Podcasts about Apollo

From rewriting Google's search stack in the early 2000s to reviving sparse trillion-parameter models and co-designing TPUs with frontier ML research, Jeff Dean has quietly shaped nearly every layer of the modern AI stack. As Chief AI Scientist at Google and a driving force behind Gemini, Jeff has lived through multiple scaling revolutions from CPUs and sharded indices to multimodal models that reason across text, video, and code.Jeff joins us to unpack what it really means to “own the Pareto frontier,” why distillation is the engine behind every Flash model breakthrough, how energy (in picojoules) not FLOPs is becoming the true bottleneck, what it was like leading the charge to unify all of Google's AI teams, and why the next leap won't come from bigger context windows alone, but from systems that give the illusion of attending to trillions of tokens.We discuss:* Jeff's early neural net thesis in 1990: parallel training before it was cool, why he believed scaling would win decades early, and the “bigger model, more data, better results” mantra that held for 15 years* The evolution of Google Search: sharding, moving the entire index into memory in 2001, softening query semantics pre-LLMs, and why retrieval pipelines already resemble modern LLM systems* Pareto frontier strategy: why you need both frontier “Pro” models and low-latency “Flash” models, and how distillation lets smaller models surpass prior generations* Distillation deep dive: ensembles → compression → logits as soft supervision, and why you need the biggest model to make the smallest one good* Latency as a first-class objective: why 10–50x lower latency changes UX entirely, and how future reasoning workloads will demand 10,000 tokens/sec* Energy-based thinking: picojoules per bit, why moving data costs 1000x more than a multiply, batching through the lens of energy, and speculative decoding as amortization* TPU co-design: predicting ML workloads 2–6 years out, speculative hardware features, precision reduction, sparsity, and the constant feedback loop between model architecture and silicon* Sparse models and “outrageously large” networks: trillions of parameters with 1–5% activation, and why sparsity was always the right abstraction* Unified vs. specialized models: abandoning symbolic systems, why general multimodal models tend to dominate vertical silos, and when vertical fine-tuning still makes sense* Long context and the illusion of scale: beyond needle-in-a-haystack benchmarks toward systems that narrow trillions of tokens to 117 relevant documents* Personalized AI: attending to your emails, photos, and documents (with permission), and why retrieval + reasoning will unlock deeply personal assistants* Coding agents: 50 AI interns, crisp specifications as a new core skill, and how ultra-low latency will reshape human–agent collaboration* Why ideas still matter: transformers, sparsity, RL, hardware, systems — scaling wasn't blind; the pieces had to multiply togetherShow Notes:* Gemma 3 Paper* Gemma 3* Gemini 2.5 Report* Jeff Dean's “Software Engineering Advice fromBuilding Large-Scale Distributed Systems” Presentation (with Back of the Envelope Calculations)* Latency Numbers Every Programmer Should Know by Jeff Dean* The Jeff Dean Facts* Jeff Dean Google Bio* Jeff Dean on “Important AI Trends” @Stanford AI Club* Jeff Dean & Noam Shazeer — 25 years at Google (Dwarkesh)—Jeff Dean* LinkedIn: https://www.linkedin.com/in/jeff-dean-8b212555* X: https://x.com/jeffdeanGoogle* https://google.com* https://deepmind.googleFull Video EpisodeTimestamps00:00:04 — Introduction: Alessio & Swyx welcome Jeff Dean, chief AI scientist at Google, to the Latent Space podcast00:00:30 — Owning the Pareto Frontier & balancing frontier vs low-latency models00:01:31 — Frontier models vs Flash models + role of distillation00:03:52 — History of distillation and its original motivation00:05:09 — Distillation's role in modern model scaling00:07:02 — Model hierarchy (Flash, Pro, Ultra) and distillation sources00:07:46 — Flash model economics & wide deployment00:08:10 — Latency importance for complex tasks00:09:19 — Saturation of some tasks and future frontier tasks00:11:26 — On benchmarks, public vs internal00:12:53 — Example long-context benchmarks & limitations00:15:01 — Long-context goals: attending to trillions of tokens00:16:26 — Realistic use cases beyond pure language00:18:04 — Multimodal reasoning and non-text modalities00:19:05 — Importance of vision & motion modalities00:20:11 — Video understanding example (extracting structured info)00:20:47 — Search ranking analogy for LLM retrieval00:23:08 — LLM representations vs keyword search00:24:06 — Early Google search evolution & in-memory index00:26:47 — Design principles for scalable systems00:28:55 — Real-time index updates & recrawl strategies00:30:06 — Classic “Latency numbers every programmer should know”00:32:09 — Cost of memory vs compute and energy emphasis00:34:33 — TPUs & hardware trade-offs for serving models00:35:57 — TPU design decisions & co-design with ML00:38:06 — Adapting model architecture to hardware00:39:50 — Alternatives: energy-based models, speculative decoding00:42:21 — Open research directions: complex workflows, RL00:44:56 — Non-verifiable RL domains & model evaluation00:46:13 — Transition away from symbolic systems toward unified LLMs00:47:59 — Unified models vs specialized ones00:50:38 — Knowledge vs reasoning & retrieval + reasoning00:52:24 — Vertical model specialization & modules00:55:21 — Token count considerations for vertical domains00:56:09 — Low resource languages & contextual learning00:59:22 — Origins: Dean's early neural network work01:10:07 — AI for coding & human–model interaction styles01:15:52 — Importance of crisp specification for coding agents01:19:23 — Prediction: personalized models & state retrieval01:22:36 — Token-per-second targets (10k+) and reasoning throughput01:23:20 — Episode conclusion and thanksTranscriptAlessio Fanelli [00:00:04]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, founder of Kernel Labs, and I'm joined by Swyx, editor of Latent Space. Shawn Wang [00:00:11]: Hello, hello. We're here in the studio with Jeff Dean, chief AI scientist at Google. Welcome. Thanks for having me. It's a bit surreal to have you in the studio. I've watched so many of your talks, and obviously your career has been super legendary. So, I mean, congrats. I think the first thing must be said, congrats on owning the Pareto Frontier.Jeff Dean [00:00:30]: Thank you, thank you. Pareto Frontiers are good. It's good to be out there.Shawn Wang [00:00:34]: Yeah, I mean, I think it's a combination of both. You have to own the Pareto Frontier. You have to have like frontier capability, but also efficiency, and then offer that range of models that people like to use. And, you know, some part of this was started because of your hardware work. Some part of that is your model work, and I'm sure there's lots of secret sauce that you guys have worked on cumulatively. But, like, it's really impressive to see it all come together in, like, this slittily advanced.Jeff Dean [00:01:04]: Yeah, yeah. I mean, I think, as you say, it's not just one thing. It's like a whole bunch of things up and down the stack. And, you know, all of those really combine to help make UNOS able to make highly capable large models, as well as, you know, software techniques to get those large model capabilities into much smaller, lighter weight models that are, you know, much more cost effective and lower latency, but still, you know, quite capable for their size. Yeah.Alessio Fanelli [00:01:31]: How much pressure do you have on, like, having the lower bound of the Pareto Frontier, too? I think, like, the new labs are always trying to push the top performance frontier because they need to raise more money and all of that. And you guys have billions of users. And I think initially when you worked on the CPU, you were thinking about, you know, if everybody that used Google, we use the voice model for, like, three minutes a day, they were like, you need to double your CPU number. Like, what's that discussion today at Google? Like, how do you prioritize frontier versus, like, we have to do this? How do we actually need to deploy it if we build it?Jeff Dean [00:02:03]: Yeah, I mean, I think we always want to have models that are at the frontier or pushing the frontier because I think that's where you see what capabilities now exist that didn't exist at the sort of slightly less capable last year's version or last six months ago version. At the same time, you know, we know those are going to be really useful for a bunch of use cases, but they're going to be a bit slower and a bit more expensive than people might like for a bunch of other broader models. So I think what we want to do is always have kind of a highly capable sort of affordable model that enables a whole bunch of, you know, lower latency use cases. People can use them for agentic coding much more readily and then have the high-end, you know, frontier model that is really useful for, you know, deep reasoning, you know, solving really complicated math problems, those kinds of things. And it's not that. One or the other is useful. They're both useful. So I think we'd like to do both. And also, you know, through distillation, which is a key technique for making the smaller models more capable, you know, you have to have the frontier model in order to then distill it into your smaller model. So it's not like an either or choice. You sort of need that in order to actually get a highly capable, more modest size model. Yeah.Alessio Fanelli [00:03:24]: I mean, you and Jeffrey came up with the solution in 2014.Jeff Dean [00:03:28]: Don't forget, L'Oreal Vinyls as well. Yeah, yeah.Alessio Fanelli [00:03:30]: A long time ago. But like, I'm curious how you think about the cycle of these ideas, even like, you know, sparse models and, you know, how do you reevaluate them? How do you think about in the next generation of model, what is worth revisiting? Like, yeah, they're just kind of like, you know, you worked on so many ideas that end up being influential, but like in the moment, they might not feel that way necessarily. Yeah.Jeff Dean [00:03:52]: I mean, I think distillation was originally motivated because we were seeing that we had a very large image data set at the time, you know, 300 million images that we could train on. And we were seeing that if you create specialists for different subsets of those image categories, you know, this one's going to be really good at sort of mammals, and this one's going to be really good at sort of indoor room scenes or whatever, and you can cluster those categories and train on an enriched stream of data after you do pre-training on a much broader set of images. You get much better performance. If you then treat that whole set of maybe 50 models you've trained as a large ensemble, but that's not a very practical thing to serve, right? So distillation really came about from the idea of, okay, what if we want to actually serve that and train all these independent sort of expert models and then squish it into something that actually fits in a form factor that you can actually serve? And that's, you know, not that different from what we're doing today. You know, often today we're instead of having an ensemble of 50 models. We're having a much larger scale model that we then distill into a much smaller scale model.Shawn Wang [00:05:09]: Yeah. A part of me also wonders if distillation also has a story with the RL revolution. So let me maybe try to articulate what I mean by that, which is you can, RL basically spikes models in a certain part of the distribution. And then you have to sort of, well, you can spike models, but usually sometimes... It might be lossy in other areas and it's kind of like an uneven technique, but you can probably distill it back and you can, I think that the sort of general dream is to be able to advance capabilities without regressing on anything else. And I think like that, that whole capability merging without loss, I feel like it's like, you know, some part of that should be a distillation process, but I can't quite articulate it. I haven't seen much papers about it.Jeff Dean [00:06:01]: Yeah, I mean, I tend to think of one of the key advantages of distillation is that you can have a much smaller model and you can have a very large, you know, training data set and you can get utility out of making many passes over that data set because you're now getting the logits from the much larger model in order to sort of coax the right behavior out of the smaller model that you wouldn't otherwise get with just the hard labels. And so, you know, I think that's what we've observed. Is you can get, you know, very close to your largest model performance with distillation approaches. And that seems to be, you know, a nice sweet spot for a lot of people because it enables us to kind of, for multiple Gemini generations now, we've been able to make the sort of flash version of the next generation as good or even substantially better than the previous generations pro. And I think we're going to keep trying to do that because that seems like a good trend to follow.Shawn Wang [00:07:02]: So, Dara asked, so it was the original map was Flash Pro and Ultra. Are you just sitting on Ultra and distilling from that? Is that like the mother load?Jeff Dean [00:07:12]: I mean, we have a lot of different kinds of models. Some are internal ones that are not necessarily meant to be released or served. Some are, you know, our pro scale model and we can distill from that as well into our Flash scale model. So I think, you know, it's an important set of capabilities to have and also inference time scaling. It can also be a useful thing to improve the capabilities of the model.Shawn Wang [00:07:35]: And yeah, yeah, cool. Yeah. And obviously, I think the economy of Flash is what led to the total dominance. I think the latest number is like 50 trillion tokens. I don't know. I mean, obviously, it's changing every day.Jeff Dean [00:07:46]: Yeah, yeah. But, you know, by market share, hopefully up.Shawn Wang [00:07:50]: No, I mean, there's no I mean, there's just the economics wise, like because Flash is so economical, like you can use it for everything. Like it's in Gmail now. It's in YouTube. Like it's yeah. It's in everything.Jeff Dean [00:08:02]: We're using it more in our search products of various AI mode reviews.Shawn Wang [00:08:05]: Oh, my God. Flash past the AI mode. Oh, my God. Yeah, that's yeah, I didn't even think about that.Jeff Dean [00:08:10]: I mean, I think one of the things that is quite nice about the Flash model is not only is it more affordable, it's also a lower latency. And I think latency is actually a pretty important characteristic for these models because we're going to want models to do much more complicated things that are going to involve, you know, generating many more tokens from when you ask the model to do so. So, you know, if you're going to ask the model to do something until it actually finishes what you ask it to do, because you're going to ask now, not just write me a for loop, but like write me a whole software package to do X or Y or Z. And so having low latency systems that can do that seems really important. And Flash is one direction, one way of doing that. You know, obviously our hardware platforms enable a bunch of interesting aspects of our, you know, serving stack as well, like TPUs, the interconnect between. Chips on the TPUs is actually quite, quite high performance and quite amenable to, for example, long context kind of attention operations, you know, having sparse models with lots of experts. These kinds of things really, really matter a lot in terms of how do you make them servable at scale.Alessio Fanelli [00:09:19]: Yeah. Does it feel like there's some breaking point for like the proto Flash distillation, kind of like one generation delayed? I almost think about almost like the capability as a. In certain tasks, like the pro model today is a saturated, some sort of task. So next generation, that same task will be saturated at the Flash price point. And I think for most of the things that people use models for at some point, the Flash model in two generation will be able to do basically everything. And how do you make it economical to like keep pushing the pro frontier when a lot of the population will be okay with the Flash model? I'm curious how you think about that.Jeff Dean [00:09:59]: I mean, I think that's true. If your distribution of what people are asking people, the models to do is stationary, right? But I think what often happens is as the models become more capable, people ask them to do more, right? So, I mean, I think this happens in my own usage. Like I used to try our models a year ago for some sort of coding task, and it was okay at some simpler things, but wouldn't do work very well for more complicated things. And since then, we've improved dramatically on the more complicated coding tasks. And now I'll ask it to do much more complicated things. And I think that's true, not just of coding, but of, you know, now, you know, can you analyze all the, you know, renewable energy deployments in the world and give me a report on solar panel deployment or whatever. That's a very complicated, you know, more complicated task than people would have asked a year ago. And so you are going to want more capable models to push the frontier in the absence of what people ask the models to do. And that also then gives us. Insight into, okay, where does the, where do things break down? How can we improve the model in these, these particular areas, uh, in order to sort of, um, make the next generation even better.Alessio Fanelli [00:11:11]: Yeah. Are there any benchmarks or like test sets they use internally? Because it's almost like the same benchmarks get reported every time. And it's like, all right, it's like 99 instead of 97. Like, how do you have to keep pushing the team internally to it? Or like, this is what we're building towards. Yeah.Jeff Dean [00:11:26]: I mean, I think. Benchmarks, particularly external ones that are publicly available. Have their utility, but they often kind of have a lifespan of utility where they're introduced and maybe they're quite hard for current models. You know, I, I like to think of the best kinds of benchmarks are ones where the initial scores are like 10 to 20 or 30%, maybe, but not higher. And then you can sort of work on improving that capability for, uh, whatever it is, the benchmark is trying to assess and get it up to like 80, 90%, whatever. I, I think once it hits kind of 95% or something, you get very diminishing returns from really focusing on that benchmark, cuz it's sort of, it's either the case that you've now achieved that capability, or there's also the issue of leakage in public data or very related kind of data being, being in your training data. Um, so we have a bunch of held out internal benchmarks that we really look at where we know that wasn't represented in the training data at all. There are capabilities that we want the model to have. Um, yeah. Yeah. Um, that it doesn't have now, and then we can work on, you know, assessing, you know, how do we make the model better at these kinds of things? Is it, we need different kind of data to train on that's more specialized for this particular kind of task. Do we need, um, you know, a bunch of, uh, you know, architectural improvements or some sort of, uh, model capability improvements, you know, what would help make that better?Shawn Wang [00:12:53]: Is there, is there such an example that you, uh, a benchmark inspired in architectural improvement? Like, uh, I'm just kind of. Jumping on that because you just.Jeff Dean [00:13:02]: Uh, I mean, I think some of the long context capability of the, of the Gemini models that came, I guess, first in 1.5 really were about looking at, okay, we want to have, um, you know,Shawn Wang [00:13:15]: immediately everyone jumped to like completely green charts of like, everyone had, I was like, how did everyone crack this at the same time? Right. Yeah. Yeah.Jeff Dean [00:13:23]: I mean, I think, um, and once you're set, I mean, as you say that needed single needle and a half. Hey, stack benchmark is really saturated for at least context links up to 1, 2 and K or something. Don't actually have, you know, much larger than 1, 2 and 8 K these days or two or something. We're trying to push the frontier of 1 million or 2 million context, which is good because I think there are a lot of use cases where. Yeah. You know, putting a thousand pages of text or putting, you know, multiple hour long videos and the context and then actually being able to make use of that as useful. Try to, to explore the über graduation are fairly large. But the single needle in a haystack benchmark is sort of saturated. So you really want more complicated, sort of multi-needle or more realistic, take all this content and produce this kind of answer from a long context that sort of better assesses what it is people really want to do with long context. Which is not just, you know, can you tell me the product number for this particular thing?Shawn Wang [00:14:31]: Yeah, it's retrieval. It's retrieval within machine learning. It's interesting because I think the more meta level I'm trying to operate at here is you have a benchmark. You're like, okay, I see the architectural thing I need to do in order to go fix that. But should you do it? Because sometimes that's an inductive bias, basically. It's what Jason Wei, who used to work at Google, would say. Exactly the kind of thing. Yeah, you're going to win. Short term. Longer term, I don't know if that's going to scale. You might have to undo that.Jeff Dean [00:15:01]: I mean, I like to sort of not focus on exactly what solution we're going to derive, but what capability would you want? And I think we're very convinced that, you know, long context is useful, but it's way too short today. Right? Like, I think what you would really want is, can I attend to the internet while I answer my question? Right? But that's not going to happen. I think that's going to be solved by purely scaling the existing solutions, which are quadratic. So a million tokens kind of pushes what you can do. You're not going to do that to a trillion tokens, let alone, you know, a billion tokens, let alone a trillion. But I think if you could give the illusion that you can attend to trillions of tokens, that would be amazing. You'd find all kinds of uses for that. You would have attend to the internet. You could attend to the pixels of YouTube and the sort of deeper representations that we can find. You could attend to the form for a single video, but across many videos, you know, on a personal Gemini level, you could attend to all of your personal state with your permission. So like your emails, your photos, your docs, your plane tickets you have. I think that would be really, really useful. And the question is, how do you get algorithmic improvements and system level improvements that get you to something where you actually can attend to trillions of tokens? Right. In a meaningful way. Yeah.Shawn Wang [00:16:26]: But by the way, I think I did some math and it's like, if you spoke all day, every day for eight hours a day, you only generate a maximum of like a hundred K tokens, which like very comfortably fits.Jeff Dean [00:16:38]: Right. But if you then say, okay, I want to be able to understand everything people are putting on videos.Shawn Wang [00:16:46]: Well, also, I think that the classic example is you start going beyond language into like proteins and whatever else is extremely information dense. Yeah. Yeah.Jeff Dean [00:16:55]: I mean, I think one of the things about Gemini's multimodal aspects is we've always wanted it to be multimodal from the start. And so, you know, that sometimes to people means text and images and video sort of human-like and audio, audio, human-like modalities. But I think it's also really useful to have Gemini know about non-human modalities. Yeah. Like LIDAR sensor data from. Yes. Say, Waymo vehicles or. Like robots or, you know, various kinds of health modalities, x-rays and MRIs and imaging and genomics information. And I think there's probably hundreds of modalities of data where you'd like the model to be able to at least be exposed to the fact that this is an interesting modality and has certain meaning in the world. Where even if you haven't trained on all the LIDAR data or MRI data, you could have, because maybe that's not, you know, it doesn't make sense in terms of trade-offs of. You know, what you include in your main pre-training data mix, at least including a little bit of it is actually quite useful. Yeah. Because it sort of tempts the model that this is a thing.Shawn Wang [00:18:04]: Yeah. Do you believe, I mean, since we're on this topic and something I just get to ask you all the questions I always wanted to ask, which is fantastic. Like, are there some king modalities, like modalities that supersede all the other modalities? So a simple example was Vision can, on a pixel level, encode text. And DeepSeq had this DeepSeq CR paper that did that. Vision. And Vision has also been shown to maybe incorporate audio because you can do audio spectrograms and that's, that's also like a Vision capable thing. Like, so, so maybe Vision is just the king modality and like. Yeah.Jeff Dean [00:18:36]: I mean, Vision and Motion are quite important things, right? Motion. Well, like video as opposed to static images, because I mean, there's a reason evolution has evolved eyes like 23 independent ways, because it's such a useful capability for sensing the world around you, which is really what we want these models to be. So I think the only thing that we can be able to do is interpret the things we're seeing or the things we're paying attention to and then help us in using that information to do things. Yeah.Shawn Wang [00:19:05]: I think motion, you know, I still want to shout out, I think Gemini, still the only native video understanding model that's out there. So I use it for YouTube all the time. Nice.Jeff Dean [00:19:15]: Yeah. Yeah. I mean, it's actually, I think people kind of are not necessarily aware of what the Gemini models can actually do. Yeah. Like I have an example I've used in one of my talks. It had like, it was like a YouTube highlight video of 18 memorable sports moments across the last 20 years or something. So it has like Michael Jordan hitting some jump shot at the end of the finals and, you know, some soccer goals and things like that. And you can literally just give it the video and say, can you please make me a table of what all these different events are? What when the date is when they happened? And a short description. And so you get like now an 18 row table of that information extracted from the video, which is, you know, not something most people think of as like a turn video into sequel like table.Alessio Fanelli [00:20:11]: Has there been any discussion inside of Google of like, you mentioned tending to the whole internet, right? Google, it's almost built because a human cannot tend to the whole internet and you need some sort of ranking to find what you need. Yep. That ranking is like much different for an LLM because you can expect a person to look at maybe the first five, six links in a Google search versus for an LLM. Should you expect to have 20 links that are highly relevant? Like how do you internally figure out, you know, how do we build the AI mode that is like maybe like much broader search and span versus like the more human one? Yeah.Jeff Dean [00:20:47]: I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. With a giant number of web pages in our index, many of them are not relevant. So you identify a subset of them that are relevant with very lightweight kinds of methods. You know, you're down to like 30,000 documents or something. And then you gradually refine that to apply more and more sophisticated algorithms and more and more sophisticated sort of signals of various kinds in order to get down to ultimately what you show, which is, you know, the final 10 results or, you know, 10 results plus. Other kinds of information. And I think an LLM based system is not going to be that dissimilar, right? You're going to attend to trillions of tokens, but you're going to want to identify, you know, what are the 30,000 ish documents that are with the, you know, maybe 30 million interesting tokens. And then how do you go from that into what are the 117 documents I really should be paying attention to in order to carry out the tasks that the user has asked? And I think, you know, you can imagine systems where you have, you know, a lot of highly parallel processing to identify those initial 30,000 candidates, maybe with very lightweight kinds of models. Then you have some system that sort of helps you narrow down from 30,000 to the 117 with maybe a little bit more sophisticated model or set of models. And then maybe the final model is the thing that looks. So the 117 things that might be your most capable model. So I think it has to, it's going to be some system like that, that is really enables you to give the illusion of attending to trillions of tokens. Sort of the way Google search gives you, you know, not the illusion, but you are searching the internet, but you're finding, you know, a very small subset of things that are, that are relevant.Shawn Wang [00:22:47]: Yeah. I often tell a lot of people that are not steeped in like Google search history that, well, you know, like Bert was. Like he was like basically immediately inside of Google search and that improves results a lot, right? Like I don't, I don't have any numbers off the top of my head, but like, I'm sure you guys, that's obviously the most important numbers to Google. Yeah.Jeff Dean [00:23:08]: I mean, I think going to an LLM based representation of text and words and so on enables you to get out of the explicit hard notion of, of particular words having to be on the page, but really getting at the notion of this topic of this page or this page. Paragraph is highly relevant to this query. Yeah.Shawn Wang [00:23:28]: I don't think people understand how much LLMs have taken over all these very high traffic system, very high traffic. Yeah. Like it's Google, it's YouTube. YouTube has this like semantics ID thing where it's just like every token or every item in the vocab is a YouTube video or something that predicts the video using a code book, which is absurd to me for YouTube size.Jeff Dean [00:23:50]: And then most recently GROK also for, for XAI, which is like, yeah. I mean, I'll call out even before LLMs were used extensively in search, we put a lot of emphasis on softening the notion of what the user actually entered into the query.Shawn Wang [00:24:06]: So do you have like a history of like, what's the progression? Oh yeah.Jeff Dean [00:24:09]: I mean, I actually gave a talk in, uh, I guess, uh, web search and data mining conference in 2009, uh, where we never actually published any papers about the origins of Google search, uh, sort of, but we went through sort of four or five or six. generations, four or five or six generations of, uh, redesigning of the search and retrieval system, uh, from about 1999 through 2004 or five. And that talk is really about that evolution. And one of the things that really happened in 2001 was we were sort of working to scale the system in multiple dimensions. So one is we wanted to make our index bigger, so we could retrieve from a larger index, which always helps your quality in general. Uh, because if you don't have the page in your index, you're going to not do well. Um, and then we also needed to scale our capacity because we were, our traffic was growing quite extensively. Um, and so we had, you know, a sharded system where you have more and more shards as the index grows, you have like 30 shards. And then if you want to double the index size, you make 60 shards so that you can bound the latency by which you respond for any particular user query. Um, and then as traffic grows, you add, you add more and more replicas of each of those. And so we eventually did the math that realized that in a data center where we had say 60 shards and, um, you know, 20 copies of each shard, we now had 1200 machines, uh, with disks. And we did the math and we're like, Hey, one copy of that index would actually fit in memory across 1200 machines. So in 2001, we introduced, uh, we put our entire index in memory and what that enabled from a quality perspective was amazing. Um, and so we had more and more replicas of each of those. Before you had to be really careful about, you know, how many different terms you looked at for a query, because every one of them would involve a disk seek on every one of the 60 shards. And so you, as you make your index bigger, that becomes even more inefficient. But once you have the whole index in memory, it's totally fine to have 50 terms you throw into the query from the user's original three or four word query, because now you can add synonyms like restaurant and restaurants and cafe and, uh, you know, things like that. Uh, bistro and all these things. And you can suddenly start, uh, sort of really, uh, getting at the meaning of the word as opposed to the exact semantic form the user typed in. And that was, you know, 2001, very much pre LLM, but really it was about softening the, the strict definition of what the user typed in order to get at the meaning.Alessio Fanelli [00:26:47]: What are like principles that you use to like design the systems, especially when you have, I mean, in 2001, the internet is like. Doubling, tripling every year in size is not like, uh, you know, and I think today you kind of see that with LLMs too, where like every year the jumps in size and like capabilities are just so big. Are there just any, you know, principles that you use to like, think about this? Yeah.Jeff Dean [00:27:08]: I mean, I think, uh, you know, first, whenever you're designing a system, you want to understand what are the sort of design parameters that are going to be most important in designing that, you know? So, you know, how many queries per second do you need to handle? How big is the internet? How big is the index you need to handle? How much data do you need to keep for every document in the index? How are you going to look at it when you retrieve things? Um, what happens if traffic were to double or triple, you know, will that system work well? And I think a good design principle is you're going to want to design a system so that the most important characteristics could scale by like factors of five or 10, but probably not beyond that because often what happens is if you design a system for X. And something suddenly becomes a hundred X, that would enable a very different point in the design space that would not make sense at X. But all of a sudden at a hundred X makes total sense. So like going from a disk space index to a in memory index makes a lot of sense once you have enough traffic, because now you have enough replicas of the sort of state on disk that those machines now actually can hold, uh, you know, a full copy of the, uh, index and memory. Yeah. And that all of a sudden enabled. A completely different design that wouldn't have been practical before. Yeah. Um, so I'm, I'm a big fan of thinking through designs in your head, just kind of playing with the design space a little before you actually do a lot of writing of code. But, you know, as you said, in the early days of Google, we were growing the index, uh, quite extensively. We were growing the update rate of the index. So the update rate actually is the parameter that changed the most. Surprising. So it used to be once a month.Shawn Wang [00:28:55]: Yeah.Jeff Dean [00:28:56]: And then we went to a system that could update any particular page in like sub one minute. Okay.Shawn Wang [00:29:02]: Yeah. Because this is a competitive advantage, right?Jeff Dean [00:29:04]: Because all of a sudden news related queries, you know, if you're, if you've got last month's news index, it's not actually that useful for.Shawn Wang [00:29:11]: News is a special beast. Was there any, like you could have split it onto a separate system.Jeff Dean [00:29:15]: Well, we did. We launched a Google news product, but you also want news related queries that people type into the main index to also be sort of updated.Shawn Wang [00:29:23]: So, yeah, it's interesting. And then you have to like classify whether the page is, you have to decide which pages should be updated and what frequency. Oh yeah.Jeff Dean [00:29:30]: There's a whole like, uh, system behind the scenes that's trying to decide update rates and importance of the pages. So even if the update rate seems low, you might still want to recrawl important pages quite often because, uh, the likelihood they change might be low, but the value of having updated is high.Shawn Wang [00:29:50]: Yeah, yeah, yeah, yeah. Uh, well, you know, yeah. This, uh, you know, mention of latency and, and saving things to this reminds me of one of your classics, which I have to bring up, which is latency numbers. Every programmer should know, uh, was there a, was it just a, just a general story behind that? Did you like just write it down?Jeff Dean [00:30:06]: I mean, this has like sort of eight or 10 different kinds of metrics that are like, how long does a cache mistake? How long does branch mispredict take? How long does a reference domain memory take? How long does it take to send, you know, a packet from the U S to the Netherlands or something? Um,Shawn Wang [00:30:21]: why Netherlands, by the way, or is it, is that because of Chrome?Jeff Dean [00:30:25]: Uh, we had a data center in the Netherlands, um, so, I mean, I think this gets to the point of being able to do the back of the envelope calculations. So these are sort of the raw ingredients of those, and you can use them to say, okay, well, if I need to design a system to do image search and thumb nailing or something of the result page, you know, how, what I do that I could pre-compute the image thumbnails. I could like. Try to thumbnail them on the fly from the larger images. What would that do? How much dis bandwidth than I need? How many des seeks would I do? Um, and you can sort of actually do thought experiments in, you know, 30 seconds or a minute with the sort of, uh, basic, uh, basic numbers at your fingertips. Uh, and then as you sort of build software using higher level libraries, you kind of want to develop the same intuitions for how long does it take to, you know, look up something in this particular kind of.Shawn Wang [00:31:21]: I'll see you next time.Shawn Wang [00:31:51]: Which is a simple byte conversion. That's nothing interesting. I wonder if you have any, if you were to update your...Jeff Dean [00:31:58]: I mean, I think it's really good to think about calculations you're doing in a model, either for training or inference.Jeff Dean [00:32:09]: Often a good way to view that is how much state will you need to bring in from memory, either like on-chip SRAM or HBM from the accelerator. Attached memory or DRAM or over the network. And then how expensive is that data motion relative to the cost of, say, an actual multiply in the matrix multiply unit? And that cost is actually really, really low, right? Because it's order, depending on your precision, I think it's like sub one picodule.Shawn Wang [00:32:50]: Oh, okay. You measure it by energy. Yeah. Yeah.Jeff Dean [00:32:52]: Yeah. I mean, it's all going to be about energy and how do you make the most energy efficient system. And then moving data from the SRAM on the other side of the chip, not even off the off chip, but on the other side of the same chip can be, you know, a thousand picodules. Oh, yeah. And so all of a sudden, this is why your accelerators require batching. Because if you move, like, say, the parameter of a model from SRAM on the, on the chip into the multiplier unit, that's going to cost you a thousand picodules. So you better make use of that, that thing that you moved many, many times with. So that's where the batch dimension comes in. Because all of a sudden, you know, if you have a batch of 256 or something, that's not so bad. But if you have a batch of one, that's really not good.Shawn Wang [00:33:40]: Yeah. Yeah. Right.Jeff Dean [00:33:41]: Because then you paid a thousand picodules in order to do your one picodule multiply.Shawn Wang [00:33:46]: I have never heard an energy-based analysis of batching.Jeff Dean [00:33:50]: Yeah. I mean, that's why people batch. Yeah. Ideally, you'd like to use batch size one because the latency would be great.Shawn Wang [00:33:56]: The best latency.Jeff Dean [00:33:56]: But the energy cost and the compute cost inefficiency that you get is quite large. So, yeah.Shawn Wang [00:34:04]: Is there a similar trick like, like, like you did with, you know, putting everything in memory? Like, you know, I think obviously NVIDIA has caused a lot of waves with betting very hard on SRAM with Grok. I wonder if, like, that's something that you already saw with, with the TPUs, right? Like that, that you had to. Uh, to serve at your scale, uh, you probably sort of saw that coming. Like what, what, what hardware, uh, innovations or insights were formed because of what you're seeing there?Jeff Dean [00:34:33]: Yeah. I mean, I think, you know, TPUs have this nice, uh, sort of regular structure of 2D or 3D meshes with a bunch of chips connected. Yeah. And each one of those has HBM attached. Um, I think for serving some kinds of models, uh, you know, you, you pay a lot higher cost. Uh, and time latency, um, bringing things in from HBM than you do bringing them in from, uh, SRAM on the chip. So if you have a small enough model, you can actually do model parallelism, spread it out over lots of chips and you actually get quite good throughput improvements and latency improvements from doing that. And so you're now sort of striping your smallish scale model over say 16 or 64 chips. Uh, but as if you do that and it all fits in. In SRAM, uh, that can be a big win. So yeah, that's not a surprise, but it is a good technique.Alessio Fanelli [00:35:27]: Yeah. What about the TPU design? Like how much do you decide where the improvements have to go? So like, this is like a good example of like, is there a way to bring the thousand picojoules down to 50? Like, is it worth designing a new chip to do that? The extreme is like when people say, oh, you should burn the model on the ASIC and that's kind of like the most extreme thing. How much of it? Is it worth doing an hardware when things change so quickly? Like what was the internal discussion? Yeah.Jeff Dean [00:35:57]: I mean, we, we have a lot of interaction between say the TPU chip design architecture team and the sort of higher level modeling, uh, experts, because you really want to take advantage of being able to co-design what should future TPUs look like based on where we think the sort of ML research puck is going, uh, in some sense, because, uh, you know, as a hardware designer for ML and in particular, you're trying to design a chip starting today and that design might take two years before it even lands in a data center. And then it has to sort of be a reasonable lifetime of the chip to take you three, four or five years. So you're trying to predict two to six years out where, what ML computations will people want to run two to six years out in a very fast changing field. And so having people with interest. Interesting ML research ideas of things we think will start to work in that timeframe or will be more important in that timeframe, uh, really enables us to then get, you know, interesting hardware features put into, you know, TPU N plus two, where TPU N is what we have today.Shawn Wang [00:37:10]: Oh, the cycle time is plus two.Jeff Dean [00:37:12]: Roughly. Wow. Because, uh, I mean, sometimes you can squeeze some changes into N plus one, but, you know, bigger changes are going to require the chip. Yeah. Design be earlier in its lifetime design process. Um, so whenever we can do that, it's generally good. And sometimes you can put in speculative features that maybe won't cost you much chip area, but if it works out, it would make something, you know, 10 times as fast. And if it doesn't work out, well, you burned a little bit of tiny amount of your chip area on that thing, but it's not that big a deal. Uh, sometimes it's a very big change and we want to be pretty sure this is going to work out. So we'll do like lots of carefulness. Uh, ML experimentation to show us, uh, this is actually the, the way we want to go. Yeah.Alessio Fanelli [00:37:58]: Is there a reverse of like, we already committed to this chip design so we can not take the model architecture that way because it doesn't quite fit?Jeff Dean [00:38:06]: Yeah. I mean, you, you definitely have things where you're going to adapt what the model architecture looks like so that they're efficient on the chips that you're going to have for both training and inference of that, of that, uh, generation of model. So I think it kind of goes both ways. Um, you know, sometimes you can take advantage of, you know, lower precision things that are coming in a future generation. So you can, might train it at that lower precision, even if the current generation doesn't quite do that. Mm.Shawn Wang [00:38:40]: Yeah. How low can we go in precision?Jeff Dean [00:38:43]: Because people are saying like ternary is like, uh, yeah, I mean, I'm a big fan of very low precision because I think that gets, that saves you a tremendous amount of time. Right. Because it's picojoules per bit that you're transferring and reducing the number of bits is a really good way to, to reduce that. Um, you know, I think people have gotten a lot of luck, uh, mileage out of having very low bit precision things, but then having scaling factors that apply to a whole bunch of, uh, those, those weights. Scaling. How does it, how does it, okay.Shawn Wang [00:39:15]: Interesting. You, so low, low precision, but scaled up weights. Yeah. Huh. Yeah. Never considered that. Yeah. Interesting. Uh, w w while we're on this topic, you know, I think there's a lot of, um, uh, this, the concept of precision at all is weird when we're sampling, you know, uh, we just, at the end of this, we're going to have all these like chips that I'll do like very good math. And then we're just going to throw a random number generator at the start. So, I mean, there's a movement towards, uh, energy based, uh, models and processors. I'm just curious if you've, obviously you've thought about it, but like, what's your commentary?Jeff Dean [00:39:50]: Yeah. I mean, I think. There's a bunch of interesting trends though. Energy based models is one, you know, diffusion based models, which don't sort of sequentially decode tokens is another, um, you know, speculative decoding is a way that you can get sort of an equivalent, very small.Shawn Wang [00:40:06]: Draft.Jeff Dean [00:40:07]: Batch factor, uh, for like you predict eight tokens out and that enables you to sort of increase the effective batch size of what you're doing by a factor of eight, even, and then you maybe accept five or six of those tokens. So you get. A five, a five X improvement in the amortization of moving weights, uh, into the multipliers to do the prediction for the, the tokens. So these are all really good techniques and I think it's really good to look at them from the lens of, uh, energy, real energy, not energy based models, um, and, and also latency and throughput, right? If you look at things from that lens, that sort of guides you to. Two solutions that are gonna be, uh, you know, better from, uh, you know, being able to serve larger models or, you know, equivalent size models more cheaply and with lower latency.Shawn Wang [00:41:03]: Yeah. Well, I think, I think I, um, it's appealing intellectually, uh, haven't seen it like really hit the mainstream, but, um, I do think that, uh, there's some poetry in the sense that, uh, you know, we don't have to do, uh, a lot of shenanigans if like we fundamentally. Design it into the hardware. Yeah, yeah.Jeff Dean [00:41:23]: I mean, I think there's still a, there's also sort of the more exotic things like analog based, uh, uh, computing substrates as opposed to digital ones. Uh, I'm, you know, I think those are super interesting cause they can be potentially low power. Uh, but I think you often end up wanting to interface that with digital systems and you end up losing a lot of the power advantages in the digital to analog and analog to digital conversions. You end up doing, uh, at the sort of boundaries. And periphery of that system. Um, I still think there's a tremendous distance we can go from where we are today in terms of energy efficiency with sort of, uh, much better and specialized hardware for the models we care about.Shawn Wang [00:42:05]: Yeah.Alessio Fanelli [00:42:06]: Um, any other interesting research ideas that you've seen, or like maybe things that you cannot pursue a Google that you would be interested in seeing researchers take a step at, I guess you have a lot of researchers. Yeah, I guess you have enough, but our, our research.Jeff Dean [00:42:21]: Our research portfolio is pretty broad. I would say, um, I mean, I think, uh, in terms of research directions, there's a whole bunch of, uh, you know, open problems and how do you make these models reliable and able to do much longer, kind of, uh, more complex tasks that have lots of subtasks. How do you orchestrate, you know, maybe one model that's using other models as tools in order to sort of build, uh, things that can accomplish, uh, you know, much more. Yeah. Significant pieces of work, uh, collectively, then you would ask a single model to do. Um, so that's super interesting. How do you get more verifiable, uh, you know, how do you get RL to work for non-verifiable domains? I think it's a pretty interesting open problem because I think that would broaden out the capabilities of the models, the improvements that you're seeing in both math and coding. Uh, if we could apply those to other less verifiable domains, because we've come up with RL techniques that actually enable us to do that. Uh, effectively, that would, that would really make the models improve quite a lot. I think.Alessio Fanelli [00:43:26]: I'm curious, like when we had Noam Brown on the podcast, he said, um, they already proved you can do it with deep research. Um, you kind of have it with AI mode in a way it's not verifiable. I'm curious if there's any thread that you think is interesting there. Like what is it? Both are like information retrieval of JSON. So I wonder if it's like the retrieval is like the verifiable part. That you can score or what are like, yeah, yeah. How, how would you model that, that problem?Jeff Dean [00:43:55]: Yeah. I mean, I think there are ways of having other models that can evaluate the results of what a first model did, maybe even retrieving. Can you have another model that says, is this things, are these things you retrieved relevant? Or can you rate these 2000 things you retrieved to assess which ones are the 50 most relevant or something? Um, I think those kinds of techniques are actually quite effective. Sometimes I can even be the same model, just prompted differently to be a, you know, a critic as opposed to a, uh, actual retrieval system. Yeah.Shawn Wang [00:44:28]: Um, I do think like there, there is that, that weird cliff where like, it feels like we've done the easy stuff and then now it's, but it always feels like that every year. It's like, oh, like we know, we know, and the next part is super hard and nobody's figured it out. And, uh, exactly with this RLVR thing where like everyone's talking about, well, okay, how do we. the next stage of the non-verifiable stuff. And everyone's like, I don't know, you know, Ellen judge.Jeff Dean [00:44:56]: I mean, I feel like the nice thing about this field is there's lots and lots of smart people thinking about creative solutions to some of the problems that we all see. Uh, because I think everyone sort of sees that the models, you know, are great at some things and they fall down around the edges of those things and, and are not as capable as we'd like in those areas. And then coming up with good techniques and trying those. And seeing which ones actually make a difference is sort of what the whole research aspect of this field is, is pushing forward. And I think that's why it's super interesting. You know, if you think about two years ago, we were struggling with GSM, eight K problems, right? Like, you know, Fred has two rabbits. He gets three more rabbits. How many rabbits does he have? That's a pretty far cry from the kinds of mathematics that the models can, and now you're doing IMO and Erdos problems in pure language. Yeah. Yeah. Pure language. So that is a really, really amazing jump in capabilities in, you know, in a year and a half or something. And I think, um, for other areas, it'd be great if we could make that kind of leap. Uh, and you know, we don't exactly see how to do it for some, some areas, but we do see it for some other areas and we're going to work hard on making that better. Yeah.Shawn Wang [00:46:13]: Yeah.Alessio Fanelli [00:46:14]: Like YouTube thumbnail generation. That would be very helpful. We need that. That would be AGI. We need that.Shawn Wang [00:46:20]: That would be. As far as content creators go.Jeff Dean [00:46:22]: I guess I'm not a YouTube creator, so I don't care that much about that problem, but I guess, uh, many people do.Shawn Wang [00:46:27]: It does. Yeah. It doesn't, it doesn't matter. People do judge books by their covers as it turns out. Um, uh, just to draw a bit on the IMO goal. Um, I'm still not over the fact that a year ago we had alpha proof and alpha geometry and all those things. And then this year we were like, screw that we'll just chuck it into Gemini. Yeah. What's your reflection? Like, I think this, this question about. Like the merger of like symbolic systems and like, and, and LMS, uh, was a very much core belief. And then somewhere along the line, people would just said, Nope, we'll just all do it in the LLM.Jeff Dean [00:47:02]: Yeah. I mean, I think it makes a lot of sense to me because, you know, humans manipulate symbols, but we probably don't have like a symbolic representation in our heads. Right. We have some distributed representation that is neural net, like in some way of lots of different neurons. And activation patterns firing when we see certain things and that enables us to reason and plan and, you know, do chains of thought and, you know, roll them back now that, that approach for solving the problem doesn't seem like it's going to work. I'm going to try this one. And, you know, in a lot of ways we're emulating what we intuitively think, uh, is happening inside real brains in neural net based models. So it never made sense to me to have like completely separate. Uh, discrete, uh, symbolic things, and then a completely different way of, of, uh, you know, thinking about those things.Shawn Wang [00:47:59]: Interesting. Yeah. Uh, I mean, it's maybe seems obvious to you, but it wasn't obvious to me a year ago. Yeah.Jeff Dean [00:48:06]: I mean, I do think like that IMO with, you know, translating to lean and using lean and then the next year and also a specialized geometry model. And then this year switching to a single unified model. That is roughly the production model with a little bit more inference budget, uh, is actually, you know, quite good because it shows you that the capabilities of that general model have improved dramatically and, and now you don't need the specialized model. This is actually sort of very similar to the 2013 to 16 era of machine learning, right? Like it used to be, people would train separate models for lots of different, each different problem, right? I have, I want to recognize street signs and something. So I train a street sign. Recognition recognition model, or I want to, you know, decode speech recognition. I have a speech model, right? I think now the era of unified models that do everything is really upon us. And the question is how well do those models generalize to new things they've never been asked to do and they're getting better and better.Shawn Wang [00:49:10]: And you don't need domain experts. Like one of my, uh, so I interviewed ETA who was on, who was on that team. Uh, and he was like, yeah, I, I don't know how they work. I don't know where the IMO competition was held. I don't know the rules of it. I just trained the models, the training models. Yeah. Yeah. And it's kind of interesting that like people with these, this like universal skill set of just like machine learning, you just give them data and give them enough compute and they can kind of tackle any task, which is the bitter lesson, I guess. I don't know. Yeah.Jeff Dean [00:49:39]: I mean, I think, uh, general models, uh, will win out over specialized ones in most cases.Shawn Wang [00:49:45]: Uh, so I want to push there a bit. I think there's one hole here, which is like, uh. There's this concept of like, uh, maybe capacity of a model, like abstractly a model can only contain the number of bits that it has. And, uh, and so it, you know, God knows like Gemini pro is like one to 10 trillion parameters. We don't know, but, uh, the Gemma models, for example, right? Like a lot of people want like the open source local models that are like that, that, that, and, and, uh, they have some knowledge, which is not necessary, right? Like they can't know everything like, like you have the. The luxury of you have the big model and big model should be able to capable of everything. But like when, when you're distilling and you're going down to the small models, you know, you're actually memorizing things that are not useful. Yeah. And so like, how do we, I guess, do we want to extract that? Can we, can we divorce knowledge from reasoning, you know?Jeff Dean [00:50:38]: Yeah. I mean, I think you do want the model to be most effective at reasoning if it can retrieve things, right? Because having the model devote precious parameter space. To remembering obscure facts that could be looked up is actually not the best use of that parameter space, right? Like you might prefer something that is more generally useful in more settings than this obscure fact that it has. Um, so I think that's always attention at the same time. You also don't want your model to be kind of completely detached from, you know, knowing stuff about the world, right? Like it's probably useful to know how long the golden gate be. Bridges just as a general sense of like how long are bridges, right? And, uh, it should have that kind of knowledge. It maybe doesn't need to know how long some teeny little bridge in some other more obscure part of the world is, but, uh, it does help it to have a fair bit of world knowledge and the bigger your model is, the more you can have. Uh, but I do think combining retrieval with sort of reasoning and making the model really good at doing multiple stages of retrieval. Yeah.Shawn Wang [00:51:49]: And reasoning through the intermediate retrieval results is going to be a, a pretty effective way of making the model seem much more capable, because if you think about, say, a personal Gemini, yeah, right?Jeff Dean [00:52:01]: Like we're not going to train Gemini on my email. Probably we'd rather have a single model that, uh, we can then use and use being able to retrieve from my email as a tool and have the model reason about it and retrieve from my photos or whatever, uh, and then make use of that and have multiple. Um, you know, uh, stages of interaction. that makes sense.Alessio Fanelli [00:52:24]: Do you think the vertical models are like, uh, interesting pursuit? Like when people are like, oh, we're building the best healthcare LLM, we're building the best law LLM, are those kind of like short-term stopgaps or?Jeff Dean [00:52:37]: No, I mean, I think, I think vertical models are interesting. Like you want them to start from a pretty good base model, but then you can sort of, uh, sort of viewing them, view them as enriching the data. Data distribution for that particular vertical domain for healthcare, say, um, we're probably not going to train or for say robotics. We're probably not going to train Gemini on all possible robotics data. We, you could train it on because we want it to have a balanced set of capabilities. Um, so we'll expose it to some robotics data, but if you're trying to build a really, really good robotics model, you're going to want to start with that and then train it on more robotics data. And then maybe that would. It's multilingual translation capability, but improve its robotics capabilities. And we're always making these kind of, uh, you know, trade-offs in the data mix that we train the base Gemini models on. You know, we'd love to include data from 200 more languages and as much data as we have for those languages, but that's going to displace some other capabilities of the model. It won't be as good at, um, you know, Pearl programming, you know, it'll still be good at Python programming. Cause we'll include it. Enough. Of that, but there's other long tail computer languages or coding capabilities that it may suffer on or multi, uh, multimodal reasoning capabilities may suffer. Cause we didn't get to expose it to as much data there, but it's really good at multilingual things. So I, I think some combination of specialized models, maybe more modular models. So it'd be nice to have the capability to have those 200 languages, plus this awesome robotics model, plus this awesome healthcare, uh, module that all can be knitted together to work in concert and called upon in different circumstances. Right? Like if I have a health related thing, then it should enable using this health module in conjunction with the main base model to be even better at those kinds of things. Yeah.Shawn Wang [00:54:36]: Installable knowledge. Yeah.Jeff Dean [00:54:37]: Right.Shawn Wang [00:54:38]: Just download as a, as a package.Jeff Dean [00:54:39]: And some of that installable stuff can come from retrieval, but some of it probably should come from preloaded training on, you know, uh, a hundred billion tokens or a trillion tokens of health data. Yeah.Shawn Wang [00:54:51]: And for listeners, I think, uh, I will highlight the Gemma three end paper where they, there was a little bit of that, I think. Yeah.Alessio Fanelli [00:54:56]: Yeah. I guess the question is like, how many billions of tokens do you need to outpace the frontier model improvements? You know, it's like, if I have to make this model better healthcare and the main. Gemini model is still improving. Do I need 50 billion tokens? Can I do it with a hundred, if I need a trillion healthcare tokens, it's like, they're probably not out there that you don't have, you know, I think that's really like the.Jeff Dean [00:55:21]: Well, I mean, I think healthcare is a particularly challenging domain, so there's a lot of healthcare data that, you know, we don't have access to appropriately, but there's a lot of, you know, uh, healthcare organizations that want to train models on their own data. That is not public healthcare data, uh, not public health. But public healthcare data. Um, so I think there are opportunities there to say, partner with a large healthcare organization and train models for their use that are going to be, you know, more bespoke, but probably, uh, might be better than a general model trained on say, public data. Yeah.Shawn Wang [00:55:58]: Yeah. I, I believe, uh, by the way, also this is like somewhat related to the language conversation. Uh, I think one of your, your favorite examples was you can put a low resource language in the context and it just learns. Yeah.Jeff Dean [00:56:09]: Oh, yeah, I think the example we used was Calamon, which is truly low resource because it's only spoken by, I think 120 people in the world and there's no written text.Shawn Wang [00:56:20]: So, yeah. So you can just do it that way. Just put it in the context. Yeah. Yeah. But I think your whole data set in the context, right.Jeff Dean [00:56:27]: If you, if you take a language like, uh, you know, Somali or something, there is a fair bit of Somali text in the world that, uh, or Ethiopian Amharic or something, um, you know, we probably. Yeah. Are not putting all the data from those languages into the Gemini based training. We put some of it, but if you put more of it, you'll improve the capabilities of those models.Shawn Wang [00:56:49]: Yeah.Jeff Dean [00:56:49]:

Rutinní nácvik startu kosmické lodi na mysu Canaveral skončil v roce 1967 neuvěřitelnou tragédií. V pátek 27. ledna 1967 shořela při přípravě prvního pilotovaného letu projektu Apollo na startovací rampě kabina.

Think you need Hollywood connections for voiceover success? Think again! Local VO work is thriving.In this video, we bust the 5 biggest myths about local voice over jobs, showing you how to land high-paying gigs, build client relationships, and turn your VO passion into a full-time career without leaving your hometown. With actionable strategies and mindset shifts, you'll learn why everything you've been told about making it in VO is probably wrong and how to fix it!.Get your FREE Local Gig Kickstart Toolkit: https://welcome.vopro.pro/local-gig-kickstart-toolkitGVAA Rate Guide: https://www.globalvoiceacademy.com/gvaa-rate-guide/#VoiceOver #VOPro #VoiceActing #VoiceoverMyths #LocalBusiness #WorkFromAnywhere #CreativeCareer #FreelanceLife #FullTimeVOLinks: (When possible, I use affiliate links and may earn a commission. See disclosure below.)▶️ Subscribe: https://vopro.pro/youtube

Rutinní nácvik startu kosmické lodi na mysu Canaveral skončil v roce 1967 neuvěřitelnou tragédií. V pátek 27. ledna 1967 shořela při přípravě prvního pilotovaného letu projektu Apollo na startovací rampě kabina. Všechny díly podcastu Portréty můžete pohodlně poslouchat v mobilní aplikaci mujRozhlas pro Android a iOS nebo na webu mujRozhlas.cz.

Los titulares de la industria del deporte, con Patricia López, de 2Playbook.Apollo refuerza su apuesta por el deporte con un nuevo fondo sectorial de 6.000 millones de dólares. Tras entrar en el capital del Atlético de Madrid, la gestora calcula que la industria necesitará hasta 50.000 millones en financiación en los próximos años. En Madrid, Rwanda y el Atlético impulsan una alianza estratégica que convierte el Riyadh Air Metropolitano en escaparate de inversión, innovación y fútbol de primer nivel. En el ámbito empresarial, Ferrari mantiene su crecimiento: su división comercial facturó un 22% más en 2025, hasta 820 millones, gracias al patrocinio, el prize money de la F1 y sus parques temáticos. En Estados Unidos, la MLS alcanza un valor medio de 650 millones por franquicia, con Inter Miami como líder con 1.450 millones. Finalmente, la Real Federación Española de Atletismo firma con Allianz como patrocinadora principal hasta 2032 para impulsar el talento y la preparación de atletas.

Johnny Pemberton (Mermaid in theaters in March! Fallout season 2 out now on Prime!) returns to make it weird! Get a free 8-count Sample Pack of LMNT’s most popular drink mix flavors with any purchase at drinkLMNT.com/weird Check out Little Saints and discover your Magic Hour. Visit littlesaints.com and use code WEIRD to get 15% Off Your First Get a free can of OLIPOP! Buy any 2 cans of Olipop in store, and they’ll pay you back for one. Works on any flavor, any retailer. Go to drinkolipop.com/WEIRD If you want to support your stress and sleep—not just track it—Apollo is worth trying. For a limited time, get $99 off the Apollo Wearable + SmartVibes bundle at apolloneuro.com/weird with code WEIR See omnystudio.com/listener for privacy information.

Where did Earth’s water come from? In this episode of Planetary Radio, we explore how scientists are answering that question by studying a remarkably well-preserved record of the early Solar System: lunar samples brought back by the Apollo missions. Host Sarah Al-Ahmed is joined by Tony Gargano, postdoctoral fellow at the Lunar and Planetary Institute with the University Space Research Association and a research affiliate at NASA’s Johnson Space Center. Gargano studies lunar rocks and regolith to understand how planets form, evolve, and acquire key ingredients like water over time. By analyzing subtle chemical fingerprints preserved in Apollo-era lunar regolith, his work helps constrain how much water meteorites could have brought to Earth and what that means for our planet’s path to habitability. The episode also features a short bonus segment with actor George Takei, recorded at the Academy Museum of Motion Pictures during a screening of “Star Trek IV: The Voyage Home.” Takei reflects on the enduring legacy of “Star Trek,” its influence on generations of scientists and explorers, and why he is excited about humanity’s return to the Moon in the Artemis era. He connects science fiction’s hopeful vision of the future with the real science helping us understand our origins today. Discover more at: https://www.planetary.org/planetary-radio/2026-earth-water-apollo-moon-dustSee omnystudio.com/listener for privacy information.

❤️ SUPPORT US and get Exclusive Perks! https://www.patreon.com/tabletoptime

Bridget, Caitlin, and Hilda are back with part 1 of "Between Two Kings" book 2 in Lindsay Straube's Split or Swallow series. And if you read book 1 (or listened to the episode) then you know how outrageous this series is, and this book does NOT disappoint. Like we're sure there's a plot we're supposed to care about, but the basilisks have entered mating season which means Tem is indulging her basilisk nature. Anyways, listen now for thoughts on part 1.  Join our Patreon for exclusive behind-the-scenes content and let's be friends!Instagram > @Booktokmademe_podTikTok > @BooktokMadeMe

ชมวิดีโอ EP นี้ใน YouTube เพื่อประสบการณ์การรับชมที่ดีที่สุด https://youtu.be/9A9mzuOwoog . ฝึกฟังเรื่องสั้นภาษาอังกฤษกับ ‘อพอลโล' เทพ Toxic กับแฟนหนุ่มคนแรกและคนเดียว! . คำนี้ดี Story รวบรวมเรื่องสั้นน่าสนใจมาเล่าเป็นภาษาอังกฤษเวอร์ชั่นเข้าใจง่าย พร้อมคำแปลภาษาไทย และไฮไลต์คำศัพท์ที่น่าสนใจ เพื่อการฝึกพัฒนาทักษะการฟัง (listening skill) . เอพิโสดนี้คุมธีมวันแห่งความรักกับเรื่องราวความรักวายๆ ในเทพนิยายกรีก ที่หลายคนอาจไม่รู้ว่า ในเทือกเขาโอลิมปัส มีความรัก BL อยู่ด้วย . ฝึกภาษาแบบสนุกๆ และได้เกร็ดความรู้อีกมากมายได้ที่ YouTube | Spotify | Apple Podcasts

On this week's episode, Emma talks about her showbizzy weekend in the Apollo and the 3Arena and Deirdre recounts yet another airport mishapThis Podcast is part of the Acast Network.Recorded at D2 Podcast StudioArtwork: Alan Bourke-TuffyThank you for listening! Follow Keep It Tight on Instagram!Hosted on Acast. See acast.com/privacy for more information.Thank you for listening! Follow Keep It Tight on Instagram! Artwork: Alan Bourke-Tuffy Hosted on Acast. See acast.com/privacy for more information.

ชมวิดีโอ EP นี้ใน YouTube เพื่อประสบการณ์การรับชมที่ดีที่สุด https://youtu.be/9A9mzuOwoog . ฝึกฟังเรื่องสั้นภาษาอังกฤษกับ ‘อพอลโล' เทพ Toxic กับแฟนหนุ่มคนแรกและคนเดียว! . คำนี้ดี Story รวบรวมเรื่องสั้นน่าสนใจมาเล่าเป็นภาษาอังกฤษเวอร์ชั่นเข้าใจง่าย พร้อมคำแปลภาษาไทย และไฮไลต์คำศัพท์ที่น่าสนใจ เพื่อการฝึกพัฒนาทักษะการฟัง (listening skill) . เอพิโสดนี้คุมธีมวันแห่งความรักกับเรื่องราวความรักวายๆ ในเทพนิยายกรีก ที่หลายคนอาจไม่รู้ว่า ในเทือกเขาโอลิมปัส มีความรัก BL อยู่ด้วย . ฝึกภาษาแบบสนุกๆ และได้เกร็ดความรู้อีกมากมายได้ที่ YouTube | Spotify | Apple Podcasts

This is Part 2! For Part 1, check the feed!This week we're looking at some of the greatest comebacks history has to offer. Where else to begin than the epic comeback that is Skoda cars! And what about the comeback of the West German economy after world war 2?! And finally, a literal comeback, Apollo 13's incredible return from the moon after disaster struck.Elsewhere, how good is the tip as a day out? Has history anything better to offer in terms of pure enjoyment? If you know, let us know: hello@ohwhatatime.com And if you want more Oh What A Time, you should sign up for our Patreon! On there you'll now find:•The full archive of bonus episodes•Brand new bonus episodes each month•OWAT subscriber group chats•Loads of extra perks for supporters of the show•PLUS ad-free episodes earlier than everyone elseJoin us at

On episode 332 of The AwardsWatch Podcast, Executive Editor Ryan McQuade is joined by AwardsWatch Editor-In-Chief Erik Anderson and AwardsWatch contributors Dan Bayer, Jay Ledbetter, and Josh Parham to go back 25 years and take a look at the 74th Academy Awards, covering the films of 2001. On this retrospective, the AW team starts the year off with a look back at a solid year of film in 2001, that brought together some of the most memorable films of the last 25 years. But the winner for Best Picture is not one held in high regard, as A Beautiful Mind took home the top prize, a make-up win for director Ron Howard after losing for Apollo 13. While the film hasn't aged well as a winner, and even as a film, the year has with spectacular films that are mentioned throughout the show like In the Mood for Love, Mulholland Drive, Gosford Park, Memento, Ghost World, The Royal Tenenbaums, In the Bedroom, Hedwig and the Angry Itch, The Lord of the Rings: The Fellowship of the Ring, and more.  In their in-depth discussion, the AW team talked about the film year of 2001, briefly discuss talk about A Beautiful Mind as a Best Picture winner, and how that speaks to the legacy of their nominates and or wins, do an extensive conversation over the below the line categories and nominees for the year, and then the new version of the AW Shoulda Woulda Coulda game, where instead of individual replacements, they must decide as a group who the nominees and winners should be in the top eight categories. The rules of the game state they can only replace two of the nominees that year from each category, except in Best Picture, where the group could replace up to three films to make up the final set of five nominated films. Like past retrospective episodes, it was a fascinating, fun conversation including spirited debates, alliances, vote swinging, celebrating various movies, performances that aren't normally talked about and more that we all hope you enjoy. You can listen to The AwardsWatch Podcast wherever you stream podcasts, from iTunes, iHeartRadio, Soundcloud, Stitcher, Spotify, Audible, Amazon Music and more. You can also listen to it on our AwardsWatch YouTube page. This podcast runs 2h06m. We will be back in next week for a review of the latest film from director Emerald Fennell, Wuthering Heights. Till then, let's get into it. Music: "Modern Fashion" from AShamaleuvmusic (intro), "B-3" from BoxCat Games Nameless: The Hackers RPG Soundtrack (outro).

Editor and singer Sally Dunkley joins Robert and Eamonn to understand why composers enjoy 'false relations', the major and minor chord at the same time. Music by Tallis, Shepherd, Byrd, Monteverdi with The Sixteen, The Cardinall's Musick, Gesualdo Six, I Fagiolini, Apollo 5 and more.Support this show http://supporter.acast.com/choral-chihuahua. Hosted on Acast. See acast.com/privacy for more information.

 HEADLINE: Tragedy and Rebirth: The Apollo 1 Fire. GUEST AUTHOR: Bob Zimmerman. SUMMARY: A tragic launchpad fire kills three astronauts, forcing NASA to admit carelessness, overhaul safety protocols, and redesign the capsule before the moon race continues. 1938

HEADLINE: Meeting the Crew: Borman, Lovell, and Anders. GUEST AUTHOR: Bob Zimmerman. SUMMARY:Profiles of the Apollo 8 astronauts reveal a mix of duty-driven command and exploration zeal, all supported by their resilient families amidst intense media scrutiny. 1968 CREW AND BACKUP CREW

HEADLINE: Leaving Earth: The Historic Launch of Apollo 8. GUEST AUTHOR: Bob Zimmerman. SUMMARY:The Saturn 5 launches successfully, and for the first time, humans leave Earth's orbit, watching their home planet shrink while navigating with primitive computers.

HEADLINE: The Bold Gamble: NASA's Decision to Go. GUEST AUTHOR: Bob Zimmerman. SUMMARY: Facing Soviet competition from Zond missions, NASA managers make the aggressive decision to send Apollo 8 to the moon early without the lunar module. 1968

Watch every episode ad-free & uncensored on Patreon: https://patreon.com/dannyjones Tim Dodd is the creator of  @EverydayAstronaut  educating everyday people on space travel, space exploration, rocket science and much more. Tim has interviewed the most notable figures in the space program including SpaceX's Elon Musk, Former NASA Admin, Jim Bridenstine, Rocket Lab's Peter Beck, Tom Markusic of Firefly, Andy Lapsa of Stoke Space, and many more. SPONSORS https://bloodflow7.com/danny - Hit the link & grab some BloodFlow7 today for 30% OFF. https://irestore.com/dannyjones - Use code DANNYJONES for exclusive savings on the iRestore Elite. https://hexclad.com/dannyjones - Find your forever cookware & get 10% off. https://amentara.com/go/dj - Use code DJ22 for 22% off. https://whiterabbitenergy.com/?ref=DJP - Use code DJP for 20% off EPISODE LINKS @EverydayAstronaut https://everydayastronaut.com FOLLOW DANNY JONES https://www.instagram.com/dannyjones https://twitter.com/jonesdanny OUTLINE 00:00 - The start of Everyday Astronaut 03:29 - Debunking the moon landing hoax 06:36 - Problems with Bart Sibrel's argument 08:08 - Getting to the moon without refueling 12:47 - The reusable rocket challenge 16:21 - Wernher von Bruan's refueling video 23:29 - Why we owe Nazis for rocket technology 30:31 - The upcoming Artemis II mission 37:30 - NASA's headquarters in Alabama 39:40 - Cost-plus contracting & SpaceX funding 41:37 - Boeing astronauts who got stuck in space 44:36 - How many satellites are in orbit 48:07 - Discovery of rocket science 53:00 - New rocket launch technologies 59:17 - Black budget propulsion technology 01:03:54 - Tim Taylor 01:10:33 - Neil Armstrong's expedition after the moon landing 01:13:48 - Best argument we DIDN'T go to the moon 01:16:45 - Explaining NASA's "erased" Apollo mission footage 01:25:46 - NASA's new administrator Jared Isaacman 01:31:16 - Elon Musk's plan for Mars & the moon 01:38:01 - The Van Allen Radiation Belts 01:46:43 - When humans will step foot on the moon again 01:51:45 - Watching a rocket launch in Kazakhstan 02:00:40 - Japanese billionaire buys ticket to fly around the moon 02:11:00 - What Russia & China are launching into space 02:14:56 - Discovery of Chernobyl fungus that eats radiation 02:21:31 - Why SpaceX is perpetually bankrupt 02:24:51 - Starship landing footage 02:31:42 - What Space Force is up to 02:37:18 - SpaceX is intentionally losing satellites 02:38:26 - OSIRIS-REx mission & asteroid space metal mining 02:44:46 - 3I/ATLAS 02:50:34 - AI is burying the truth 02:59:01 - Flat Earth & space mission conspiracies 03:00:51 - Jeff Bezos' Saturn V recovery mission 03:05:29 - Regenerative cooling on rocket boosters 03:09:34 - Why Artemis is going to the moon's south pole 03:14:06 - Mystery aircraft that deflected a missile 03:20:28 - Secrets astronauts are keeping from us 03:28:00 - Why the Apollo post-flight conference is not weird at all 03:36:01 - Debunking Bart Sibrel's "smoking gun" of moon landing hoax 03:42:01 - Inside the Apollo 11 lunar module 03:48:42 - The windows on Apollo 11 03:54:45 - Logistics of faking the moon landing Learn more about your ad choices. Visit podcastchoices.com/adchoices

This week we're looking at some of the greatest comebacks history has to offer. Where else to begin than the epic comeback that is Skoda cars! And what about the comeback of the West German economy after world war 2?! And finally, a literal comeback, Apollo 13's incredible return from the moon after disaster struck.Elsewhere, how good is the tip as a day out? Has history anything better to offer in terms of pure enjoyment? If you know, let us know: hello@ohwhatatime.com And if you want more Oh What A Time, you should sign up for our Patreon! On there you'll now find:•The full archive of bonus episodes•Brand new bonus episodes each month•OWAT subscriber group chats•Loads of extra perks for supporters of the show•PLUS ad-free episodes earlier than everyone elseJoin us at

We have a fresh round of new drill results to report this morning from Collective Mining, Talisker Resources, Pacifica Silver and Azimut Exploration. This episode of Mining Stock Daily is brought to you by... Revival Gold is one of the largest pure gold mine developer operating in the United States. The Company is advancing the Mercur Gold Project in Utah and mine permitting preparations and ongoing exploration at the Beartrack-Arnett Gold Project located in Idaho. Revival Gold is listed on the TSX Venture Exchange under the ticker symbol “RVG” and trades on the OTCQX Market under the ticker symbol “RVLGF”. Learn more about the company at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠revival-dash-gold.com⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Vizsla Silver is focused on becoming one of the world's largest single-asset silver producers through the exploration and development of the 100% owned Panuco-Copala silver-gold district in Sinaloa, Mexico. The company consolidated this historic district in 2019 and has now completed over 325,000 meters of drilling. The company has the world's largest, undeveloped high-grade silver resource. Learn more at⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠https://vizslasilvercorp.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Equinox has recently completed the business combination with Calibre Mining to create an Americas-focused diversified gold producer with a portfolio of mines in five countries, anchored by two high-profile, long-life Canadian gold mines, Greenstone and Valentine. Learn more about the business and its operations at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠equinoxgold.com⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Integra Resources is a growing precious metals producer in the Great Basin of the Western United States. Integra is focused on demonstrating profitability and operational excellence at its principal operating asset, the Florida Canyon Mine, located in Nevada. In addition, Integra is committed to advancing its flagship development-stage heap leach projects: the past producing DeLamar Project located in southwestern Idaho, and the Nevada North Project located in western Nevada. Learn more about the business and their high industry standards over at integraresources.com

Ari Sussman of Collective Mining walks listeners through today's new drill results at Apollo, the flagship deposit within the company's Guayabales Project in Colombia. Today's results expanded the gold and tungsten mineralized horizon close to surface. Ari also provides comments on the latest drill results deep into Apollo at the Ramp Zone, where the company continues to find exception gold grades.

…ON TODAYS PROGRAM… MERCEDES CAUSE PANIC! RIVAL TEAMS LOOK FOR FIA INTERVENTION BEFORE START OF SEASON. ALL EYES ON ADRIAN NEWEY AND ASTON MARTIN'S EXTREME NEWEY DESIGN BLOWING PEOPLES MIND! WILLIAMS COULD BE SAND BAGGING... AND, FERNANDO STILL THINKING OF THE TRIPLE CROWN!! THIS WEEK'S NASIR HAMEED CORNER…MORE VINTAGE BANTER BETWEEN THE HOST AND NASIR…THIS WEEKS SPECIAL GUEST: OLIVIER PANIS! Olivier Panis, originally from Oullins, Lyon, is a former French Formula One driver. Early in his career, Panis began with karting, progressing through several junior series before moving up to the French Formula 3 series. By 1990, he secured 4th place in the championship and achieved runner-up status the following year. After karting, Panis competed in two seasons of F3000. His initial season involved challenges with the Apamotox team's stubborn Lola car, while the second season saw him racing for the highly viewed DAMS Equipe team. His perseverance paid off when he was crowned champion, setting the stage for his entry into Formula 1 with Ligier. At 27, Panis joined the French-based Ligier F1 team in 1994. He secured a surprise second-place finish at Hockenheim that season, ending the season 11th overall in the Drivers' Championship. He continued to impress, securing another unexpected second place at the 1995 Australian Grand Prix, despite trailing two laps behind the leader, and finished 8th in the championship. Panis's most astonishing triumph came at the 1996 Monaco Grand Prix, where he drove his way to victory in treacherously wet conditions. It marked Ligier's first win in 15 years—their last—and was the first French victory in a French car at Monaco in 66 years. However, apart from this win, Panis failed to finish higher than fifth for the remainder of the season. In 1997, racing for Prost, who had bought Ligier, Panis showed promise, placing third in the championship standings after six races. Unfortunately, a crash in Canada broke his leg, sidelining him for eight races. He returned for the season's last three races and finished ninth in the championship. The 1998 season was less successful for Panis, who struggled to score points under Prost's management. He earned only a single point across the following season, leading to the end of his relationship with the team. Panis then considered an offer from Williams but opted to test for McLaren instead, which kept his presence in the paddock despite a full-time drive. He joined BAR in 2001, although the team didn't meet his expectations, finishing 14th for two consecutive seasons. In 2003, Panis moved to the new Toyota team to provide his experience and mentor his teammate, Cristiano da Matta. Although he improved in qualifying, his overall results mirrored his previous seasons, finishing 14th once again. Panis continued with Toyota through 2004, his tenth year in Formula One. He announced his retirement in October of that year, effective after the 2004 Japanese Grand Prix. He stayed with Toyota as a test driver through 2005 and 2006, ending his F1 career at age 37, with five podiums and 76 career points from 157 starts. Olivier Panis Formula One World Championship career. F1 Career 1994–1999, 2001–2004 Teams Ligier, Prost, BAR, Toyota Entries 158 (157 starts) Championships 0 Wins 1 Podiums 5 Career points 76 Pole positions 0 Fastest laps 0 First entry 1994 Brazilian Grand Prix First win 1996 Monaco Grand Prix Last win 1996 Monaco Grand Prix Last entry 2004 Japanese Grand Prix Olivier Panis Teammates 13 Teammates Involvement First Year Last Year Eric Bernard 13 1994  Johnny Herbert 1 1994  Franck Lagorce 2 1994  Aguri Suzuki 6 1995  Martin Brundle 11 1995  Pedro Diniz 16 1996  Shinji Nakano 10 1997  Jarno Trulli 34 1998 2005 Jacques Villeneuve 34 2001 2002 Cristiano da Matta 28 2003 2004 Ricardo Zonta 16 2004  Ryan Briscoe 5 2004  Ralf Schumacher 1 2005 HSR Pistons and Props Presented by the Alan Jay Automotive Network Returns to Sebring February 13-15. SEBRING, Fla. (Feb. 5, 2026) – Historic Sportscar Racing (HSR) Pistons & Props Presented by the Alan Jay Automotive Network kicks-off the 2026 HSR racing season next weekend at Sebring International Raceway, Feb. 13-15. The must-attend event once again celebrates Sebring's rich sports car racing heritage and notable aviation history with four days of on-track action and an airplane "fly-in" of retro civilian and military aircraft from the World War II era and last half century. HSR Pistons & Props Presented by the Alan Jay Automotive Network honors the legendary Mobil 1 Twelve Hours of Sebring sports car race, which runs for the 74th time March 21, and Sebring International Raceway's patriotic aviation history. Hendricks Field, on which Sebring International Raceway stands, was built as a United States Army Air Forces training base during World War II. One plane scheduled to appear is a Beechcraft T-34 Mentor owned and piloted by Bob Hahnemann, who could be the first HSR Pistons & Props participant to take part in both the winged and four-wheel activity. An accomplished pilot and sports car racing competitor, Hahnemann is listed as a co-driver with his son, Matt Hahnemann, in Friday afternoon's B.R.M Chronographes Legacy Enduro in their 2007 No. 111 Porsche 997 GT3 Cup car. Just after the race, Bob will taxi from the adjacent Sebring Regional Airport down the raceway's Ulmann Straight (backstretch) in the T-34, joining a quality lineup of other must-see airplanes and accomplished pilots in a parade to the paddock. Positioned inside the Sebring paddock, the planes will be on display and available for viewing from Friday at 4:30 p.m. through late morning on Sunday. The Beechcraft T-34 Mentor was a post-World War II trainer that was a learning workhorse for thousands of cadets for more than 25 years.  It was used in the Air Force until the 1960s and a go-to in the Navy well into the 1970s. The senior Hahnemann and his partner, Len Tucker, purchased the plane four years ago from legendary NASA astronaut and United States Air Force Colonel Frank Borman, Commander of Apollo 8.  Apollo 8 was the first mission to fly around the Moon. Also a test pilot – and former President of Eastern Airlines – Borman put his own high-performance enhancements on the T-34, installing a Continental IO-550, which was the largest engine you could put in a Mentor. The twin "SU" lettering as the plane's nickname – SU SU IX – also continued Borman's tradition of using the first letters of his wife Susan's name on his aircraft. On the HSR competition side, a highlight of the overall entry list is a nice turnout of entries in the HSR Sasco Vintage Cup for Groups 2 and 3. Home to small-bore racing machines that deliver big-time competition, Sasco Vintage Cup features many unique and eclectic race cars. One particularly rare entry is the Olthoff Racing 1960 No. 26 GSM Dart driven by Englishman John Spiers.  The GSM was built in South Africa by Glass Sport Motor company.  The company, which manufactured the Dart from 1959 until 1962, got its name – Glass Sport – given its use of fiberglass.  The lightweight production sports cars were generally used for racing. The No. 26 has been modified to feature a full flip-top front end and left-hand drive. Power comes from a Ford 1600 Kent engine – produced in Kent, England – with twin side-draft carburetors. Spiers will battle with a top trio of British-built Ginettas, including frequent HSR race winner and podium finishers Hervey Parke in his 1965 No. 11 Ginetta G4 prepared by Michael's Vintage Racing. Michael Oritt drives a similar 1961 No. 82 Ginetta G4 while Thomas Grudovich completes the quick Ginetta contingent in his 1966 No. 425 Ginetta G4. Another favorite small-bore British contender could be the comeback story of the weekend. Accomplished HSR driver Kenneth Greenberg was uninjured in a heavy Turn 1 accident in December's season-ending HSR event at Sebring, but his Air Power Racing 1964 No. 324 Morgan Plus 4 was nearly a total write off. Weston Farmer and the team at Air Power quickly went to work non-stop, and Greenberg and the Morgan are entered in the Vintage Cup sprints and B.R.M Legacy Enduro. Farmer reports many hours are still ahead before traveling to Sebring next week from the team shop in St. Augustine, Fla. after the Morgan's frame was destroyed and even the engine block was cracked in the incident. The team bought a similar 1967 Morgan chassis as a donor car, and the roll cage was completed last week. Oil lines, fuel lines and electrical systems are going in this week and a rebuilt engine recently arrived. For complete information on HSR Sebring Pistons & Props Presented by the Alan Jay Automotive Network, including the event schedule and entry lists, visit www.HSRrace.com/sebring-pistons-and-props.  For tickets, visit www.SebringRaceway.com.  

The Space Show Presents Mark Whittington, Sunday, Feb. 1, 2026Quick Summary:Our program began with followed by conversations about the Artemis II mission and space technology advancements. The group explored various aspects of space exploration, including Iran's space program, lunar missions, and the development of space suits and launch systems. The discussion concluded with conversations about AI in journalism, nuclear reactor technologies, and the upcoming Artemis II mission, including its potential for public celebration and media coverage.Detailed SummaryOur Zoom program Wisdom Team discussed the Artemis II mission, with Mark expressing confidence in its success. They also talked about the challenges of keeping up with rapidly advancing technology and the stock market, particularly in the space industry. David expressed interest in doing a show about space-related ETFs but felt that Andrew, who focuses on his own ETF, might not be the best person for it given there now a plethora of space focused ETFs.After introducing Mark Whittington as the guest for the Sunday space show, we discussed Iran's space program. Mark explained that while Iran's space program is not robust by global standards, it serves dual purposes, potentially threatening global security. He suggested that if Iran were to transition to a more freedom-oriented regime, preserving and developing their space program could be beneficial for economic growth and inspiring young Iranians to pursue STEM subjects.Mark continued talking about Iran's space program, highlighting its development of launch vehicles and satellites, and its potential to support a nuclear program. He suggested that a post-Islamic Republic Iran could benefit economically from continuing its space program and joining international initiatives like the Artemis Accords. David inquired about the U.S. government's perspective on Iran's space program, and Mark noted that concerns primarily focus on its nuclear capabilities. They briefly touched on the potential for military action against Iranian spaceports and the possibility of Reza Pahlavi returning as a stabilizing figure in Iran. The conversation concluded with a brief mention of the upcoming Artemis II mission.Mark discussed the Artemis II mission in some detail, addressing concerns about the heat shield and NASA's confidence in its workarounds. He highlighted the mission's potential impact on American society, comparing media coverage then and now, and expressed hope that Artemis II would be a significant story. Mark also noted the mission's duration of 10 days, including a loop around the moon, and emphasized the diversity of the crew. David shared a question from a listener about potential lunar payload or surface interaction during Artemis II, which Mark clarified does not involve landing on the moon but rather a loop around it.We continued focusing on the Artemis program and lunar exploration. Mark explained that CubeSats will launch with Orion but won't be lunar landers, and SpaceX and Blue Origin are developing lunar landers. The next Starship test is expected in 5 weeks, aiming to refuel in low Earth orbit and land on the moon. John Jossy inquired about ESA's life support system test, which Mark confirmed is part of the systems test in low Earth orbit. Marshall asked about reusability of the Space Launch System's solid rocket boosters, to which Mark replied they are not planned to be reusable due to the infrequent use of the system.Our team discussed public excitement and historical significance of the Artemis II mission, with Mark noting that while Artemis II should be the story of the year, public awareness and support may be lower than during the Apollo missions. They discussed the upcoming Starship test in 5 weeks as a potential rival for public attention, and explored the possibility of live TV coverage during the mission, including the crew's perspective of Earth rise. Mark suggested that the crew should be allowed to express their personal thoughts and experiences during the mission rather than following a pre-determined script.Our team also discussed the inspirational impact of Earthrise, with Marshall sharing his perspective on the spiritual connection to celestial objects. Mark and Marshall exchanged views on the role of mathematics and physics in understanding the universe, while David inquired about public perceptions of space exploration versus Earth's economic concerns. Mark highlighted the potential economic benefits of space travel and SpaceX's upcoming IPO, as well as Elon Musk's plans for AI data centers in orbit, powered by a network of satellites. The conversation concluded with a discussion on the development of space-based solar power and its potential to overcome the limitations of solar energy on Earth, with our guest emphasizing the importance of diverse energy sources like nuclear and natural gas.Mark discussed the development of lunar spacesuits, noting that Axiom Space is the prime contractor and progress is ongoing, with suits expected to be ready by 2028. He also addressed the potential merger between SpaceX and Tesla, suggesting it would create a holding company with separate divisions, and discussed the development of Optimus robots for space exploration. Mark highlighted the success of NASA Administrator Bill Nelson, who was confirmed after a tumultuous process, and expressed optimism about Artemis II's upcoming launch and its potential to generate momentum for future space missions. He also touched on the challenges of transitioning from SLS to commercial systems for future Artemis missions, noting that while there are concerns about delays, the goal remains to build a lunar base.Our group discussed historical SpaceX launch operations, particularly focusing on the Falcon 1 rocket launches from Kwajalein Atoll in the Pacific. David explained that while Kwajalein offered advantages like reduced orbital velocity requirements due to its equatorial location, the 8,000-mile supply line and classified military operations at the site ultimately led SpaceX to shift operations to other locations including Vandenberg and Boca Chica. The discussion concluded with Marshall recalling the Celestas Memorial payload incident, where a Falcon 1 rocket failed to reach orbit and instead crashed into the Marianas Trench, though the exact crash location was never publicly disclosed by SpaceX.John Hunt proposed a fallback plan for Starship's on-orbit refueling, involving an unmanned variant with an exploration upper stage as a third stage, to reduce costs for lunar missions. Mark and David discussed the potential of nuclear power and propulsion in space, as well as the integration of such technologies into Starship for Mars missions. David emphasized the importance of having a plan for implementing new ideas, cautioning against presenting alternatives without a clear path forward. The conversation also touched on the role of AI in managing information overload for executives like Elon Musk, with Marshall sharing insights from his experience with AI in research and business.Mark described the limitations and potential of AI in journalism, emphasizing the need for human oversight in verifying sources. Ajay shared information about advanced nuclear reactor technologies, including Generation 4 reactors and molten salt reactors, highlighting their safety features and reduced waste production. Mark expressed interest in learning more about these reactors. The group agreed to continue the discussion if time permitted, with Mark mentioning his upcoming focus on the Artemis II mission and other space-related stories.Mark continued promoting the upcoming Artemis II mission, which is scheduled for a wet dress rehearsal followed by a potential launch on February 8th, after the Super Bowl. He shared his experience as a space writer and author, mentioning his books about lunar exploration. The group discussed the potential for a ticker tape parade and public celebration if the mission is successful, with Mark agreeing to write about this possibility in his Sunday newsletter. Dr. Ajay and others expressed interest in subscribing to Mark's newsletter, which is distributed through The Hill newspaper. David did not think a ticker tape parade was in the cards, so to speak.Special thanks to our sponsors:American Institute of Aeronautics and Astronautics, Helix Space in Luxembourg, Celestis Memorial Spaceflights, Astrox Corporation, Dr. Haym Benaroya of Rutgers University, The Space Settlement Progress Blog by John Jossy, The Atlantis Project, and Artless EntertainmentOur Toll Free Line for Live Broadcasts: 1-866-687-7223 (Not in service at this time)For real time program participation, email Dr. Space at: drspace@thespaceshow.com for instructions and access.The Space Show is a non-profit 501C3 through its parent, One Giant Leap Foundation, Inc. To donate via Pay Pal, use:To donate with Zelle, use the email address: david@onegiantleapfoundation.org.If you prefer donating with a check, please make the check payable to One Giant Leap Foundation and mail to:One Giant Leap Foundation, 11035 Lavender Hill Drive Ste. 160-306 Las Vegas, NV 89135Upcoming Programs:Broadcast 4502 Zoom Steve Wolfe, Elizabeth Change | Tuesday 10 Feb 2026 700PM PTGuests: Steven WolfeZoom: Steve Wolfe , Elizabeth Change on the Beyond Earth Upcoming Symposium and more BE newsBroadcast 4503: Hotel Mars with Rick Fisher | Wednesday 11 Feb 2026 930AM PTGuests: Rick Fisher, John Batchelor, Dr. David LivingstonRick Brings us news regarding United States and China are also locked in a contest regarding Solar System domination between China and the USBroadcast 4504 Zoom: Frank Pietronigro | Friday 13 Feb 2026 930AM PTGuests: Frank PietronigroZoom: Frank discusses the Zero Gravity Arts Commission and moreBroadcast 4506 Zoom Open Lines | Sunday 15 Feb 2026 1200PM PTGuests: Dr. David LivingstonOpen Lines discussion. All topics welcome Get full access to The Space Show-One Giant Leap Foundation at doctorspace.substack.com/subscribe

(0:00) Intro(2:04) About the podcast sponsor: The American College of Governance Counsel(2:50) Start of interview(3:51) Betsy's origin story(9:14) The HealthSouth Board Scandal(16:35) Her preference when picking what boards to serve on(17:30) Insights VC-backed Boards and role and profile of the independent director in this context(21:20) Insights on PE-backed Boards and role and profile of the independent director in this context(25:35) Navigating International Board Dynamics. Her experience on boards of Volvo and Schneider Electric.(30:57) The Rise of Private Markets. Example of Atlas Air (Apollo backed). IPOs in 2026.(35:07) AI's Impact on the Market and other macro trends(38:10) Founder-Led Companies and Governance (including dual-class share structures).(42:25) The Impact of Geopolitics on Governance(45:11) The Impact of Politicization on Governance. Examples of Budweiser, Google, Netflix, and the mission-driven approach by Coinbase.(50:09)  Adapting to Accelerating Change as Directors. The problem with incrementalist "custodian" directors in times of disruption. "It's really about being change-adaptive and comfortable making decisions with incomplete information. You look at someone like Musk, he's making decisions when he has 60% of the information. Most boards want 95% before they'll move. That's the fundamental challenge."(55:58) Books that have greatly influenced her life ("the best business book"):Good to Great, by Jim Collins (2001)(56:16) Her mentors. Craig Billings (CEO Wynn Resorts), Michael Steen (CEO Atlas Air Cargo), Jean-Pascal Tricoire (Chairman, Schneider), her mom ("her biggest mentor").(57:06) On the current state of shareholder activism(57:58) Quotes that she thinks of often or lives her life by "Perfect is the enemy of good enough." (58:19) An unusual habit or an absurd thing that she loves: she's a compulsive note-taker (plus, her recommended policy for directors)(1:00:12) The living person she most admires: Elon MuskBetsy Atkins has served on more than 38 public company boards and through 17 IPOs, in addition to scores of PE and VC-backed company boards. She brings a rare perspective shaped by crisis situations, international board service, and rapid technological change. She currently serves on the boards of Wynn Las Vegas, GoPuff, and the Google Cloud Advisory Board. You can follow Evan on social media at:X: @evanepsteinLinkedIn: https://www.linkedin.com/in/epsteinevan/ Substack: https://evanepstein.substack.com/__To support this podcast you can join as a subscriber of the Boardroom Governance Newsletter at https://evanepstein.substack.com/__Music/Soundtrack (found via Free Music Archive): Seeing The Future by Dexter Britain is licensed under a Attribution-Noncommercial-Share Alike 3.0 United States License

Bei der Apollo-11 Mondlandung 1969 war Frank Sinatra mit "Fly Me to the Moon" symbolisch mit dabei. Angeblich befand sich der Song auf einer Kassette von NASA-Astronaut Buzz Aldrin. Roland erzählt, wie er als 8-Jähriger mit roten Wangen vor dem TV saß.

China is reportedly urging banks to curb USTs exposure amid market risk, Bloomberg reports, citing sources; guidance does not apply to China's state holdings of US Treasuries.Japanese PM Takaichi's LDP party won a landslide victory at the snap election on Sunday, securing a super majority; JPY bid, JGBs lower and Nikkei 225 soars.European bourses are broadly firmer, whilst US equity futures move lower; Nikkei 225 soars post-LDP victory.USD hit on China-USTs report, JPY strengthens post-LDP, whilst GBP lags on regional political woes.JGBs set a bearish tone for global fixed income, with USTs also dragged on the China-USTs report; Gilts digest the McSweeney resignation and reports that PM Starmer faces further pressure to resign.WTI and Brent are flat. Precious metals continue to rebound as the PBoC buys gold for a 15th consecutive month.Looking ahead, highlights include US Consumer Inflation Expectations (Jan), BoC Market Participants Survey. Speakers include ECB's Lane & Lagarde, Fed's Waller & Bostic, Earnings from Apollo, Becton Dickinson, Loews, On Semiconductor & Cleveland-Cliffs.Read the full report covering Equities, Forex, Fixed Income, Commodites and more on Newsquawk

APAC stocks began the week higher after last Friday's rally on Wall St, where the DJIA topped the 50k level for the first time.The Nikkei 225 also hit a fresh record high after PM Takaichi's landslide election victory and supermajority.China is reportedly urging banks to curb US Treasuries exposure amid market risk, Bloomberg reports, citing sources; guidance does not apply to China's state holdings of US Treasuries.European equity futures indicate a positive cash market open with Euro Stoxx 50 futures up 0.4% after the cash market closed higher by 1.2% on Friday.Highlights include Swiss Consumer Confidence (Jan), Norwegian GDP (Q4), Mexican Inflation (Jan), US Consumer Inflation Expectations (Jan), BoC Market Participants Survey. Speakers include ECB's Lane & Lagarde, Fed's Waller & Bostic, Earnings from Apollo, Becton Dickinson, Loews, On Semiconductor & Cleveland-Cliffs.Read the full report covering Equities, Forex, Fixed Income, Commodites and more on Newsquawk

What if the biggest lever you have today isn't another action plan—but one decision? In this episode, Bill Sherman talks with Apollo Emeka, who calls himself "the big decisions guy," and traces how that identity started early—when Apollo was effectively handed the power to choose school or not as a kid, and felt the real-world consequences of deciding either way. Apollo's path is anything but linear: military service, Iraq deployment, an FBI internship, and a mindset shaped by high-stakes environments where "what could go wrong?" isn't drama—it's a discipline. He shares a vivid example: after his family was impacted by the Eaton fire in Altadena and evacuated, they stress-tested a radical idea (moving to Panama) by asking that question seriously, researching risks, and acting fast once no deal-breakers showed up. A turning point came when Apollo commissioned a third party to interview his clients and surface where his real impact was. The message was consistent: decision-making. That clarity gave him permission to drop the "other consulting stuff" and go all-in on helping leaders make better decisions faster—then validating the shift publicly and operationally (including flipping his website). You'll hear practical tools, not theory. Apollo describes how most leaders' stated goals score shockingly low on a fulfillment scale—often a 6 or 7—because they're inherited, socially pressured, or "sensible," not energizing. That insight becomes the doorway to choosing goals you actually want, not goals you can defend. He also lays out what he calls a "big decision" framework: it must be a 10/10 on fulfillment, read like a toddler's run-on sentence (because it forces your competing life priorities onto the same page), make other decisions easier, and be bold enough that people might call you crazy. Apollo reads his own big decision statement—including the ambition to build scale through a best-selling book, a top podcast, and bigger stages, while protecting what matters at home. Finally, Apollo names the hidden saboteurs that keep smart people stuck: the "decision monsters." He trains clients to stop living in "can / should / could," and to recognize three common blockers—feasibility, worthiness, and social judgment—so leaders can choose with intention instead of permission. Three Key Takeaways: • Make one "big decision" that simplifies everything else. A real big decision is designed to be high-fulfillment (a 10/10), bold enough to feel uncomfortable, and specific enough that future choices get easier because they can be measured against it. • Stop chasing goals you can defend and start choosing goals you actually want. Apollo argues many leaders rate their current goals at only a 6–7 on fulfillment because they're inherited, socially expected, or "sensible." The fix is to re-select goals based on energy and meaning—not optics. • Name the "decision monsters" before they run the meeting in your head. He calls out the common traps—living in "can/should/could," fear about feasibility, doubts about worthiness, and worry about social judgment. Once you label the blocker, you can choose directly instead of negotiating with it. If this week's episode got you thinking about making one clear decision that cuts through noise, you'll get even more value from Lee Caraher's conversation—because it lives in the same territory: clarity under pressure and the choices leaders make when the old playbook stops working. Lee digs into how to lead across generations without the drama, how to shift your approach when talent and expectations change, and what to do when a business model needs a reset. Listen to sharpen your decision filters, reduce second-guessing, and walk away with practical moves you can use immediately.

“Hybrid is the ultimate in creative solutions,” says Apollo's co-head of equity and head of hybrid, Matt Nord, when discussing the firm's fast-growing hybrid strategy and the types of solutions that may allow it to triple in scale over the coming years. Nord joins Bloomberg Intelligence's Noel Hebert on the latest episode of Credit Crunch to talk about controlling for downside, the benefits of scale and why the lack of realizations across private equity is creating a need for more tailored funding solutions. The two also discuss Apollo's investment ethos, including its value orientation and pursuit of “excess return per unit of risk.” The Credit Crunch podcast is part of BI's FICC Focus series.

Welcome to The Daily Wrap Up, an in-depth investigatory show dedicated to bringing you the most relevant independent news, as we see it, from the last 24 hours (2/6/26). As always, take the information discussed in the video below and research it for yourself, and come to your own conclusions. Anyone telling you what the truth is, or claiming they have the answer, is likely leading you astray, for one reason or another. Stay Vigilant. !function(r,u,m,b,l,e){r._Rumble=b,r[b]||(r[b]=function(){(r[b]._=r[b]._||[]).push(arguments);if(r[b]._.length==1){l=u.createElement(m),e=u.getElementsByTagName(m)[0],l.async=1,l.src="https://rumble.com/embedJS/u2q643"+(arguments[1].video?'.'+arguments[1].video:'')+"/?url="+encodeURIComponent(location.href)+"&args="+encodeURIComponent(JSON.stringify([].slice.apply(arguments))),e.parentNode.insertBefore(l,e)}})}(window, document, "script", "Rumble");   Rumble("play", {"video":"v737u4c","div":"rumble_v737u4c"}); Video Source Links (In Chronological Order): Saudi Arabia's Airstrikes in Yemen Killed at Least 13 Civilians in January - News From Antiwar.com Rate of Israeli Strikes on Lebanon at Highest Level Since Ceasefire - News From Antiwar.com US Launches Its 27th Airstrike in Somalia of the Year - News From Antiwar.com Trump is blasting away at Somalia with zero effect | Responsible Statecraft Pentagon Inks Massive $200 Million Deal to Buy Israeli Cluster Weapons (17) Ryan Rozbiani on X: "Every Iranian and American MUST WATCH THIS NEW

Jeffrey Epstein's entanglement with Leon Black and Larry Summers runs through the Jeffrey Epstein VI Foundation and its flagship project, the Institute for New Economic Thinking (INET), born out of the wreckage of the 2008 financial crisis. Black, the billionaire Apollo founder, bankrolled INET with roughly $25 million and installed himself as its chief patron, while Summers — fresh off his controversial presidency at Harvard and a career bouncing between Wall Street and Washington — became one of its intellectual faces. Epstein, already a convicted sex offender by 2008, quietly emerged as a financial conduit and behind-the-scenes broker for INET and its affiliates, using donor networks, shell foundations, and elite access to move money and cultivate influence. Through Epstein's foundation, funds were routed into academic projects, conferences, and research hubs that placed him back inside elite academic circles that had supposedly shut him out, laundering his reputation through economics, philanthropy, and intellectual respectability.What makes the IPI/INET web so corrosive is how thoroughly it fused money, power, and reputational cover. Black would later admit paying Epstein $158 million for “tax advice,” an explanation so implausible it collapsed under its own weight, while Summers maintained institutional ties to projects and donors connected to Epstein long after his 2008 conviction was public record. Epstein was not a peripheral donor — he was a facilitator, recruiter, and fixer who connected hedge-fund money, Ivy League legitimacy, and political access in a closed loop that insulated all participants from scrutiny. The IPI ecosystem gave Epstein exactly what he needed after Florida: proximity to young academics, international travel, visa sponsorships, and an elite shield that made him look like a disgraced financier turned reformed intellectual benefactor. It wasn't an accident, and it wasn't ignorance — it was a deliberate system where billionaires, former Treasury secretaries, and a convicted predator all found mutual benefit inside the same polished academic machine.to contact  me:bobbycapucci@protonmail.com

According to newly reported emails between Jeffrey Epstein and Leon Black, Epstein pressed Black with aggressive financial demands for years, particularly around 2015 to 2016. Epstein repeatedly insisted on annual payments of roughly US$40 million for providing tax-and-estate-planning services, seeking an upfront US$25 million plus multiple US$5-million bi-monthly installments. He chastised Black's children and financial advisers, calling them incompetent and saying that their actions had created a “really dangerous mess.”While Black had engaged Epstein for advisory services and reportedly paid over US$150 million over a period of time, the correspondence underscores how Epstein sought to impose unusually high compensation and used personal attacks and pressure tactics. Black maintains that Epstein's role was limited to legitimate financial work, and investigations (such as the independent review by law firm Dechert LLP) found no conclusive wrongdoing by Black, though substantial payments and tax-planning strategies remain under scrutiny from the U.S.   to contact me:bobbycapucci@protonmail.comsource:Jeffrey Epstein sent nasty emails to Apollo founder Leon Black demanding millions of dollars

Rissi Palmer, the GRAMMY-nominated singer-songwriter who has performed at the White House, the Apollo and the Grand Ole Opry, shares her newest EP, “Perspectives,” and previews this weekend's Opry 100 celebration.Become a supporter of this podcast: https://www.spreaker.com/podcast/tavis-smiley--6286410/support.

Snart ska en raket iväg som är tänkt att ta människor längre bort från jorden än någonsin. Nästa steg är att åter landa på månen, och allt fler länder ger sig in i den nya rymdkapplöpningen. Lyssna på alla avsnitt i Sveriges Radios app. NASA:s raket som står på startplattan i Florida är den största hittills, och den har aldrig tidigare lyft med en besättning. Detsamma gäller Orion-kapseln, där de fyra astronauterna ska sitta. Dessutom ligger europeiska ESA bakom en viktig del, för första gången. Men varför är månen, som ingen människa satt sin fot på sedan 1972, nu åter så intressant? Vi hör svaren på det, hur resorna Artemis II och III ska gå till, och hur det hela också är en förberedelse för bemannade turer till Mars.Medverkande: Kristine Dannenberg, ansvarig för utforskande och tillträde till rymden, Rymdstyrelsen.Reporter: Björn Gunérbjorn.guner@sr.seProducent: Lars Broströmlars.brostrom@sr.se

On Thursday's show: City of Houston Controller Chris Hollins discusses the city's deficit spending, due in part to the continuing need to pay overtime to staff police and fire departments.Also this hour: The delay in the Artemis II mission stirs memories of how the Apollo program played out decades ago. We look back on our last missions to the moon.Then, the nonprofit Trees for Houston celebrates the planting of its one millionth tree today. We visit the site of that celebration.And a recent report finds incomes are growing and poverty is shrinking, but living in Texas is still less affordable. We talk about why that's the case.Watch

"Every year I hear people complaining that the NFL makes lousy picks for the Super Bowl halftime show. If the picks are lousy then ratings must tank. But they do not. In fact the halftime show has never been better watched. We have a long list of ratings and demographics to show that the NFL seems to know what they are doing."

"We are coming into tax season so Tammy and will talk about paying the government. The HITS Act is now in full swing. Foreign governments are changing their tax codes for musicians and we also have a list of what you might not have known was tax deductible."

Back on this day in 1971 Apollo 14 astronauts walked on the moon. This was the third time astronauts landed on the moon.

Unlock voice over jobs hiding in your hometown. Most actors miss these secret, local VO goldmines!Discover proven steps to land high-paying, local voice over gigs, even if you're just getting started.Say goodbye to endless online auditions and learn actionable strategies to build a thriving VO career in your own community.Get your FREE Local Gig Kickstart Toolkit: https://welcome.vopro.pro/local-gig-kickstart-toolkit#voiceover #voiceacting #localgigs #VOPro #careeradvice #actorlife #auditiontipsLinks: (When possible, I use affiliate links and may earn a commission. See disclosure below.)▶️ Subscribe: https://vopro.pro/youtube

Kevin Nealon (new special Loose in the Crotch on YouTube! Hiking with Kevin!) returns to make it weird again! Go to quince.com/weirdo for free shipping on your order and 365-day returns Sign up and get 10% off at betterhelp.com/weirdo If you want to support your stress and sleep—not just track it—Apollo is worth trying. For a limited time, get $99 off the Apollo Wearable + SmartVibes bundle at apolloneuro.com/weird with code WEIRDSee omnystudio.com/listener for privacy information.

This edition of the IPS DEPROGRAM, recorded on February 2, 2026, explores the intersections of ancient holidays, systemic media deception, and the "manufacturing of history" through space programs and entertainment. The host argues that the public is being conditioned through predictive programming and manufactured news events to accept a controlled, "hyper-reality".The Ritual of Groundhog Day and Candlemas: The host connects Groundhog Day to the Christian holiday Candlemas and the pagan holiday Imbolc, marking the midpoint between the solstices. He dismisses the groundhog as a "ceremonial prop," claiming the weather outcome is actually determined by a human committee.Moon Mission Skepticism (Artemis II vs. Apollo): Drawing a parallel between the current Artemis II mission and the Apollo missions, the host questions why modern space agencies lack the confidence to land on the moon today if they achieved it decades ago. He posits that the original missions were "manufactured miracles" and that NASA is currently waiting for more sophisticated generative AI and LED technology to pull off a more convincing fake landing.Predictive Programming and Fictional Counterparts: The broadcast analyzes how world leaders and events are foreshadowed in movies. For instance, characters in films like The Madness or Megalopolis are seen as media counterparts to figures like Elon Musk and Donald Trump, used to habituate the public to specific narratives like assassinations or "mind viruses".The Rebrand of "Trutherville" and Content Theft: The host critiques the "alt-media" ecosystem, specifically targeting "content thieves" on platforms like X who regurgitate information from smaller channels without credit. He emphasizes the need for a new branch of media outside of both controlled mainstream and alternative outlets.Systemic Fakery and Psychological Condition: A significant portion of the show discusses how hoaxes—such as the Charlie Kirk "assassination" event—are used as litmus tests for public discernment. The host argues that the public is trapped in a "Mystery Science Theater 3000" scenario, forced to watch "bad science fiction" until they lose their ability to distinguish reality from theatre."Obviously, the groundhog is just a ceremonial prop. It doesn't really determine anything. It's just tradition.""My theory as to why they haven't landed on the moon again in all these decades is that the rear screen projection wouldn't fool anybody today.""History is lies agreed upon... Space with its angels, which we call aliens, is an acceptable version of heaven that everybody agrees on.""The correct answer is entertainment is synchronized with fake news events... they're manufacturing history.""We're collectively in the Mystery Science Theater 3000. Forced by evil scientists to watch bad B-movies until we break."Clarifying Question: Are you looking for a more detailed analysis of the specific "number codes" mentioned (like 33 and 113) for a research project, or should I focus on the media deconstruction aspect for a general summary?Topic BreakdownNotable Quotes

In this episode, Scott Becker reviews the YTD performance of the largest private equity and alternative asset managers, ranking Carlyle, Apollo, Blackstone, TPG, and KKR from best to worst.

Once Upon a Time in the world of professional boxing....Apollo Creed was the champion of the world.  He eventually lost his title to Rocky Balboa (Sylvester Stallone who would receive an Oscar nomination for his performance in this film) who would also lose the title before winning it back with Apollo as his trainer.  They remained very good friends right up until the day that Apollo tragically lost his life during an exhibition match against the feared Russian boxer, Ivan Drago.  But did you know that Apollo actually had a son right before he died??That son is Adonis Johnson (Michael B. Jordan) and he has now become a boxer himself to the consternation of his mother (Phylicia Rashad), only he has been struggling to find some one who can train him.  So one day he travels across the country to Philadelphia to seek out the former champion whom his father trained when he needed him the most.....and after some hesitation, Rocky decides to train him as he works his way up the professional boxing circuit.  Along the way, Adonis also finds a companion in Philly with the lovely singer Bianca (Tessa Thompson) when suddenly, he has the opportunity to fight for the SAME boxing title which his father previously had.  What results is one of the most beloved sports dramas of recent years directed by future Oscar-nominee Ryan Coogler (Black Panther, Sinners) as his first major studio film.  Host: Geoff GershonEdited By Ella GershonProducer: Marlene Gershon  Send us a textSupport the showhttps://livingforthecinema.com/Facebook:https://www.facebook.com/Living-for-the-Cinema-Podcast-101167838847578Instagram:https://www.instagram.com/livingforthecinema/Letterboxd:https://letterboxd.com/Living4Cinema/

In this episode of SpaceTime, we explore new insights into the origins of Earth's water, groundbreaking discoveries beneath the surface of Jupiter's volcanic moon Io, and how tectonic plate movements may have influenced Earth's climate throughout history.New Clues on Earth's Water OriginsA recent study published in the Proceedings of the National Academy of Sciences reveals that asteroid and comet impacts could only account for a small fraction of Earth's water supply. By analyzing oxygen isotopes in lunar regolith collected during the Apollo missions, researchers found that the early Earth likely retained little to no water during its formative years. This challenges long-held beliefs and suggests that the majority of Earth's water must have originated from other sources, rather than being delivered by celestial bodies.Unprecedented Volcanic Activity on IoNASA's Juno spacecraft has captured remarkable data on Io, the most volcanically active body in our solar system. Observations from a December flyby revealed the most energetic eruption ever detected on Io, affecting a vast area of 65,000 square kilometers. The findings indicate that interconnected magma reservoirs beneath Io's surface are responsible for this extraordinary volcanic activity, providing new insights into the moon's geological dynamics and evolution.Tectonic Plates and Earth's ClimateA new study suggests that carbon released from shifting tectonic plates may have played a significant role in Earth's climatic transitions, rather than volcanic activity as previously thought. Researchers reconstructed carbon movements over the last 540 million years, providing evidence that carbon emissions from mid-ocean ridges were the primary drivers of climate shifts between ice ages and warmer periods. This research reshapes our understanding of past climate dynamics and offers valuable insights for future climate models.www.spacetimewithstuartgary.com✍️ Episode ReferencesProceedings of the National Academy of SciencesJournal of Geophysical Research PlanetsCommunications Earth and EnvironmentBecome a supporter of this podcast: https://www.spreaker.com/podcast/spacetime-your-guide-to-space-astronomy--2458531/support.

Matt is joined by Dhara Patel from the National Space Centre to explore Artemis II ; the first crewed mission around the Moon in over 50 years. They discuss Europe's crucial role in the mission, the historic “firsts” represented by the Artemis crew. Matt also takes a quick look at the remarkable engineering heritage behind Orion's propulsion system including Shuttle-era engines and chemical propellants whose story stretches back to Apollo.

Cristina Gomez discusses Stanford professor Dr. Garry Nolan's shocking lab discovery involving UAP materials with impossible isotope ratios, his claims of White House threats, and the classified truth behind the Clementine lunar mission, including Apollo 14 astronaut Edgar Mitchell's private confession about why speaking publicly would be treason.To see the VIDEO of this episode, click or copy link - https://youtu.be/Z7un7zxRWI0Visit my website with International UFO News, Articles, Videos, and Podcast direct links -www.ufonews.co0:00 - UFO Stories That Connect0:52 - Beings In His Bedroom2:05 - UFO Material Sat For 5 Years3:38 - Isotopes Beyond Human Tech5:14 - Non-Humans Were Here First6:56 - NASA Hides Evidence8:04 - Lunar Bases On The Moon9:46 - Astronaut Said TreasonBecome a supporter of this podcast: https://www.spreaker.com/podcast/strange-and-unexplained--5235662/support.

America's Favorite Rocket Scientist™ and Big Brain Smart Head™ Bob Luzenski joins Rafe to discuss NASA's official remembrances during this time of year and hopes and aspirations for the upcoming Artemis II launch.  After some solemn discussions of the fallen astronauts from the Apollo 1 fire of 1967, the Challenger accident of 1986, and the Columbia accident of 2003, discussion turns to the Artemis project to return people to the moon.  Artemis II has half-a-dozen open launch windows over the next several months and NASA is planning some aggressive space flight tests with this next launch to set the stage for a human footprints again on the lunar surface in the next couple of years.  Bob details the plans for the mission while providing needed context to what happened in the late 1960s with the Apollo flights.Post-production note:  Bob mentions in this episode a previous Buf episode in which we discussed his witnessing the Challenger disaster of 1986.  That episode is Episode 19, "Slipping the Surly Bonds of Earth", originally released January 26, 2021; it can be found in the list of previous episodes of the Buf.*****As always, you can reach the Buf at bufnagle@bufnagle.com*****As you know, this is an independent podcast so your hosts also carry all the expenses of running this podcast. As such, some of you have asked how you can help out. Well, here's the answer: support us on Buy Me a Coffee:https://buymeacoffee.com/bufnagleOn this page, you can do a really nice thing like send us a couple dollars to help cover the cost of recording and hosting and microphones and research and all that. Any little bit really helps! Thank you in advance!!!

Description With Artemis II about to launch, Enjoy Stuff looks back at humanity's long obsession with the Moon, from ancient myths to modern sci-fi classics. Jay and Shua explore moon-centric movies, music, comics, and real-life lunar history, proving that whether it's inspiring astronauts or werewolves, there's always been something in the way she moons.   News The Muppet Show arrives on Disney+ as a special event   Check out our TeePublic store for some enjoyable swag and all the latest fashion trends What we're Enjoying Shua has been watching Percy Jackson and the Olympians Season 2 and found it very faithful to the book, even if the pacing felt a bit slow at times. He's still enjoying the journey and is happy to know a third season is already on the way. Jay is diving into Star Trek: Starfleet Academy on Paramount+ and was pleasantly surprised by its quality and tone. Holly Hunter stands out, and Jay appreciates that the show doesn't talk down to younger viewers while still feeling very much like Star Trek.     Sci-Fi Saturdays -  This week on Sci-Fi Saturdays Jay looks at Life (2017), a tense and claustrophobic sci-fi thriller about scientists aboard the ISS discovering a rapidly evolving life form from Mars. It's gripping and well made, but definitely not something you throw on when you're looking for a feel-good movie night.Read his article on RetroZap.com. And make sure to play around with the interactive map on MCULocationScout.com. Plus, you can tune in to SHIELD: Case Files where Jay and Shua talk about great stuff in the MCU.   Enjoy The Moon!  This week, Jay and Shua take a giant leap through the Moon's long history in fiction, film, and real-world exploration. From early writers imagining lunar voyages to Hollywood's evolving obsession with moon bases, secret missions, and alien mysteries, the Moon has always been a perfect sci-fi playground.   They also look at how real lunar exploration influenced pop culture, from the Apollo missions to the upcoming Artemis II launch, and why the Moon continues to inspire stories across science fiction, fantasy, horror, music, and comics. Whether it's astronauts, superheroes, or werewolves, the Moon always finds a way to steal the spotlight.   Are you excited to take steps for a man and woman, and another giant leap for mankind. Let us know! First person that emails me with the subject line, "Houston, no problem" will get a special mention on the show.  Let us know. Come talk to us in the Discord channel or send us an email to EnjoyStuff@RetroZap.com