Podcasts about Selo

Razmišljate li i vi o odlasku iz grada na selo? Od pandemije korona virusa sve je primjetniji trend bijega stanovnika većih gradova u manja mjesta ili na selo i to ne samo u Njemačkoj. Prednosti su mnoge ali nakon nekog vremena i mane izlaze na vidjelo. Koje? I zašto neki opet sele u grad? Nenad Kreizer razgovara s novinarkom Ivom Hanzen koje je nedavno udobni život u gradu zamijenila životom u prirodi a reporterka Maja Marić navodi aspekte preseljavanja urbanog stanovništva u ruralne sredine. Von Nenad Kreizer.

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

Seorang pendaki bernama Aditio Rahmat Eko Prayudi (38) asal Kabupaten Bogor meninggal dunia saat melakukan pendakian di Gunung Merbabu melalui jalur Selo, Boyolali, Jawa Tengah. Korban yang sempat melakukan registrasi resmi pada pukul 10.12 WIB memilih balik kanan ke basecamp karena merasa kondisi fisiknya tidak fit, namun tiba-tiba pingsan pada pukul 11.58 WIB berdasarkan pantauan CCTV. Pihak Balai Taman Nasional Gunung Merbabu bersama tim gabungan langsung melakukan penanganan medis dan evakuasi ke Puskesmas Selo, hingga akhirnya korban dinyatakan meninggal dunia pada pukul 12.55 WIB akibat henti jantung.Narasumber: > Alfa Sandy Aprazah > (Kepala Seksi Pengelolaan Taman Nasional Wilayah II Taman Nasional Gunung Merbabu)

#tocadodragao #2026 #podcast #ouvintes #emails #comentarios #comedia #discussao #musica #cerebro #youtube #tretaEpisódio de hoje: Mais um selinho! PLOFFFT! Tá colado! Vamos pra mais uma leitura de e-mails caótica como vocês gostam!ENTRE NA COMUNIDADE DO TOCA! ⁠⁠https://cesber.wixsite.com/tocadodragao⁠⁠REDES SOCIAIS E MUITO MAIS!https://beacons.ai/tocadodragaoFAÇA SUA DOAÇÃO #APOIE a TOCA a partir de R$ 10,00/ mês - Estamos no Apoia.se!https://apoia.se/atocadodragaoDOADORES DE JULHO/2026 PAULO DEROS ELVE, BRUNO BRAZ, RODRIGO SILVA, MARCIA REGINA BERNARDES, MASON YEON, PAULA GESTAL, GABRIEL SCHADE, LELE DANTAS, JOÃO PANDA, CEZAR AUGUSTO, ANTHONY MARTINS, BRENDA NASCIMENTO, MISTER DOVAH, LENHORMAR, VICTOR FERNANDES, MATHEUS TRENTINIAgradecemos aos Inscritos do Podcast que fizeram suas doações pelo PICPAY nosso e-mail: tocadodragaopodcast@gmail.comGRUPO DO TELEGRAM https://t.me/+fn75BRye8sY2NDExGRUPO DO WHATSAPPhttps://chat.whatsapp.com/KUtDsVnnv7w6hcseloXqCQCASTERS NESSE EPISÓDIO: Richard (Ricky, O Bardo) e Rodrigo SilvaMÚSICAS ORIGINAIS DO TOCA #Compositor: Caio Varalta / Tema do Podcast: "Entrando na Toca" - Todos os Direitos Reservados

No Ponto de Partida React desta sexta-feira (17), Luiza Silvestrini e Pedro Doria comentam as opiniões da audiência do Meio sobre o sistema eleitoral brasileiro, os reflexos dele na composição do Congresso e, consequentemente, na destinação do orçamento, e sobre a proposta do Tribunal Superior Eleitoral de conceder um selo aos institutos de pesquisa que mais se aproximarem do resultado das eleições.See omnystudio.com/listener for privacy information.

Confira os destaques de Os Pingos nos Is desta terça-feira (14):No editorial do programa Os Pingos nos Is desta terça-feira (14), o apresentador André Marsiglia voltou a falar da decisão do ministro Alexandre de Moraes, que proibiu o senador e pré-candidato à Presidência Flávio Bolsonaro (PL-RJ) de visitar o pai, o ex-presidente Jair Bolsonaro, por 90 dias após a leitura de uma carta. Marsiglia aponta um grave erro de interpretação jurídica na medida: as restrições impostas ao ex-presidente o proíbem de usar redes sociais, mas não impedem que terceiros utilizem seus próprios canais para divulgar suas ideias ou ler suas cartas. Ainda segundo sua análise, a decisão do STF abre margem para que o caso seja explorado politicamente por Flávio para sensibilizar seu eleitorado. "Se o Judiciário vai entrar de sola nessas eleições, que entre com o pé direito. Ou melhor, entre com os dois", conclui. A Pesquisa Futura/Apex divulgada nesta terça-feira (14) mostra que a atuação do Supremo Tribunal Federal (STF) é desaprovada por 52,3% dos brasileiros e aprovada por 35,9%. Outros 11,8% dos entrevistados não souberam responder. O levantamento apresenta a percepção da população sobre o desempenho da Corte. A bancada do Os Pingos Nos Is comenta os números. Um levantamento do Tribunal Superior Eleitoral (TSE) aponta que nove de 29 partidos brasileiros sobrevivem de doações eleitorais, principalmente de seus filiados. Isso ocorre porque as siglas não possuem acesso ao Fundo Partidário ou recebem recursos insuficientes para cobrir suas despesas. Reportagem: André Anelli. A Polícia Federal (PF) concluiu um dos inquéritos sobre desvios em aposentadorias e indiciou o ex-presidente do Instituto Nacional do Seguro Social (INSS), Alessandro Antônio Stefanutto, o ex-procurador-geral da autarquia, Virgílio Antônio Ribeiro Filho, o ex-diretor de benefícios André Fidelis e outros investigados por suspeita de corrupção. A Associação Brasileira de Empresas de Pesquisa (Abep) criticou a proposta do Tribunal Superior Eleitoral (TSE), idealizada pelo presidente da Corte, ministro Nunes Marques, de criar um "Selo de Acurácia Eleitoral". A iniciativa visa premiar os institutos de pesquisa que apresentarem maior proximidade com o resultado oficial das urnas. Reportagem: Thaís Sprovieri. A defesa de Mauro Cid se reuniu com o empresário e ex-banqueiro Daniel Vorcaro, dono do Banco Master, na Papudinha, em Brasília, onde está preso. O encontro irritou os advogados de Vorcaro. Você confere essas e outras notícias em Os Pingos nos Is.

(00:00:00) Xadrez Verbal #467 OTAN e racismo na Copa do Mundo (00:05:40) Giro de Notícias #01 (00:30:35) Coluna Aberta: Cúpula da OTAN (01:09:40) Efemérides: A Semana na História (01:17:05) Match: América Latina (01:54:40) Xeque: Bacia do Pacífico e Copa do Mundo 2026 (02:35:30) Giro de Notícias #02 (02:58:50) Peões da Semana (03:02:35) Sétimo Selo (03:12:40) Música de Encerramento O racismo de uma senadora paraguaia criou uma crise diplomática com a França! Passamos por essa e outras notícias internacionais da Copa do Mundo.Também repercutimos a cúpula da OTAN, além de observamos o movimento das peças no sempre complicado tabuleiro do Oriente Médio.No mais, demos aquele tradicional pião pela nossa quebrada latino-americana, com destaque para a crise política na Colômbia.Use o código XADREZ na Você Europeu para ter condições exclusivas: https://voceeuropeu.com.br/xadrez/Conheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

(00:00:00) Xadrez Verbal #466 Duas Convidadas e Duas Semanas (00:08:30) Giro de Notícias #01 (00:28:45) Coluna Aberta #01: Marc Bloch no Panteão Francês, com Fran Becher (00:55:45) Coluna Aberta #02: Brasil no G7, com Carolina Marins (01:23:00) Efemérides: A Semana na História (01:43:20) Coluna Aberta #03: Oriente Médio (02:38:40) Europa e Copa do Mundo 2026 (04:09:45) Xeque: América Latina (05:17:45) Gambito da Dama: Alan Greenspan (05:33:00) Giro de Notícias #03 (05:48:45) Peões da Semana (05:50:15) Sétimo Selo (06:05:45) Música de Encerramento #01 (06:08:45) Música de Encerramento #02 Recebemos a professora Fran Becher para uma análise sobre a introdução do historiador francês Marc Bloch no Panteão de Paris.Também conversamos com a jornalista Carolina Marins sobre a participação brasileira na 52ª Cúpula do G7, realizada na cidade francesa de Evian.No mais, observamos o movimento das peças no sempre complicado tabuleiro do Oriente Médio e demos uma volta pelo Velho Continente, com destaque para a renúncia do primeiro-ministro britânico Keir Starmer.Por fim, demos aquele tradicional pião pela quebrada latino-americana, repercutindo os terremotos na Venezuela e os resultados eleitorais na Colômbia e Peru.Cuide de sua saúde mental com a Psicólogos Brasil: https://www.psicologosbr.com/Apoie a cientista brasileira Glaucia Lidiane da Silva: https://www.vakinha.com.br/vaquinha/ajude-glaucia-a-fazer-justica-contra-os-espanhois-pelo-plagio-do-metodo-da-taylorConheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Em 2026, iniciativa do Ministério Público do Estado da Bahia (MPBA), em parceira com os Tribunais de Contas do Estado (TCE/BA) e dos Municípios (TCM/BA), registrou mais R$ 615 milhões com atrações artísticas.

Episodio 374.¿Cómo decir esto suavemente? Mi podcast de bañarme, lo descubrí de noche con lágrimas en los ojos. Siendo honesto, cosa que a mi me cuesta mucho, hay algo ahí que no entiendo y no quiero entender porque son consejos no financieros sobre finanzas de la época favorita de Marx. ¿Sabes qué? Voy a escribir una canción sobre latencia.

#tocadodragao #2026 #podcast #ouvintes #emails #comentarios #comedia #discussao #videogamesEpisódio de hoje: Mais um selo do nosso programa especial, dedicados a vocês nos incríveis ouvintes! Bora pra mais um Toca Postal! Selo nº 12!ENTRE NA COMUNIDADE DO TOCA! ⁠⁠https://cesber.wixsite.com/tocadodragao⁠⁠REDES SOCIAIS E MUITO MAIS!https://beacons.ai/tocadodragaoFAÇA SUA DOAÇÃO #APOIE a TOCA a partir de R$ 10,00/ mês - Estamos no Apoia.se!https://apoia.se/atocadodragaoDOADORES DE MAIO/2026 PAULO DEROS ELVE, BRUNO BRAZ, RODRIGO SILVA, MARCIA REGINA BERNARDES, MASON YEON, PAULA GESTAL, GABRIEL SCHADE, LELE DANTAS, JOÃO PANDA, CEZAR AUGUSTO, ANTHONY MARTINS, ANDRIA SEDREZ, BRENDA NASCIMENTO, MISTER DOVAH, LENHORMAR, VICTOR FERNANDESAgradecemos aos Inscritos do Podcast que fizeram suas doações pelo PICPAY nosso e-mail: tocadodragaopodcast@gmail.comGRUPO DO TELEGRAM https://t.me/+fn75BRye8sY2NDExGRUPO DO WHATSAPPhttps://chat.whatsapp.com/KUtDsVnnv7w6hcseloXqCQCASTERS NESSE EPISÓDIO: Richard (Ricky, O Bardo) e Rodrigo Silva (Capivara Molhada)MÚSICAS ORIGINAIS DO TOCA #Compositor: Caio Varalta / Tema do Podcast: "Entrando na Toca" - Todos os Direitos Reservados

“Sabe, porém, isto: que nos últimos dias sobrevirão tempos trabalhosos. Porque haverá homens amantes de si mesmos, avarentos, presunçosos, soberbos, blasfemos, desobedientes a pais e mães, ingratos, profanos, Sem afeto natural, irreconciliáveis, caluniadores, intemperantes, cruéis, sem amor para com os bons, Traidores, obstinados, orgulhosos, mais amigos dos deleites do que amigos de Deus, Tendo aparência de piedade, mas negando a eficácia dela. Destes afasta-te.” II Timóteo 3:1-5“Tu, porém, tens seguido a minha doutrina, modo de viver, INTENÇÃO, fé, longanimidade, amor, paciência, Perseguições e aflições tais quais me aconteceram...” II Timóteo 3:10-11

Polarizo com o fim de brincar Estabilizando a ilusão Selo o processo da magia com o tom lunar do desafio Eu sou guiada pelo poder da Abundância

(00:00:00) Xadrez Verbal #462 Xi Anfitrião e Protestos na Bolívia (00:06:30) Giro de Notícias #01 (00:21:40) Coluna Aberta: visitas a Xi Jinping e Oriente Médio (01:09:40) Efemérides: A Semana na História (01:17:45) Match: Europa (01:46:50) Giro de Notícias #02 (01:53:15) Xeque: América Latina (03:02:25) Gambito da Dama: presidência do FED (03:10:40) Giro de Notícias #03 (03:25:00) Peões da Semana (03:26:30) Sétimo Selo (03:38:40) Música de Encerramento Em pouco mais de 6 meses de governo, o presidente boliviano Rodrigo Paz se vê ameaçado por protestos e tentamos explicar as diversas motivações para isso, além de outras notícias da nossa quebrada latino-americana.Já na China, Xi Jinping recebeu Trump e Putin separadamente. Comentamos os desdobramentos dessas visitas e um possível acordo com o Irã. No mais, repercutimos mais uma polêmica edição do Eurovision, com vitória inédita da Bulgária, em nossa volta pelo Velho Continente.Use o cupom XADREZVERBAL na Academia Guhan de Mandarim: https://tinyurl.com/fy65wbbbConheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

"E, chamando os Seus discípulos, disse-lhes: Em verdade vos digo que esta pobre viúva deitou mais do que todos os que deitaram na arca do tesouro; Porque todos ali deitaram do que lhes sobejava, mas esta, da sua pobreza, deitou tudo o que tinha, todo o seu sustento." Marcos 12:43-44

A Associação de Pais e Amigos de Autistas (AMA) de Lauro Müller segue trabalhando para expandir os atendimentos multidisciplinares a crianças com Transtorno do Espectro Autista (TEA) no município, especialmente aquelas em situação de vulnerabilidade social. A principal ferramenta para isso é o projeto Selo Empresa Amiga do Autista, que funciona por meio de um sistema de adoção solidária. Atualmente, a instituição atende 45 crianças com suporte de psicólogo, fonoaudiólogo e pedagogo. Em entrevista ao repórter Álvaro Souza, a presidente da AMA, Fabiana Cataneo, atualizou o andamento da iniciativa e confirmou a adesão das duas primeiras empresas parceiras: uma conquistou o selo ouro e outra o selo bronze. "Esses dois selos já nos dão uma ajuda. Com essa adesão, a gente vai conseguir dar uma aumentada nos nossos números de atendimento de psicologia", destacou Fabiana. Como ajudar? Como as demandas são altas e há fila de espera, a AMA reforça a necessidade de novos parceiros comerciais. Empresas interessadas podem entrar em contato com a associação para se tornarem mantenedoras diretas dos serviços terapêuticos. Além disso, pessoas físicas também podem contribuir de forma espontânea. A instituição disponibiliza frequentemente um QR Code em suas redes sociais para doações de qualquer valor, fundamentais para a manutenção dos trabalhos e a realização de eventos beneficentes, como rifas e almoços. Ouça a entrevista completa: 

Ouça a reportagem de Ronaldo Araújo sobre a entrega do selo pela Universidade Federal de Sergipe em parceria com a Defensoria Pública do Estado.

Sérgio Martins é jornalista e crítico musical. Ele apresenta a coluna Conversas Musicais às 3ªs, 8h, no Jornal Eldorado.See omnystudio.com/listener for privacy information.

⁠⁠Enio Augusto⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ e ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Marcos Buosi⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ trazem as notícias do mundo da corrida com os comentários, informações, opiniões e análises mais pertinentes, peculiares e inesperadas no Redação PFC. Escute, informe-se e divirta-se.⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠SEJA MEMBRO DO CANAL!!!⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

#tocadodragao #2026 #podcast #ouvintes #emails #comentarios #comedia #discussao #videogamesEpisódio de hoje: Mais um selo do nosso programa especial, dedicados a vocês nos incríveis ouvintes! Bora pra mias um Toca Postal!ENTRE NA COMUNIDADE DO TOCA! ⁠⁠https://cesber.wixsite.com/tocadodragao⁠⁠REDES SOCIAIS E MUITO MAIS!https://beacons.ai/tocadodragaoFAÇA SUA DOAÇÃO #APOIE a TOCA a partir de R$ 10,00/ mês - Estamos no Apoia.se!https://apoia.se/atocadodragaoDOADORES DE MAIO/2026 PAULO DEROS ELVE, BRUNO BRAZ, RODRIGO SILVA, MARCIA REGINA BERNARDES, MASON YEON, PAULA GESTAL, GABRIEL SCHADE, LELE DANTAS, JOÃO PANDA, CEZAR AUGUSTO, ANTHONY MARTINS, ANDRIA SEDREZ, BRENDA NASCIMENTO, MISTER DOVAH, LENHORMAR, VICTOR FERNANDESAgradecemos aos Inscritos do Podcast que fizeram suas doações pelo PICPAY nosso e-mail: tocadodragaopodcast@gmail.comGRUPO DO TELEGRAM https://t.me/+fn75BRye8sY2NDExGRUPO DO WHATSAPPhttps://chat.whatsapp.com/KUtDsVnnv7w6hcseloXqCQCASTERS NESSE EPISÓDIO: Richard (Ricky, O Bardo) e Rodrigo Silva (Capivara Molhada)MÚSICAS ORIGINAIS DO TOCA #Compositor: Caio Varalta / Tema do Podcast: "Entrando na Toca" - Todos os Direitos Reservados

"Porque todos os que são guiados pelo Espírito de Deus, esses são filhos de Deus.Porque não recebestes o espírito de escravidão, para outra vez estardes em temor, mas recebestes o Espírito de adoção de filhos, pelo qual clamamos: Aba, Pai.O próprio Espírito testifica com o nosso espírito que somos filhos de Deus.E, se nós somos filhos, somos logo herdeiros também, verdadeiramente herdeiros de Deus, e co-herdeiros de Cristo: se é certo que com Ele padecemos, para que também com Ele sejamos glorificados." Romanos 8:14-17

A Maratona do Rio agora é selo elite da World Athletics, um passo importante dos vários que ela vai ter que dar tornar realidade o sonho de se tornar Major. Vamos conversar sobre isso?Nossos cupons e links - https://cnoar.run/cuponsO Corrida no Ar News é produzido diariamente e postado por volta das 6 da manhã.

(00:00:00) Xadrez Verbal #459 Emirados Árabes saem da OPEP (00:06:55) Giro de Notícias #01 (00:27:30) Coluna Aberta: entrevista com João Paulo Charleaux (01:14:05) Efemérides: A Semana na História (01:19:55) Match: Oriente Médio (02:13:00) Xeque: América Latina (03:00:15) Gambito da Dama: Brasil "queridinho" dos investidores (03:07:15) Giro de Notícias #02 (03:35:15) Peões da Semana (03:37:50) Sétimo Selo (03:50:50) Música de Encerramento Recebemos o analista político, escritor e jornalista João Paulo Charleaux para um papo sobre seu novo livro "As Regras da Guerra". Também repercutimos a saída dos Emirados Árabes Unidos da Organização dos Países Exportadores de Petróleo e os impactos desta decisão para o bloco comercial. No mais, demos aquele tradicional pião pela nossa quebrada latino-americana, com destaque para a captura do principal narcotraficante do CJNG, após a morte do Mencho, e também o maior ataque terrorista da Colômbia em décadas, faltando quase um mês para as eleições presidenciais no país vizinho.#PubliAlura Aprenda tecnologia com a Alura com nosso desconto: https://alura.tv/xadrezverbalUse o cupom XADREZVERBAL na Academia Guhan de Mandarim: https://academiaguhan.com.br/Conheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

A presidente da Associação de Pais e Amigos de Autistas do Município de Lauro Müller (AMA/LM), Fabiana Cataneo, e a assistente social Teresinha Ribeiro, participaram nesta quarta-feira (29) de entrevista no programa Cruz de Malta Notícias para lançar oficialmente o projeto “Selo Empresa Amiga do Autista” e apresentar seus principais objetivos à comunidade. A iniciativa tem como foco estabelecer parcerias com empresas locais para garantir o atendimento multidisciplinar de crianças com Transtorno do Espectro Autista (TEA), especialmente aquelas em situação de vulnerabilidade social. A proposta funciona por meio de um sistema de “adoção solidária”, no qual empresas tornam-se mantenedoras diretas dos serviços terapêuticos oferecidos pela instituição. De acordo com a AMA/LM, o projeto surge diante da necessidade urgente de ampliar o acesso a tratamentos especializados. Atualmente, muitas famílias enfrentam dificuldades financeiras para custear terapias essenciais, o que pode comprometer o desenvolvimento das crianças e gerar sobrecarga familiar. Além disso, há uma demanda reprimida no município, com 19 crianças aguardando na fila por atendimento. O objetivo geral do programa é garantir assistência gratuita e contínua para 45 crianças já atendidas pela entidade, além de possibilitar a inclusão gradual daqueles que ainda aguardam por acompanhamento especializado. Entre os objetivos específicos estão a manutenção das terapias para evitar retrocessos cognitivos, o suporte psicossocial às famílias e a promoção da conscientização sobre o autismo no ambiente corporativo. As empresas interessadas podem aderir ao projeto por meio de cotas de apadrinhamento, com investimento mensal por criança atendida. No modelo “Selo Prata”, é garantido o atendimento de três crianças, com cobertura de sessões em áreas como psicologia, fonoaudiologia e pedagogia, além de acompanhamento social das famílias. Já o “Selo Ouro” prevê o apadrinhamento de seis crianças. Com o selo, as empresas participantes passam a ser reconhecidas pelo compromisso com a inclusão social e o apoio ao desenvolvimento de crianças com TEA no município. Ouça a entrevista completa:

O que acontece quando uma advogada, pianista clássica de alma, decide largar tudo e virar DJ? Não é só uma história de coragem. É um estudo de caso de branding do mais puro que já trouxe aqui no Branding Tudo Podcast.Nesse episódio eu converso com DJ Anne Louise: baiana, formada em Direito pela Federal da Bahia, ex-músico de banda clássica, e hoje uma das DJs de tribal mais reconhecidas do mundo, tocando em mais de 30 países e quatro continentes. Tudo isso construído a partir de autenticidade, propósito e uma clareza de marca que muita empresa grande não tem.A gente mergulha fundo em como a Anne construiu a marca Anne Louise como uma terceira pessoa, separando a pessoa física da empresa. Como surgiu o manifesto "Missionária da Felicidade" dentro de um avião, no auge de um burnout que ela nem sabia que estava vivendo. Como ela usou o YouTube durante a pandemia pra conquistar o mercado asiático sem nunca ter ido buscar as datas: são as datas que vêm até ela.Se você está construindo uma marca pessoal ou quer entender o que é autenticidade de verdade aplicada ao negócio, esse episódio foi feito pra você.

(00:00:00) Xadrez Verbal #458 Sob a Sombra da Suástica (00:07:30) Giro de Notícias #01 (00:20:55) Coluna Aberta #01: Sob a sombra da suástica, com Fran Becher (01:19:35) Efemérides: A Semana na História (01:23:10) Coluna Aberta #02: Oriente Médio (02:12:10) Match: América Latina (03:36:15) Xeque: Europa (03:50:15) Giro de Notícias #02 (03:57:30) Peões da Semana (03:58:50) Sétimo Selo (04:08:40) Música de Encerramento Recebemos a historiadora e professora Fran Becher para falar sobre a França na Segunda Guerra Mundial e a infância e juventude neste contexto.Também seguimos acompanhando as negociações entre EUA e Irã, além da crescente crise eleitoral no Peru e a vitória de Rumen Radev na Bulgária.#PubliAlura Aprenda tecnologia com a Alura com nosso desconto: https://alura.tv/xadrezverbalCuide de sua saúde mental com a Psicólogos Brasil: https://www.psicologosbr.com/Conheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Compre a regata oficial do Corrida no ArMasculina - https://cnoar.run/RegataMasculinaCNAFeminina - https://cnoar.run/RegataFemininaCNAFaça camisetas, regatas, sacochilas e muito mais com a Ecco Bolsas e Camisetas - https://eccobolsas.com.br/Participe da Asics Run Challenge (use o cupom CORRIDANOAR10 para ter 10% de desconto na inscrição) - https://cnoar.run/Runchallenge10 Milhas Garoto recebe selo da World Athletics, Meia de Balneario Camburiu também, só que um selo é diferente do outro e vou explicar pra vocês; Maratona do Paraná recebeu o selo ouro da CBAt e só falta a de Aracaju pra deixar o circuito Brasil Gigante "ourificado"; Apple virou parceira da MAratona de Londres. Nossos links - https://linktr.ee/corridanoarO Corrida no Ar News é produzido diariamente e postado por volta das 6 da manhã.

O SELO DA LIBERDADE REJEITADA - 12/04/2026 - PR. RUBENS MARTIMSeja muito bem-vindo!Neste Podcast você vai encontrar tempos preciosos de orações e pregações inspiradas pelo Espírito Santo de Deus, o único capaz de transformar os corações por inteiro, conduzindo para uma profunda intimidade com o Pai e equipando sua vida para falar do amor de Jesus a todos os povos.Que o fogo do Espírito Santo queime em seu coração em cada vídeo/aúdio, despertando sua vida para o chamado de Deus para estes últimos dias.Inscreva-se no canal e compartilhe para que mais pessoas sejam abençoadas pela pregação do evangelho do Reino de Jesus Cristo.Leia a bíblia!https://www.instagram.com/casa.comunidadecrista/https://www.instagram.com/rubens_martim/https://www.instagram.com/falaprapsico.julianamartins/

(00:00:00) Xadrez Verbal #456 Cessar-fogo no Irã (00:06:30) Giro de Notícias #01 (00:27:25) Coluna Aberta: Oriente Médio (01:45:10) Efemérides: A Semana na História (01:53:45) Match: América Latina (02:56:55) Xeque: Europa (03:30:50) Gambito da Dama: PIB americano e seus componentes (03:38:45) Giro de Notícias #02 (03:50:20) Peões da Semana (03:52:55) Sétimo Selo (04:07:20) Música de Encerramento Analisamos o frágil cessar-fogo anunciado entre EUA, Israel e Irã, mediado pelo Paquistão, além de outras notícias do sempre complicado tabuleiro do Oriente Médio.Recebemos novamente Fernanda Simas, agora integrada ao time do XV, para dar aquele tradicional pião pela nossa quebrada latino-americana, com destaque para as prévias eleitorais no Peru.No mais, demos uma volta pelo Velho Continente, também em ritmo eleitoral na Hungria, em um pleito que diz muito em relação ao futuro da União Europeia.#PubliAlura Aprenda tecnologia com a Alura com nosso desconto: https://alura.tv/xadrezverbalCuide de sua saúde mental com a Psicólogos Brasil: https://www.psicologosbr.com/Conheça a Carta Global de Fernanda Simas: https://www.cartaglobal.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Znate li ko je čovek koji je vratio Harija Potera u Srbiju, a čiji je otac ispisao najlepše stihove našeg detinjstva? Gost 361. epizode je Dobrivoje Lale Erić, čovek impresivne karijere i neverovatne životne priče. Kroz topao i nostalgičan razgovor sa Ivanom, Lale nas vodi na putovanje od romantičnog odrastanja u Bariču i specifičnog odnosa sa ocem, čuvenim pesnikom Dobricom Erićem, do urbanih beogradskih priča iz devedesetih. Otkriva nam kako je teklo njegovo obrazovanje, od ljubavi prema arheologiji, preko uloge Istraživačke stanice Petnica kao oaze slobode, do studija istorije umetnosti. Saznajemo kako je Srbija dobila prava na Harija Potera upravo zahvaljujući njemu, ali i kako već 15 godina uspešno spaja nauku, umetnost i društvo kroz rad u Centru za promociju nauke. Ovo je priča o radoznalosti, nepresušnoj energiji i želji za igrom koja menja svet oko nas. O čemu smo pričali: - Najava razgovora - Početak razgovora - Smilies pitanje: Šta je hteo da bude kad poraste? - Okruženje tokom odrastanja - Kako je biti sin Dobrice Erića - Selo i školski dani - Banovo brdo nekada - Gde na studije - Fakultetski dani - Život nakon revolucije - Pravac posle faksa - Iz izdavaštva u muzej - Centar za promociju nauke - Problemi posle Kovida - Zaključak razgovora Realizaciju ove epizode podržali su naši prijatelji i sponzori: - Epson Srbija - https://www.epson.rs - Orion telekom - https://oriontelekom.rs - Smilies - https://smilies.rs Hvala na poverenju i podršci! Podržite nas na BuyMeACoffee: https://bit.ly/3uSBmoa Pročitajte transkript ove epizode: https://bit.ly/3Pr43b5 Posetite naš sajt i prijavite se na našu mailing listu: http://bit.ly/2LUKSBG Prijavite se na naš YouTube kanal: http://bit.ly/2Rgnu7o Pratite Pojačalo na društvenim mrežama: FB: https://www.facebook.com/PojacaloRS/ IG: https://www.instagram.com/pojacalo.rs/ X: https://x.com/PojacaloRS LN: https://www.linkedin.com/company/pojacalo TikTok: https://www.tiktok.com/@pojacalo.rs

Compre a regata oficial do Corrida no ArMasculina - https://cnoar.run/RegataMasculinaCNAFeminina - https://cnoar.run/RegataFemininaCNAFaça camisetas, regatas, sacochilas e muito mais com a Ecco Bolsas e Camisetas - https://eccobolsas.com.br/Recorde provas brasileiras com selo da World Athltetics; Maratona de Boston aumentou as ondas de largada; Maratona e Meia Internacional de Belo Horizonte recebeu selo ouro da CBAt; Forcell agora é o gel oficial da MAratona de Porto Alegre; Lupo vai expandir circuito de corrida por todas as regiões do Brasil.Nossos links - https://linktr.ee/corridanoarO Corrida no Ar News é produzido diariamente e postado por volta das 6 da manhã.

(00:00:00) Xadrez Verbal #453 EUA x Venezuela... também no Beisebol (00:05:30) Giro de Notícias #01 (00:23:40) Coluna Aberta: Oriente Médio (01:31:25) Efemérides: A Semana na História (01:41:00) Match: Giro Esportivo e Velho Continente (02:26:00) Xeque: América Latina (03:57:40) Giro de Notícias #02 (04:14:25) Peões da Semana (04:15:55) Sétimo Selo (04:29:35) Música de Encerramento Infelizmente, mais uma semana de guerra no Irã! Atualizamos e analisamos as últimas notícias sobre o conflito e também sobre o Oriente Médio.Repercutimos a mudança no comando militar da Venezuela no nosso tradicional pião pela quebrada latino-americana.No mais, fizemos outro giro esportivo e recebemos novamente o Ubiratan Leal, amigo de longa data, para comentar sobre a conquista do Mundial de Beisebol pela seleção venezuelana sobre os EUA, em Miami, dois meses após a abdução de Nicolás Maduro.Aprenda tecnologia com a Alura com nosso desconto! #PubliAlura: https://alura.tv/xadrezverbalUse o cupom XADREZVERBAL na Academia Guhan de Mandarim: https://academiaguhan.com.br/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Nueva cita con la ciudadanía en Díselo a tu alcalde, que en este primer programa de 2026 mantiene su nueva dinámica para atender las demandas vecinales distrito a distrito. En esta ocasión, el turno es para el Distrito 5 de Bilbao, con especial atención a Atxuri, cuyos vecinos y vecinas trasladan sus preguntas en Hoy por Hoy Bilbao-Bizkaia al alcalde, Juan Mari Aburto. El distrito engloba los barrios de Atxuri, Bilbao La Vieja, Casco Viejo, Iturrialde, La Peña, Miribilla, San Adrián, San Francisco, Buia, Solokoetxe y Zabala, que irán pasando por el programa en próximas ediciones

Tema SELO DA VITÓRIA: FAMÍLIAS RESTAURADAS. Domingo - Manhã 18/01/2026 Pregador: Bp. André Barbosa. Este sermão traz uma palavra profunda sobre recomeço, identidade e restauração, baseada na promessa de Deus a Zorobabel. Aprendemos que Deus trabalha com aliança e selo, não com sentimentos passageiros. Muitas vezes, o desânimo espiritual nos paralisa e chamamos medo de prudência, adiando aquilo que Deus já determinou. O Senhor nos lembra que o tempo é hoje, que Ele nos escolheu e nos selou com autoridade espiritual para reconstruir não apenas estruturas, mas destinos, famílias e gerações. Mesmo quando tudo parece em ruínas, Deus se levanta como nossa proteção e garante que a glória do que Ele fará será maior do que o passado. Uma mensagem poderosa para quem precisa romper com a procrastinação espiritual e viver um novo tempo de restauração total em Deus. Aula 05 Módulo – 68 Seminário: Construindo uma casa sobre a Rocha

Tema – O Ano do Selo de Deus. Passagem de Ano 2025/26 Pregador : Ap. Miguel Ângelo Naquele dia, diz o SENHOR dos Exércitos, tomar-te-ei, ó Zorobabel, filho de Salatiel, servo meu, diz o SENHOR, e te farei como um anel de selar, porque te escolhi, diz o SENHOR dos Exércitos. (Ageu2:23) Neste culto profético de passagem de ano, somos conduzidos a uma revelação poderosa para atravessar 2026 com fé, segurança e autoridade espiritual. À luz das Escrituras e do ministério dos profetas Ageu e Zacarias, Deus declara que não despreza os pequenos começos, pois Ele já vê o fim glorioso preparado para os Seus escolhidos. Mesmo diante de um ano anunciado como turbulento, o Senhor libera uma palavra de direção, governo espiritual e reconstrução. 2026 é declarado como o ano do selo de Deus, um tempo em que o povo será estabelecido como anel de selar, símbolo de autoridade, aliança e legitimação espiritual. Deus promete restaurar vidas, famílias, propósitos e valores que foram interrompidos. Este é um chamado para viver acima do medo, firmados na Palavra viva, caminhando com clareza, fé e esperança, sabendo que o Senhor irá abalar céus e terra para cumprir Suas promessas.

Amanhã é dia de Volta da Pampulha e Maratona de Valência. E mano, e a NB42k de PoA, que subiu para selo elite da World Athletics?Assine a nossa newsletter e fique sempre bem informado - https://corridanoar.com/newsletterNossos links - https://linktr.ee/corridanoarO Corrida no Ar News é produzido diariamente e postado por volta das 6 da manhã.

(00:00:00) Xadrez Verbal #443 Negociações sobre a Ucrânia (00:02:00) Giro de Notícias #01 (00:14:00) Coluna Aberta: Proposta de paz para a Guerra na Ucrânia (01:18:05) Efemérides: A Semana na Históri (01:25:15) Match: América Latina (02:28:15) Xeque: G20 (03:33:20) Giro de Notícias #02 (03:44:20) Peões da Semana (03:47:05) Sétimo Selo (03:55:20) Música de Encerramento Analisamos a mais nova rodada de negociações sobre um fim para a guerra na Ucrânia.Também demos aquele tradicional pião pela nossa quebrada latino-americana, com crise palaciana na Bolívia e a prévia das eleições em Honduras.Finalmente, repercutimos a 20ª reunião de cúpula do G20, realizada em Joanesburgo, na África do Sul.Aproveite a Black November da Alura: https://alura.tv/xadrezverbalAgende uma reunião com a Rio Claro Investimentos: https://rioclaro.com.br/xadrez-verbal/Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Palestra realizada no dia 12/10/2024 durante a 68ª Feira do Livro Espírita de Uberaba/MG, no Memorial Chico Xavier.

(00:00:00) Xadrez Verbal #439 Happy Birthday, Senhor Presidente (00:03:40) Giro de Notícias #01 (00:22:00) Coluna Aberta: Oriente Médio (00:42:15) Efemérides: A Semana na História (00:50:50) Match: América Latina (02:14:35) Giro de Notícias #02 (02:25:05) Xeque: Bacia do Pacífico (03:02:35) Gambito da Dama: Juros, inflação e câmbio (03:10:30) Giro de Notícias #03 (03:17:45) Peões da Semana (03:19:15) Sétimo Selo (03:27:55) Música de Encerramento Lula se encontrou com Trump na Malásia e repercutiremos este e outros encontros bilaterais durante a 47ª Cúpula da ASEAN, além de outras notícias da bacia do Pacífico.Também observamos o movimento das peças no sempre complicado tabuleiro do Oriente Médio, destacando as reações aos planos israelenses em relação à Cisjordânia.No mais, demos aquele tradicional pião pela nossa quebrada latino-americana, com novo presidente na Bolívia e a vitória governista nas eleições legislativas argentinas.Participe da Imersão IA Alura com Google Gemini: https://alura.tv/xadrezverbal-imersao-dev-2025Use o cupom XADREZVERBAL na Academia Guhan de Mandarim: https://academiaguhan.com.br/Use o cupom XADREZVERBAL50 para ter 50% de desconto no plano de saúde do seu pet na Petlove: https://saude.petlove.com.br/?promocao=influencer&utm_source=spotify&utm_medium=influencer&utm_campaign=xadrezverbalCampanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Da Serra da Barriga ao Congresso Nacional, Lélia Gonzalez chega ao topo de sua luta no fim dos anos 1980. Mas enquanto conquista vitórias históricas, o corpo começa a cobrar o preço de décadas de luta.APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta  ou orelo.cc/historiapretaChave Pix: historiapreta@gmail.comLOJAAcesse loja.historiapreta.com.br e vista nossa história.FICHA TÉCNICAPesquisa e roteiro: Thiago AndréApresentação: Thiago AndréRedes sociais e Gerência da comunidade: Carolina FerreiraIdentidade Visual: Raimundo BrittoNos siga nas redes sociais no twitter @historiapreta e no Instagram @historia_pretaBIBLIOGRAFIAAlex Ratts e Flávia Rios. Lélia Gonzalez. Selo negro, 2014.Sueli Carneiro. Lélia Gonzalez: um retrato. Zahar, 2024                                           APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta OU orelo.cc/historiapretaChave Pix: historiapreta@gmail.com

(00:00:00) Xadrez Verbal #438 Protestos em Madagascar (00:04:40) Giro de Notícias #01 (00:24:05) Coluna Aberta #01: entrevista com o embaixador Alessandro Candeas (01:35:40) Coluna Aberta #02: Oriente Médio (02:19:55) Efemérides: A Semana na História (02:24:50) Match: América Latina (03:38:30) Xeque: África (04:14:05) Gambito da Dama: Nobel de Economia (04:23:50) Giro de Notícias #02 (04:42:45) Peões da Semana (04:43:50) Sétimo Selo (04:54:05) Música de Encerramento Realizamos uma entrevista exclusiva com o embaixador Alessandro Candeas, embaixador brasileiro na Palestina de 2020 a 2024.No mais, a bandeira do pirata que estica de One Piece continua presente em protestos pelo Mundo, dessa vez em Madagascar e trataremos da conjuntura política da ilha africana, além da queda do presidente Andry Rajoelina.Também demos aquele tradicional pião pela nossa quebrada latino-americana, com destaque para o pedido de demissão do chefe militar nos EUA na região e os protestos no Equador e Peru.Por fim, a professora Vivian Almeida repercute os laureados com o Nobel de Economia!Aproveite o Guia do Mochileiro Tech da Alura: https://alura.tv/xadrezverbal-guia-2025Campanha e comunicado sobre nosso amigo Pirulla: https://www.pirulla.com.br/

Esse bate-papo foi gravado em junho de 2025, a Carla Douglass e a Fabi Gutierrez receberam o Morris Livitak. Ele é fundador e CEO da Maturi, plataforma líder no Brasil para profissionais 50+. Ele também é colunista da Fast Company Brasil e leciona em cursos de pós-graduação em gerontologia nos Hospitais Albert Einstein e Sírio Libanês, e no curso de Diversidade da Escola de Negócios da ABERJE.Somos fãs dele e do trabalho que ele e seu time realizam com a Maturi. O Morris retorna ao Bate-Papo Carlotas, nosso primeiro papo foi em 2019 (!). Dessa vez falamos sobre a Diversidade Geracional, sobre o Selo que eles desenvolveram, o Festival Maturi e como cada um de nós pode se envolver nessa causa.Visite a Manturi: https://www.maturi.com.br/

In this heartwarming catch-up episode, your hosts Shannon Shine and Selomon sit down to share what they've been up to behind the scenes—from art and books to school and spiritual work.

My laboratory has made major contributions to our understanding of non-canonical functions for protein kinases by discovering diverse and unanticipated biochemical activities that are performed by this protein superfamily. Protein kinases have been studied for decades and play important roles in many physiological and pathological processes. The textbook view is that these enzymes transfer phosphate from ATP to protein substrates in a process termed phosphorylation. However, my laboratory has overturned this paradigm by discovering new catalytic activities of atypical protein kinases and pseudokinases. For example, we discovered that the predicted pseduokinases SelO, SidJ and nsp12 catalyze AMPylation, polyglutamylation and mRNA capping, respectively. These results have revealed important new insights into the cellular response to oxidative stress and the pathogenic mechanisms employed by bacterial and viral pathogens. Our work on eukaryotic, prokaryotic, and viral kinases has exposed the catalytic versatility of the protein kinase fold and suggests that pseudoenzymes should be analyzed for alternative catalytic activities. In this lecture, I will present our recent discovery of kinases responsible for isoprenoid salvage.

Depois do protesto no Teatro Municipal, o Movimento Negro percebe a necessidade de expandir para além de São Paulo e se espalhar pelo Brasil. Lélia Gonzalez assume a responsabilidade de viajar pelo país e formar novos militantes, mas encontra desafios internos.APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta  ou orelo.cc/historiapretaChave Pix: historiapreta@gmail.comLOJAAcesse loja.historiapreta.com.br e vista nossa história.FICHA TÉCNICAPesquisa e roteiro: Thiago AndréApresentação: Thiago AndréRedes sociais e Gerência da comunidade: Carolina FerreiraIdentidade Visual: Raimundo BrittoNos siga nas redes sociais no twitter @historiapreta e no Instagram @historia_pretaBIBLIOGRAFIAAlex Ratts e Flávia Rios. Lélia Gonzalez. Selo negro, 2014.Sueli Carneiro. Lélia Gonzalez: um retrato. Zahar, 2024                                           APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta OU orelo.cc/historiapretaChave Pix: historiapreta@gmail.com

Nous sommes le 22 mars 1917, près de Saint-Pétersbourg. C'est là, que le soir venu, sur ordre du nouveau gouvernement révolutionnaire, on exhume un corps qui a été enterré trois mois plus tôt, au début de l'année, dans une chapelle en construction, près du palais de Tsarskoïe Selo, le Versailles russe. Ce corps, on va le bruler et en disperser les cendres dans la forêt avoisinante (la légende dit que seul le cercueil s'enflamma, que la dépouille resta intacte). Ce corps, celui de Raspoutine. Avec Vincent Genin, historien sujets traités : Raspoutine, Saint-Pétersbourg, Tsarskoïe Selo, révolutionnaire, Merci pour votre écoute Un Jour dans l'Histoire, c'est également en direct tous les jours de la semaine de 13h15 à 14h30 sur www.rtbf.be/lapremiere Retrouvez tous les épisodes d'Un Jour dans l'Histoire sur notre plateforme Auvio.be :https://auvio.rtbf.be/emission/5936 Intéressés par l'histoire ? Vous pourriez également aimer nos autres podcasts : L'Histoire Continue: https://audmns.com/kSbpELwL'heure H : https://audmns.com/YagLLiKEt sa version à écouter en famille : La Mini Heure H https://audmns.com/YagLLiKAinsi que nos séries historiques :Chili, le Pays de mes Histoires : https://audmns.com/XHbnevhD-Day : https://audmns.com/JWRdPYIJoséphine Baker : https://audmns.com/wCfhoEwLa folle histoire de l'aviation : https://audmns.com/xAWjyWCLes Jeux Olympiques, l'étonnant miroir de notre Histoire : https://audmns.com/ZEIihzZMarguerite, la Voix d'une Résistante : https://audmns.com/zFDehnENapoléon, le crépuscule de l'Aigle : https://audmns.com/DcdnIUnUn Jour dans le Sport : https://audmns.com/xXlkHMHSous le sable des Pyramides : https://audmns.com/rXfVppvN'oubliez pas de vous y abonner pour ne rien manquer.Et si vous avez apprécié ce podcast, n'hésitez pas à nous donner des étoiles ou des commentaires, cela nous aide à le faire connaître plus largement. Hébergé par Audiomeans. Visitez audiomeans.fr/politique-de-confidentialite pour plus d'informations.

Em plena ditadura, Lélia Gonzalez ajuda a fundar o Movimento Negro Unificado. Mobiliza lideranças, ocupa as ruas e transforma protesto em organização. Surgia ali um novo capítulo da luta antirracista no Brasil.APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta  ou orelo.cc/historiapretaChave Pix: historiapreta@gmail.comLOJAAcesse loja.historiapreta.com.br e vista nossa história.FICHA TÉCNICAPesquisa e roteiro: Thiago AndréApresentação: Thiago AndréRedes sociais e Gerência da comunidade: Carolina FerreiraIdentidade Visual: Raimundo BrittoNos siga nas redes sociais no twitter @historiapreta e no Instagram @historia_pretaBIBLIOGRAFIAAlex Ratts e Flávia Rios. Lélia Gonzalez. Selo negro, 2014.Sueli Carneiro. Lélia Gonzalez: um retrato. Zahar, 2024                                                       APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta OU orelo.cc/historiapretaChave Pix: historiapreta@gmail.com

A perda do marido leva Lélia Gonzalez a um reencontro profundo com suas raízes negras. Começava ali uma revolução interior que logo transbordaria para o mundo.APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta  ou orelo.cc/historiapretaChave Pix: historiapreta@gmail.comLOJAAcesse loja.historiapreta.com.br e vista nossa história.FICHA TÉCNICAPesquisa e roteiro: Thiago AndréApresentação: Thiago AndréRedes sociais e Gerência da comunidade: Carolina FerreiraIdentidade Visual: Raimundo BrittoNos siga nas redes sociais no twitter @historiapreta e no Instagram @historia_pretaBIBLIOGRAFIAAlex Ratts e Flávia Rios. Lélia Gonzalez. Selo negro, 2014.Sueli Carneiro. Lélia Gonzalez: um retrato. Zahar, 2024                                                                                                                                                  APOIEEste episódio só foi possível graças a contribuição generosa de nossos apoiadores. Se você gosta do nosso trabalho, considere nos apoiar em apoia.se/historiapreta OU orelo.cc/historiapretaChave Pix: historiapreta@gmail.com