With an Onboard Touchscreen https://www.cined.com/arri-cforce-max-replaces-the-cforce-plus-twice-as-fast-15-smaller-and-with-an-onboard-touchscreen/   Have feedback on this episode? Email us at podcast@cined.com or leave a comment below.

Fast Hours has entered the witness protection program. Same Drew. Same Rory. But fewer syllables and more chaos.In this episode, Drew Brucker and Rory Flynn officially drop “Midjourney” from the podcast name and relaunch as Fast Hours, a broader home for the creative AI ecosystem: image models, video models, LLMs, vibe coding, Claude, ChatGPT, Midjourney, and whatever tool drops five minutes after they hit publish. Naturally, the rebrand lasts about four minutes before they're elbows-deep in GPT-Image-2, OpenAI's new ChatGPT image model that quietly showed up and immediately started making designers question their calendar, career choices, and relationship with kerning.The big topic: GPT-Image-2 is shockingly good with text, typography, brand systems, visual decks, product mockups, and multi-image outputs. Rory walks through how he used ChatGPT and Claude to create a custom typeface from visual references, generate a premium typography presentation, extract geometry, and turn the whole thing into usable font files. Drew then shows how he turned his own handwriting into a working typeface, because apparently “personal brand” now includes making your lowercase g file a tax asset.They also dig into the uncomfortable middle ground of AI creative work: when it saves time, when it still needs human judgment, why anti-AI panic and AI hype both miss the point, and why the real advantage is context. Not prompts. Not magic buttons. Context.The episode also covers GPT-Image-2 vs Nano Banana Pro, richer color rendering, micro-text improvements, AI-generated sports graphics, brand kit concepts, Freepik settings, Claude Design, 4K video generation, Kling, Veo 3.1, Seedance, and the strange reality that a custom brand typeface can now go from “that'll be $150K” to “Rory did it before lunch.”Basically, it's an episode about the exact moment creative production stops feeling like a tool demo and starts feeling li ke a factory someone accidentally left unlocked.---⏱️ Fast Hour00:00 Fast Hours is (re)born03:36 Going tool-agnostic04:34 GPT-Image-2 quietly drops05:31 Text becomes the unlock07:31 The AI backlash returns10:57 Hype, fear, and the middle12:10 Typography gets weird14:50 What custom fonts cost15:43 GPT-Image-2 vs Nano Banana17:39 Rory's font experiment18:47 Fiddleheads become a typeface19:39 Building the type deck20:36 The nine-slide image unlock21:14 Geometry, spacing, and logic22:11 Turning images into font files23:02 Micro-text gets better24:19 Claude builds the font package26:41 The revision loop changes27:50 Context is the silver bullet32:11 Drew makes a handwriting font35:35 Why designers obsess over type37:52 Reverse-engineering prompts39:51 Richer color and sports graphics41:27 Fixing artifacts and details42:37 Nano Banana vs GPT-Image-2 tests44:26 Sports realism gets scary good45:27 Why teams need this now46:43 Freepik settings and ratios48:36 Testing, tokens, and limits49:44 Brand kits and rebrand concepts53:19 Google I/O and the next model53:48 Veo 3.1 falls behind55:04 Kling adds native 4K56:40 Character sheets and macros58:07 Rebrands as visual prototypes01:00:53 Building a reference library01:01:36 Three weeks in a row01:02:58 Claude Design tease01:03:37 Tell your local [fill in the blank] spam finale

In this episode I sit down with my friend Sirio, one of the most creative AI minds I know, to break down Seedance V2. Sirio walks us through the exact use cases, prompts, and tactics he's using to build on top of this model inside his platform Enhancor, covering multi-input generation, virtual try-ons, ad translation, AI influencers with lip sync, video extension, and 3D product template replacement. I wanted this to go beyond the "look how cool this is" tutorials and focus on how creators and founders can actually build businesses, run ads, and produce creative assets with it. By the end, you'll have a practical playbook for Seedance V2 and a clear view of where it fits alongside other models like Kling 3, Veo, and fine-tuned options. Timestamp 00:00 – Intro 02:22 – Demo 1: Replacing Characters and Background in a Green Screen Scene 08:03 – Prompting Tactics and Optimize Prompts 09:45 – Demo 2: Virtual Try-On in Montreal (Minus 30 Degrees) 13:05 – Demo 3: Ad Translation and Character Replacement (Chinese to English) 16:02 – Demo 4: 3D Product Template with Brand Texture Swap 18:40 – Demo 5: Video Extension and Filling in the Middle 20:55 – Demo 6: AI Influencers and Prompting Realistic Emotion 29:31 – What Happens to Adobe Over the Next Five Years Key Points Seedance V2 is the first widely available video model to support true multi-input generation — up to two images, two videos, and an audio file combined in a single prompt. Treat Seedance V2 as a video editor, not just a generator: character swap, background swap, text preservation, ad translation, and template population all work from natural-language prompts. Seedance rewards highly specific prompts; I pair my own draft with Claude Opus 4.6 to optimize prompts for vision models. Strong source reference images remain the single biggest quality lever — the model mimics taste from what you feed it. For AI influencers and lip sync, describe muscle movements and emotional transitions rather than simply labeling an emotion like "sad" or "happy." Seedance V2 is the current default for editing and generating video, yet other models (Kling 3 for cinematic feel, Enhancer V4 for talking-head realism) still win on specific use cases. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND SIRIO ON SOCIAL Enhancor AI: https://www.enhancor.ai Instagram: https://www.instagram.com/heysirio/ Youtube: https://www.youtube.com/@SirioBerati

Rain fell down but the goals did not in Seattle for this round two matchup between the USWNT and Japan. Sam and Kling break down all the action including Captain Claire Hutton and the return of Tierna Davidson.Get your TWG merch here: https://mibcourage.co/4c7wOCBSUBSCRIBE TO THE WOMEN'S GAME NEWSLETTER: https://mibcourage.co/42X5HpBSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

See omnystudio.com/listener for privacy information.

Lucas Rizzotto is one of the most distinctive artists working at the intersection of technology and human experience. He built Where Thoughts Go, a VR piece that proved genuine connection was possible inside a headset when everyone said it wasn't. He followed it with Pillow, a mixed reality app designed around the bedroom. He then spent months letting an AI algorithm run his life — wearing Mantra smart glasses, building a surveillance and memory system on himself, and documenting it as an ongoing series on Instagram and TikTok. Now he's making a live cinematic experience called Escape the Internet, which he calls Broadway crossed with a video game crossed with standup comedy. It premiered as a ghost debut at South by Southwest this year.Mike Boland, analyst and founder of AR Insider, sits in for Rony Abovitz in this episode. The conversation opens on the Rec Room shutdown — $250 million raised, a $3.5 billion valuation, and now a wind-down. The panel connects the collapse to a pattern: VR has always been an exotic pursuit sold as a mainstream one, and the unit economics of concurrent immersive social spaces are nearly impossible. The discussion moves to OpenAI shutting down Sora, the AI video generation race between Google VO3 and Kling, the rise of AI slop in social feeds, and Lucas confirming he quit LinkedIn because it's unreadable.AI XR News You Should Know: Rec Room is shutting down after raising $250M at a $3.5B peak valuation. Snapchat is acquiring its remaining assets. OpenAI closed down Sora, overwhelmed by competition from Google VO3 and Kling. AI-only social feeds from Meta and Grok are not gaining traction — users are tuning them out.Key Moments:[05:37] – Ted's thesis: VR is an exotic pursuit that was never going to be mainstream, and Rec Room would have been healthier if it accepted that early[07:33] – Lucas: Ready Player One was the worst thing to happen to XR — it gave executives a fictional roadmap to fund[18:38] – Ted asks whether Apple can do for mixed reality what it did for the smartphone — and the panel is skeptical[27:42] – Mike on physics as the hard ceiling: Moore's Law doesn't apply to waveguides and optics the way it applies to chips[29:02] – Lucas explains why he dropped display glasses for his wearable AI experiment — they increase engineering complexity by 50x[32:17] – Lucas's AI-controlled life series: a complex algorithm watches him, mines personal data, and tells him what to do to find happiness — including an unplanned trip to Lithuania[34:12] – Ted asks if the experiment is a net positive or negative. Lucas: neutral if you're in control, net negative if Meta or OpenAI are running the system[37:52] – Lucas on convenience as a death by a thousand cuts: he optimized his life in Berlin to have everything within three minutes and became miserable[41:00] – Charlie on Where Thoughts Go: assigned it to students every semester; it only works if you surrender to it[47:15] – Escape the Internet: hundreds of people in a movie theater, all on their phones, playing a shared cinematic narrative. Lucas calls it a modern version of church[53:40] – The standup model applied to software: Lucas tested Escape the Internet at SXSW and cut 50% of the material that didn't get a reactionThis conversation sits at the intersection that the AI XR Podcast lives for: technology as creative material, not just commercial tool. Lucas's view that we've been building things people use all the time when we should be building things that blow their minds for two hours and then get out of the way is one of the sharper critiques of the attention economy you'll hear this year.This episode is brought to you by Zappar and Mattercraft — the leading visual development environment for building immersive 3D web experiences on mobile, headsets, and desktop. Mattercraft now includes an AI assistant that helps you design, code, and debug in real time, right in your browser. Start building at mattercraft.io. Subscribe to the AI XR Podcast so you never miss a conversation.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Today we're sharing a podcast we love over here at RE: The U.S. Soccer Podcast hosted by none other than the legendary Meghan Klingenberg. Each week Kling does fascinating deep-dive interviews with U.S. Soccer icons -- kinda like Fresh Air, but for soccer. Christen was a guest. Alyssa Naeher was on the show. She recently recorded with Kristine Lilly. And Tobin! She sat down with Kling and chatted about her career, their favorite memories of playing together, and how they've dealt with retirement. If you like this episode, just search up “U.S. Soccer Podcast” on YouTube or wherever you listen to podcasts. To learn more about listener data and our privacy practices visit: https://www.audacyinc.com/privacy-policy Learn more about your ad choices. Visit https://podcastchoices.com/adchoices

In this episode, Ray Cochrane digs into a new study showing AI is literally frying workers’ brains, then unpacks Anthropic’s wildest month ever – from a 1,487% user surge to Pentagon retaliation to a leaked model called Mythos. Also covered: OpenAI kills Sora after burning $15 million a day, OpenClaw’s terrifying security holes, Apple axing the Mac Pro, ARM’s first-ever production CPU, and why King Tut’s dagger was forged from a meteorite. – Want to start a podcast? It’s easy to get started! Sign-up at Blubrry – Thinking of buying a Starlink? Use my link to support the show. Subscribe to the Newsletter. Email Ray if you want to get in touch! Like and Follow Geek News Central’s Facebook Page. Support my Show Sponsor: Best Godaddy Promo Codes Get 1Password Full Summary Cochrane opens the show with a study that puts a name to something most AI-heavy workers have already felt. From there, the episode moves through one of the most turbulent months in AI industry history, touching on corporate ethics, national security, hardware shortages, and ancient archaeology. AI Use at Work Is Causing “Brain Fry” A study from Boston Consulting Group and UC Riverside surveyed 1,500 full-time US workers and found that 14% experience what researchers call “AI brain fry” – mental fatigue from excessive AI tool oversight. Those affected report 33% more decision fatigue, 39% more major errors, and an increase in intent to quit from 25% to 34%. Notably, productivity peaks at one to three AI tools and drops off at four or more. Cochrane relates this directly to his own workflow, often running two to four tools side by side. However, he pushes back on the doom framing. He argues that context switching across multiple projects and rubber-stamping AI output without review are the real sources of fry. His takeaway: either work more slowly with greater intent, or use the accelerated pace to reclaim free time. Anthropic’s Wild Month: Exodus, Pentagon, and Mythos Claude sessions surged by roughly 1,487% from mid-January to early March, knocking ChatGPT off the top spot in the app store for the first time. ChatGPT uninstalls spiked nearly 300%, one-star reviews exploded 775% in a single day, and a boycott movement called “Quit GPT” has grown to between 2.5 and 4 million participants. The catalyst was OpenAI stepping in to take the Pentagon defense deal that Anthropic had publicly declined. Cochrane is firmly against automated domestic surveillance and autonomous weaponry, noting that the models are not reliable enough for such responsibilities. OpenAI tried to walk it back, but the Electronic Frontier Foundation called their language “weasel words.” Meanwhile, the Department of Defense slapped Anthropic with a supply chain risk label – a national security designation previously reserved for hostile foreign companies. Anthropic sued the Trump administration. Then Microsoft filed a legal brief in Anthropic’s defense, joined by 149 former judges, dozens of Google and OpenAI employees, and nearly two dozen retired generals. On top of all that, security researchers discovered an unsecured data cache exposing nearly 3,000 unpublished Anthropic files, including a model code-named Mythos (also called Capybara). Internal documents describe it as a step change in capabilities, scoring dramatically higher than Opus 4.6 on coding, reasoning, and cybersecurity. Then Anthropic’s source code leaked publicly as well. Sponsor: GoDaddy Economy hosting is $6.99/month, WordPress hosting is $12.99/month, and domains are $11.99. Both hosting plans include a free domain, professional email, and SSL certificate. Go to geeknewscentral.com/godaddy for the best pricing and to directly support this independent show. OpenAI Shuts Down Sora Video App OpenAI announced on March 24th that it is killing Sora, its AI video-generation app. Downloads cratered from 3.3 million in November to 1.1 million by February. The real numbers are brutal: Sora was costing roughly $15 million per day to run against a total lifetime revenue of just $2.1 million. The Sora web and app experience ends April 26th, with the API shutting down September 24th. Additionally, the Disney partnership – a billion-dollar deal meant to validate AI in Hollywood – collapsed completely. Deep fakes of Martin Luther King Jr. and Robin Williams appeared almost immediately despite guardrails, and both families protested publicly. Cochrane notes that competitors like Runway, Pika, and Kling are still operating, and suspects Hollywood will pivot to generating scene backgrounds rather than full content. OpenClaw Is a Security Nightmare Cochrane’s personal OpenClaw install started making outbound requests flagged by his ISP – with no changes or new skills installed. He shut it down and plans to wipe the device entirely. The broader picture is alarming. A January 2026 audit found 512 vulnerabilities in OpenClaw, eight critical. Twenty-six percent of community skills contain at least one vulnerability. Oasis Security discovered a vulnerability chain called “Clawjacked” where any website can silently take full control of a developer’s agent. Between March 18th and 21st alone, nine additional vulnerabilities were disclosed, several of which were rated 9.9 out of 10. Cochrane draws a direct parallel to the browser extension era: supply chain attacks hidden as helpful tools. Claude Code Auto Mode: AI Policing AI Anthropic published details on a new “auto mode” for Claude Code after finding that users approve 93% of permission prompts – essentially mashing “yes.” Auto mode replaces manual approvals with a two-layer defense: an input scanner to detect prompt injection and a second AI model that monitors the first and decides whether to allow each action. The safety checker can only see what the user asked for and what the AI is trying to do. It cannot see the AI’s reasoning, so the AI cannot talk its way past the check. However, Cochrane notes it still misses about one in six dangerous actions (17%), and the fundamental question remains: if the base layer can get infected, so can the checker. Qwen Overtakes Llama as Most-Deployed Self-Hosted LLM RunPod’s 2026 State of AI report, based on usage data from 183 countries, reveals that Alibaba’s Qwen has overtaken Meta’s Llama as the most popular self-hosted AI model. Llama 4 has barely been adopted, with users sticking to version 3 because it just works. Additionally, vLLM now powers 40% of all AI endpoints, NVIDIA’s latest GPU usage scaled 25x last year, and nearly 70% of AI image work runs through ComfyUI. Cochrane sees Qwen winning on merit and argues that is how open source should work. AI Data Centers Are Taking All the CPUs Too AI data centers are not just consuming GPUs and memory anymore – CPUs are now being strained too. Intel server CPU lead times have stretched from two weeks to six months. AMD typically occurs at 8 to 10 weeks. Server CPU demand is projected to jump 15% in 2026, but Intel’s output capacity is growing in single digits. The shift from chatbots to autonomous AI agents is changing the hardware ratio, since agents require far more CPU power to coordinate tasks and call tools. TSMC is prioritizing more profitable AI chips over regular CPUs. Cochrane warns that consumers and businesses are effectively subsidizing the AI boom through higher prices and longer waits. AMD Ryzen 9 9950X3D2: First Dual-Cache X3D CPU AMD announced the Ryzen 9 9950X3D2, the first CPU with dual-cache X3D technology. It arrives April 22nd with 208MB of total cache and a 200W TDP – up from the current model. However, AMD is unusually honest, calling the gains “modest,” ranging from 5-13% depending on the workload. Notably, they have not released gaming benchmarks, which is conspicuous for an X3D chip. Cochrane owns a single X3D chip and sees no reason to upgrade. ARM Launches “AGI” CPU After 35 years of licensing chip designs to Apple, Qualcomm, Samsung, and NVIDIA, ARM has launched its first production silicon: a 136-core server chip co-developed with Meta as the lead customer. ARM’s stock jumped about 16% on the news. You can pack over 8,000 cores in a single air-cooled rack, or over 45,000 with liquid cooling. Volume shipments begin by the end of 2026. Cochrane appreciates the move but calls the “AGI” branding marketing hype. The bigger story is ARM transitioning from blueprint designer to direct competitor against Intel, AMD, and NVIDIA in data centers – while still licensing to the companies it now competes against. Apple Discontinues the Mac Pro Apple removed the Mac Pro from its website and confirmed that no future model is planned. The $6,999 machine had not been updated since the 2023 M2 Ultra model. Apple is pointing professionals toward the Mac Studio with its M4 Ultra chip, with an M5 Ultra refresh expected later this year. They also discontinued the $700 wheels kit, $300 feet kit, and Pro Display XDR the same week. Cochrane says good riddance – the Mac Studio covers what 90% of users need. Apple’s AI Pin: An AirTag-Sized Wearable Reports suggest Apple is developing an AirTag-sized wearable AI pin with cameras, microphones, and wireless charging. It would clip to clothing or hang as a necklace, running as an iPhone accessory powered by an upgraded Siri with Google’s Gemini AI. A possible 2027 release is expected alongside iOS 27, though development is early and could be canceled. Cochrane ties this to a broader shift: data collection moving from the application layer to physical devices. Apple employees internally refer to the device as “the eyes and ears of the iPhone.” He warns that always-on wearable cameras, combined with existing AI-powered surveillance poles, are pushing society deeper into mass data collection without meaningful consent. Quantum Entanglement Speed Measured for the First Time Scientists at TU Wien’s Institute of Theoretical Physics, led by Professor Joachim Burgdorfer, measured how fast quantum entanglement happens for the first time. The answer: about 232 attoseconds – a billionth of a billionth of a second. The research was published in Physical Review Letters in late 2024 and is now circulating widely. Einstein called quantum entanglement “spooky action at a distance.” Turns out it is not instantaneous – just extraordinarily fast. This measurement technique opens the door to quantum cryptography and quantum computing. However, Cochrane clarifies: this does not mean faster-than-light communication. Entanglement links particles but does not transmit information through space. Bronze Age Iron Artifacts Came From Outer Space Geochemical analysis by French scientist Albert Jambon, originally published in the Journal of Archaeological Science in 2017, confirmed that virtually all Bronze Age iron artifacts were made from meteorites. The artifacts span Egypt, Turkey, Syria, and China, including beads dating to 3200 BCE and the famous dagger from King Tut’s tomb, dating to around 1350 BCE. The story resurfaced after researchers published new findings this month on fragments of meteoritic iron weapons from China’s Sanxingdui sacrificial site. Bronze Age people lacked the technology to smelt iron ore, but meteoritic iron arrived in a metallic state, ready to be forged. Cochrane closes the episode, noting that ancient civilizations were working with extraterrestrial material before they could produce their own iron – resourcefulness that deserves respect. Cochrane wraps up the show by thanking GoDaddy for over twenty years of partnership and reminding listeners to subscribe, sign up for the newsletter, and reach out via email. The post Agentically Frying your Brain using AI #1861 appeared first on Geek News Central.

What happens when you sit down with a lighting designer who's saving 90% of his time using AI—and ask him to show you exactly how he does it? In this episode of LytePOD, host Sam Koerbel travels to Dubai to sit down with Faraz Izhar, a lighting designer who has transformed his entire workflow using artificial intelligence—not as a replacement for creativity, but as a power tool that amplifies it. This isn't a conversation about theory or hype. It's a candid, deeply practical look at how AI is being used right now to create cinematic presentations, automate boring tasks, and unlock creative possibilities that simply weren't feasible six months ago. Faraz reveals why prompting is the new soft skill of the design era, why AI agents are already handling luminaire schedules and technical documentation, and why the best measure of success isn't the rendering—it's how fast you can iterate, explore, and communicate your vision to clients in ways that make them feel the project before it's built. He walks through the entire process: how he uses Midjourney to create custom mood images tied directly to project narratives, how Kling and Google Veo transform static renders into cinematic sequences that show transitions from dusk to night, and how Suno generates soundtracks that elevate presentations into immersive experiences. But this conversation goes deeper. It's about the tension between automation and intuition, the risk of cultural homogenization, and why the human element must remain at the forefront—even as machines learn faster than we ever imagined. Faraz shares why guardrails matter more than speed, why AI hallucinates and how to catch it, and why the industry needs to embrace this technology now—not because it's perfect, but because the designers who don't will be left behind.

What happens when you sit down with a lighting designer who's saving 90% of his time using AI—and ask him to show you exactly how he does it?In this episode of LytePOD, host Sam Koerbel travels to Dubai to sit down with Faraz Izhar, a lighting designer who has transformed his entire workflow using artificial intelligence—not as a replacement for creativity, but as a power tool that amplifies it. This isn't a conversation about theory or hype. It's a candid, deeply practical look at how AI is being used right now to create cinematic presentations, automate boring tasks, and unlock creative possibilities that simply weren't feasible six months ago.Faraz reveals why prompting is the new soft skill of the design era, why AI agents are already handling luminaire schedules and technical documentation, and why the best measure of success isn't the rendering—it's how fast you can iterate, explore, and communicate your vision to clients in ways that make them feel the project before it's built. He walks through the entire process: how he uses Midjourney to create custom mood images tied directly to project narratives, how Kling and Google Veo transform static renders into cinematic sequences that show transitions from dusk to night, and how Suno generates soundtracks that elevate presentations into immersive experiences.But this conversation goes deeper. It's about the tension between automation and intuition, the risk of cultural homogenization, and why the human element must remain at the forefront—even as machines learn faster than we ever imagined. Faraz shares why guardrails matter more than speed, why AI hallucinates and how to catch it, and why the industry needs to embrace this technology now—not because it's perfect, but because the designers who don't will be left behind.

In this episode, Dr. Christopher Wolfe discusses the Vision Source D25 program with Mick Kling, exploring its origins, benefits, and how it fosters community and value among independent optometrists. Discover how the program reinvests in practices and strengthens the independent optometry community.   Chapters 00:00 Weather Talk and Introduction 01:11 Unpacking the D25 Program 02:04 The Origin Story of Vision Source 05:10 The Value of Community in Vision Source 09:30 Understanding the D25 Program 11:06 Member Dividends Explained 17:42 The Mechanics of D25 Participation 28:49 Future Opportunities and Closing Thoughts  resources Vision Source Official Website - https://visionsource.com Insight Member Platform - https://visionsource.com/insight Mick Kling on LinkedIn - https://www.linkedin.com/in/mickkling/ Amir Kashnevis on LinkedIn - https://www.linkedin.com/in/amirkashnevis/  guest links LinkedIn - https://www.linkedin.com/in/mickkling/   ---------------------- For our listeners, use the code 'EYECODEMEDIA22' for 10% off at check out for our Premiere Billing & Coding bundle or our EyeCode Billing & Coding course. Sharpen your billing and coding skills today and leave no money on the table! questions@eyecode-education.com https://coopervision.com/our-company/news-center/press-release/coopervision-and-aoa-join-forces-launch-myopia-collective Go to MacuHealth.com and use the coupon code PODCAST2024 at checkout for special discounts  Show Sponsors: CooperVision MacuHealth

OpenAI's GPT-5.4 is the newest state-of-the-art AI model. But for how long? Sam Altman and team have to fend off Dario Amodei's Anthropic in more ways than one.  GPT-5.4 beats humans at computer use, hallucinates 33% less, and costs half what Opus 4.6 does for coding. Meanwhile Anthropic just got blacklisted by the Pentagon, Claude shot to number one in the App Store, and Dario went on CBS looking like he aged five years. Plus Kling Motion Control 3.0, Grok Imagine Extend, Notebook LM cinematic overviews, Ben Affleck's secret AI startup sold to Netflix, and a device that jams AI from recording you. THE KING WEARS THE CROWN UNTIL NEXT TUESDAY. THAT'S HOW THIS WORKS NOW. #ai #ainews #openai Come to our Discord: https://discord.gg/muD2TYgC8f Join our Patreon: https://www.patreon.com/AIForHumansShow AI For Humans Newsletter: https://aiforhumans.beehiiv.com/ Follow us for more on X @AIForHumansShow Join our TikTok @aiforhumansshow To book us for speaking, please visit our website: https://www.aiforhumans.show/ // Show Links //   OpenAI's GPT-5.4 is here… https://openai.com/index/introducing-gpt-5-4/ CUA (Computer Use Agent) OpenAI Demo https://x.com/OpenAIDevs/status/2029620984853188738?s=20) ChatGPT 5.3 Instant (the emojis are back) https://x.com/OpenAI/status/2028893701427302559?s=20 Dario Amodei Rips Sam, Calls OpenAI "Mendacious" And Other Things About Trump in Memo https://www.theinformation.com/articles/read-anthropic-ceos-memo-attacking-openais-mendacious-pentagon-announcement?rc=c3oojq&shared=d3dcdfe79bb39c22 Sam Altman Response To Backlash  https://x.com/sama/status/2028640354912923739?s=20 Voice Mode For Claude Code https://x.com/trq212/status/2028628570692890800?s=20 AI for Humans Link Blog https://www.aiforhumans.show/linkblog Kling Motion Control 3.0 https://x.com/Kling_ai/status/2029391461880574136?s=20 Solid Examples https://x.com/maxescu/status/2029223608854434000?s=20 Grok Imagine Extend Video https://x.com/grok/status/2028571874188259337?s=20 Gavin's Grok Extended video: https://x.com/gavinpurcell/status/2029423165571776587?s=20 Cinematic Overviews for NotebookLM https://x.com/NotebookLM/status/2029240601334436080?s=20 Justine example https://x.com/venturetwins/status/2029592813277683750?s=20 Netflix Buys Ben Affleck's Secretive AI Company https://variety.com/2026/film/news/netflix-acquires-ben-affleck-ai-filmmaking-startup-interpositive-1236679498/ Spectre 1 Stops AI Audio Recordings IRL https://x.com/aidaxbaradari/status/2028864606568067491?s=20 The Shape Store Viral Tiktok/IG https://www.tiktok.com/@a.i.solation/video/7612337372667071774?_r=1&_t=ZT-94RCk4vtqYg Bilawal Sidhu's Operation Fury Visualization With Actual Geolocation Data https://youtu.be/rXvU7bPJ8n4?si=n4aoCsgHLQ6bq432 Infinite Favicons https://www.favicons.art/ Open Source Vibe Coded Music Video Maker https://x.com/blizaine/status/2029330454122242106?s=20 New AI For Humans Website! https://www.aiforhumans.show/

Matchday 1 of the SheBelieves Cup has wrapped and Sam and Kling are here to recap all the action! We saw a 2-0 USWNT win against Argentina that included a captain Lindsey Heaps goal AND Emma Hayes in a cowboy hat. Listen now before the USWNT take on Canada for game 2 on Wednesday.TWG Do it Live!, is presented by the great friends at Olipop, use code GVFC for a buy one get one free offer at drinkolipop.com/GVFCSUBSCRIBE TO THE WOMEN'S GAME NEWSLETTER: https://mibcourage.co/42X5HpBNEW TWG MERCH HERE: https://mibcourage.co/3MxW4rNSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Get our AI Video Guide: https://clickhubspot.com/dth Episode 97: How close are we to a world where AI-generated videos are indistinguishable from reality? Matt Wolfe (https://x.com/mreflow) and Joe Fier (linkedin.com/in/joefier) dive deep into Seedance 2.0—ByteDance's new AI video model that could outpace giants like Sora and Veo. Joe, a marketing and business expert known for his hands-on approach and insights into AI's rapid evolution, helps to break down the five most fascinating developments in the AI space this week. They tackles game-changing AI advances: Seedance 2.0's mind-blowing video generation for ads and motion graphics, the rollout of Google's Veo 3.1 in Google Ads, the GPT-5.3 Codex Spark coding model built on specialized inference chips, Gemini's DeepThink model for scientific research, and the early rollout of ChatGPT ads. Check out The Next Wave YouTube Channel if you want to see Matt and Nathan on screen: https://lnk.to/thenextwavepd — Show Notes: (00:00) Seedance 2.0 arrives – AI video generation blurs reality, ad creation moves fast. (03:03) Google's Veo 3.1 powers video ads, advertisers can now generate clips directly from image uploads. (05:33) Comparison of Runway, Kling, Veo, and Sora—head-to-head prompt showdown. (07:00) Motion graphics and explainers—AI's take on the creative industry. (08:35) US vs. China—Copyright, IP, and training data debates. (12:10) Deepfake and video authenticity—why we now default to skepticism. (13:30) Google's edge in visual AI via YouTube's massive corpus. (14:39) The next frontier: Longer, more consistent video generation. (15:14) Where do humans fit in? Taste, storytelling, and creative direction. (18:30) GPT-5.3 Codex Spark—coding models on Cerebras inference chips, demo generating a website in 18 seconds. (24:34) AI tool comparisons—Codex vs. Cursor vs. Claude Code. (25:12) Speed as the key bottleneck breaker in creative and technical workflows. (28:02) Google's Gemini DeepThink—state-of-the-art research, advanced coding and physics capabilities. (32:52) Gemini demo attempt—3D-printable STL file and solving the three-body problem. (33:20) ChatGPT rolls out ads—impact on monetization and user trust. (40:02) Google's ad history—how “sponsored” is becoming harder to distinguish. (44:02) Democratizing AI access via ad-supported models. (45:03) Matt Schumer's viral article—why AI is moving even faster than most people realize. (51:11) Tools that build tools—AGI's path and the new role for humans. (53:12) Real-world skills and taste—where humanity still wins (for now). (54:01) Final thoughts—wake up, pay attention, and stay on the leading edge. — Mentions: Seedance 2.0: https://www.seedance.com/ ByteDance: https://www.bytedance.com/ CapCut: https://www.capcut.com/ Veo: https://deepmind.google/models/veo/ Runway: https://runwayml.com/ ChatGPT Codex: https://chatgpt.com/codex Matt Schumer's Viral Article: https://www.mattshumer.com/blog/ai-changes-everything Super Bowl Claude Commercial: https://www.anthropic.com/news/super-bowl-ad Get the guide to build your own Custom GPT: https://clickhubspot.com/tnw — Check Out Matt's Stuff: • Future Tools - https://futuretools.beehiiv.com/ • Blog - https://www.mattwolfe.com/ • YouTube- https://www.youtube.com/@mreflow — Check Out Nathan's Stuff: Newsletter: https://news.lore.com/ Blog - https://lore.com/ The Next Wave is a HubSpot Original Podcast // Brought to you by Hubspot Media // Production by Darren Clarke // Editing by Ezra Bakker Trupiano

Enterprise XR hasn't disappeared, it has quietly moved into places where it saves time, reduces errors and changes how people work every day. On this episode of the AI XR Podcast, Charlie Fink and Rony Abovitz talk with Boston Consulting Group partner Kristi Woolsey, who leads BCG's immersive practice, about how XR plus AI is already being used for training, maintenance, onboarding, retail and architecture inside some of the world's most conservative organizations.Kristi shares a Swiss Rail project where field technicians wear lightweight AR glasses that recognize who they are and which train car they are standing in front of, pull the correct procedures from internal systems and use AI to turn thick manuals into simple task checklists.She explains how this leads to double-digit efficiency gains for both experienced and new workers, and how a small behavior design choice – automatic logging for headset users versus manual end-of-shift paperwork for everyone else – helped overcome skepticism on the front line. Drawing on her background as a physical-space architect, she also describes how VR and rapidly improving 3D tools are changing the way companies design stores, offices and buildings before anything physical is built.AI XR News you should know, Charlie and Rony cover Anthropic's massive new funding round and ethics turbulence, Chinese generative video tools like Seed Dance 2 and Kling that put TV-quality visuals in reach of “garage Spielbergs,” and Meta's reported seven million Ray-Ban and Oakley AI smart glasses sold – early signals of where wearable AI and XR are really headed.Key Moments01:03 – Anthropic's huge raise and what the ethics departure might signal05:08 – Seed Dance 2 and Kling showcase a new level of generative video08:35 – Meta's seven million smart glasses and the reality behind that number12:10 – Why wearable AI may be the real “last mile” of turning us into cyborgs15:28 – Inside the early metaverse tours Kristi and Rony built for enterprises20:27 – How BCG's VR onboarding keeps new hires engaged months before day one23:30 – Swiss Rail's AR and AI maintenance assistant and what it actually does on site27:05 – Designing XR systems that give value to both the business and frontline workers30:29 – Using VR as a lab for retail and workplace behavioral strategy33:06 – How AI-generated 3D models point toward “build every space digitally first”This episode shows how “metaverse” ideas have turned into practical tools: XR plus AI is cutting training times, improving maintenance quality and letting companies experiment with spaces before they exist. Kristi's examples make it clear that the real action is in careful workflow design, not flashy avatars.This episode is brought to you by Zappar, creators of Mattercraft, the leading visual development environment for building immersive 3D web experiences for mobile, headsets and desktop. https://mattercraft.io/Mattercraft combines the power of a game engine with the flexibility of the web and now includes an AI assistant that helps you design, code and debug in real time, right in your browser. To explore what's possible with AI-powered XR on the web, start building smarter with Mattercraft from Zappar.Listen to “Enterprise XR Meets AI: How Smart Glasses, Digital Twins and Holodecks Are Quietly Changing Work – Kristi Woolsey” on the AI XR Podcast and follow the show for new episodes every week.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Anthropic drops Opus 4.6. Twenty minutes later, OpenAI fires back with GPT-5.3 Codex. This is the AI agentic coding arms race and it's moving fast. Both AI models are writing code that can write itself now. OpenAI is using 5.3 to improve its own tooling. Opus 4.6 is "voicing discomfort with being a product." We tested both and break down what actually matters for people building stuff.  Plus Kling 3.0 is out (and harder to prompt than you think), OpenClaw bots are hiring humans on rent-a-human.ai, Roblox launches prompt-to-3D creation, and robots are now doing 130K step challenges in negative 47 degree weather. THE MODELS ARE IMPROVING THEMSELVES NOW. EVERYTHING IS FINE. Come to our Discord: https://discord.gg/muD2TYgC8f Join our Patreon: https://www.patreon.com/AIForHumansShow AI For Humans Newsletter: https://aiforhumans.beehiiv.com/ Follow us for more on X @AIForHumansShow Join our TikTok @aiforhumansshow To book us for speaking, please visit our website: https://www.aiforhumans.show/ // Show Links // Anthropic's Claude Opus 4.6 https://www.anthropic.com/news/claude-opus-4-6 Orchestrating Agents in Claude Code https://x.com/lydiahallie/status/2019469032844587505?s=20 Opus 4.6 Beats Humans at analyzing complex human science docs https://x.com/_simonsmith/status/2019502742209769540?s=20 OpenAI GPT-5.3 Codex https://openai.com/index/introducing-gpt-5-3-codex/ 5.3 Codex First model instrumental in creating itself https://x.com/deredleritt3r/status/2019475360438493597 OpenAI Frontier https://openai.com/index/introducing-openai-frontier/ Anthropic's Superbowl Ads https://x.com/tomwarren/status/2019039874771550516?s=20 GPT-5 connected to an autonomous lab to do experiments https://x.com/OpenAI/status/2019488071134347605?s=20 OpenClaw https://openclaw.ai/ Rent-A-Human https://rentahuman.ai/bounties Kling 3.0 = Really good model https://x.com/Kling_ai/status/2019064918960668819?s=20 Kling 3.0 Moonlanding Mockumentary https://x.com/Kling_ai/status/2019228615775604784?s=20 PJ Ace's Way of Kings Intro https://x.com/PJaccetturo/status/2019072637192843463?s=20  We Are The Art | Brandon Sanderson's Keynote Speech https://youtu.be/mb3uK-_QkOo?si=EgKBjxZf4GE4DYIJ Gavin's Kling Fail https://x.com/gavinpurcell/status/2019436331999588371?s=20 FIGMA VECTOR AI https://x.com/moguzbulbul/status/2019106665732403708?s=20 Grok Imagine 1.0 Officially Launches https://x.com/xai/status/2018164753810764061?s=20 Roblox Launches 4D Creation https://x.com/Roblox/status/2019221624604750238 Unitree Robot Walks Across The Tundra (-47C!!) https://x.com/War_Radar2/status/2018315065414635813?s=20 KinectIQ's Humanoid Framework https://youtu.be/Y2DhzLPGdwY?si=iWibCGoc_h53yZz3 The LooksMaxxor https://x.com/Gossip_Goblin/status/2018362969025884282?s=20 Midi-Survivor https://x.com/measure_plan/status/2019082789379858577?s=20

Get PJ's free AI Video Production Stack + Workflow: https://clickhubspot.com/whs Ep. 393 233 million views in just three days — can AI-generated ads really replace million-dollar productions? Kipp, Kieran, and guest, PJ Accetturo, of Genre.ai, dive into the wild world of AI-powered commercial workflows and the viral David Beckham ad that's turning heads across the industry. Learn more about AI-driven creative teams, the tools behind photorealistic video production, and the emerging future—where hyper-niche stories thrive and challenger brands outsmart the incumbents. Mentions PJ Accetturo https://www.linkedin.com/in/pj-accetturo-b3b693129/ Genre.ai https://www.genre.ai/ Figma https://www.figma.com/ Nano Banana Pro https://gemini.google/overview/image-generation/ Freepik https://www.freepik.com/ai/image-generator Veo 3.1 https://gemini.google/overview/video-generation/ Kling https://klingai.com/global/ ElevenLabs https://elevenlabs.io/ Get our guide to build your own Custom GPT: https://clickhubspot.com/customgpt We're creating our next round of content and want to ensure it tackles the challenges you're facing at work or in your business. To understand your biggest challenges we've put together a survey and we'd love to hear from you! https://bit.ly/matg-research Resource [Free] Steal our favorite AI Prompts featured on the show! Grab them here: https://clickhubspot.com/aip We're on Social Media! Follow us for everyday marketing wisdom straight to your feed YouTube: ​​https://www.youtube.com/channel/UCGtXqPiNV8YC0GMUzY-EUFg  Twitter: https://twitter.com/matgpod  TikTok: https://www.tiktok.com/@matgpod  Join our community https://landing.connect.com/matg Thank you for tuning into Marketing Against The Grain! Don't forget to hit subscribe and follow us on Apple Podcasts (so you never miss an episode)! https://podcasts.apple.com/us/podcast/marketing-against-the-grain/id1616700934   If you love this show, please leave us a 5-Star Review https://link.chtbl.com/h9_sjBKH and share your favorite episodes with friends. We really appreciate your support. Host Links: Kipp Bodnar, https://twitter.com/kippbodnar   Kieran Flanagan, https://twitter.com/searchbrat  ‘Marketing Against The Grain' is a HubSpot Original Podcast // Brought to you by Hubspot Media // Produced by Darren Clarke.

A Journey in Holistic WellnessIn This Podcast: Clinical herbalist Kimberly Kling returns to discuss regenerative health in a highly toxic modern world. Drawing from personal experience, clinical practice, and ecological awareness, she explains how petrochemicals, industrial agriculture, and environmental toxins disrupt human health—especially the gut microbiome, mitochondria, and detox pathways. The conversation moves from root causes to practical, accessible steps people can take, including food choices, herbs, lifestyle shifts, and community action. Throughout, the focus remains on empowerment, resilience, and reconnecting with plant wisdom rather than fear.Our Guest: Kimberly is a clinical herbalist and the guiding force behind joyful roots in Southern Arizona where she helps her community locally and beyond cultivate inner wellness through earth centered herbal care, rooted in a deep reverence for the healing power of plants. Kimberly's journey began in childhood, crafting magical plant stews and foraging connections with Michigan's native flora. Her background in landscape architecture and engineering provided a foundation for understanding the intricate relationships between plants, people, and the land. However, it was motherhood and a personal health crisis that led to her clinical herbalism deepening her passion for holistic wellness. Now, Kimberly integrates traditional wisdom with modern herbal practices, empowering others to reconnect with plant wisdom for vibrant health and wellbeing.Medical Disclaimer: In today's episode we are talking about our health. The information provided in this podcast is for general information and entertainment purposes only. It does not constitute medical advice, diagnosis or treatment. We are not medical doctors and no medical doctor/patient relationship is formed. Always seek advice from your qualified medical doctor regarding questions you may have about your medical condition.Key Topics & EntitiesKimberly KlingJoyful RootsClinical herbalismEnvironmental toxinsPetrochemicalsHaber-Bosch ProcessGlyphosate, Diquat, ParaquatGut microbiomeMitochondrial healthAutoimmune illness (lupus)AntioxidantsLiver detoxificationRegenerative agricultureFood forestsKey Questions AnsweredWhy are modern humans experiencing chronic illness earlier than previous generations?Because exposure to synthetic chemicals, petrochemicals, pesticides, plastics, and food additives has rapidly increased over the last ~150 years, overwhelming biological systems that evolved alongside natural substances.How do pesticides and herbicides affect the body if they're “safe for humans”?They often harm microbial...

How can your brand build genuine loyalty that translates into long-term business value when customer expectations continually evolve?Agility demands a deep understanding of your customer, a willingness to experiment, and the ability to pivot quickly when needed.Today, we're going to talk about building loyalty and trust through focused innovation, specifically within the competitive landscape of the baby care market. To help me discuss this topic, I'd like to welcome, Lindsey Kling, SVP Brand Marketing + Partnerships at Coterie. About Lindsey Kling Lindsey has 15 years of experience in scaling brands from startup to unicorn status, leveraging her expertise in strategic partnerships and marketing to drive growth and innovation in competitive markets. Prior to Coterie, she served as Head of Partnerships at Away where she built and led the partnerships function, forging brand-defining deals with partners such as Serena Williams, American Express and United Airlines. Before Away, Lindsey led the partnerships strategy for Uber and Uber Eats, developing foundational alliances that contributed to the brand's growth and ascent to a $100b valuation.Lindsey is a graduate of New York University and resides in Northern California with her husband and two children, Blaire and Asher. Lindsey Kling on LinkedIn: https://www.linkedin.com/in/lindseykling/ Resources Coterie: https://www.coterie.com/ The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://www.teksystems.com/versionnextnow Catch the future of e-commerce at eTail Palm Springs, Feb 23-26 in Palm Springs, CA. Go here for more details: https://etailwest.wbresearch.com/ Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstromDon't miss a thing: get the latest episodes, sign up for our newsletter and more: https://www.theagilebrand.showCheck out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company Hosted on Acast. See acast.com/privacy for more information.

How can your brand build genuine loyalty that translates into long-term business value when customer expectations continually evolve? Agility demands a deep understanding of your customer, a willingness to experiment, and the ability to pivot quickly when needed. Today, we're going to talk about building loyalty and trust through focused innovation, specifically within the competitive landscape of the baby care market. To help me discuss this topic, I'd like to welcome, Lindsey Kling, SVP Brand Marketing + Partnerships at Coterie. About Lindsey Kling Lindsey has 15 years of experience in scaling brands from startup to unicorn status, leveraging her expertise in strategic partnerships and marketing to drive growth and innovation in competitive markets. Prior to Coterie, she served as Head of Partnerships at Away where she built and led the partnerships function, forging brand-defining deals with partners such as Serena Williams, American Express and United Airlines. Before Away, Lindsey led the partnerships strategy for Uber and Uber Eats, developing foundational alliances that contributed to the brand's growth and ascent to a $100b valuation.Lindsey is a graduate of New York University and resides in Northern California with her husband and two children, Blaire and Asher. Lindsey Kling on LinkedIn: https://www.linkedin.com/in/lindseykling/ Resources Coterie: https://www.coterie.com/ The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://www.teksystems.com/versionnextnow Catch the future of e-commerce at eTail Palm Springs, Feb 23-26 in Palm Springs, CA. Go here for more details: https://etailwest.wbresearch.com/ Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstromDon't miss a thing: get the latest episodes, sign up for our newsletter and more: https://www.theagilebrand.showCheck out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company