Subscribe
Share
Share
Embed
TBPN is a live tech talk show hosted by John Coogan and Jordi Hays, streaming weekdays from 11–2 PT on X and YouTube, with full episodes posted to Spotify immediately after airing.
Described by The New York Times as “Silicon Valley’s newest obsession,” TBPN has interviewed Mark Zuckerberg, Sam Altman, Mark Cuban, and Satya Nadella. Diet TBPN delivers the best moments from each episode in under 30 minutes.
It's Model Mayhem, folks. It's Model Mayhem. We got tons and tons of new AI model releases, and it's a great week to be into AI. I think all the lab leaders, they got together. They said, you know what?
Speaker 1:People just love AI. Let's give them more
Speaker 2:Even more.
Speaker 1:Let's all team up to launch new AI for them the same week so that everyone has something just to be happy about. You
Speaker 2:know? That's right.
Speaker 1:That's right. We got Anthropic Fable 5.1. We got New Spark 1.3. We got Gemini 3.8 Flash. Open AI's tees in Astra.
Speaker 1:Is it GPT-six? There's a six where the s goes in one of the videos. People will figure it out. But the model mayhem is continuing. Everyone got back from their long summers, their vacations.
Speaker 1:They said, we gotta we gotta launch
Speaker 2:something new. Time to ship.
Speaker 1:We gotta update this stuff.
Speaker 2:And, Grock, Claude. Yeah. And Oh, yeah. The opening were all down Yeah. This morning.
Speaker 2:Many Astra might have escaped.
Speaker 1:Kind of a deflock moment for AI, maybe? What do you think is going on?
Speaker 2:Possibly.
Speaker 1:Do we actually understand why all of the different models went down at the same time? Because it's easy if it's like AWS went down and it took
Speaker 2:down a bunch east...
Speaker 1:Yeah. But why
Speaker 3:is Gemini... Think it was US East one.
Speaker 1:Why is Gemini down then?
Speaker 2:No. No. Gemini wasn't down.
Speaker 1:Oh, Gemini was never down....
Speaker 2:Whole thing.
Speaker 1:Oh. Okay. Okay. Amazon, clearly very critical to the to the global Internet. Good luck to the folks over at Amazon that are fighting the good fight.
Speaker 2:Keep... Announcements. One.
Speaker 1:Which we got?
Speaker 2:Shermott's birthday.
Speaker 1:It
Speaker 3:is. Big five o. Should we sing full happy birthday?
Speaker 2:I think full happy birthday. Happy birthday to you. Happy birthday dear chumma. Happy birthday to you. Fantastic.
Speaker 2:Almost...
Speaker 1:All in conference is coming to Los Angeles, I believe. Couple weeks. Very exciting. Sign up.... For our invites.
Speaker 2:So, wait for our invites.
Speaker 1:I think we... I think we've been invited to pay and go if you want to go.
Speaker 2:Good to know. Equally important.
Speaker 1:Yes.
Speaker 2:TBPN's road...
Speaker 1:Christmas. How many days...
Speaker 3:Twelve days out.
Speaker 1:Hundred and twelve days out.
Speaker 2:Twelve days out.
Speaker 1:Hundred and twelve
Speaker 2:days. Gotta say, it's feeling like it's going a little slow. Yeah. It it... I...
Speaker 2:I wish there would... I I wish I was feeling more pace.
Speaker 1:Think about it this way. We're only thirteen days away from double digits. That's a big moment. That's a moment everyone's gonna be talking about on the road to Christmas. Yeah.
Speaker 1:When we get to 99 till Christmas, that's Yeah. When you can start a countdown.
Speaker 2:That's
Speaker 1:big. If you get a really big advent calendar, you can basically start that. Let me give you the roundup on the model mayhem that's going on. So, Anthropic launch, Claude Fable 5.1 alongside the restricted Claude Mythos 5.1. Google released Gemini 3.8 Flash and a cybersecurity focused version.
Speaker 1:That's good news. Meta released Muse Spark 1.3. So it might not be that much of a surprise that Anthropic seems to have the strongest model of the three with Fable 5.1 scoring 66 on the artificial intelligence index. That is the bar chart that everyone has been posting in this cycle. It feels like we're sort of maybe getting to the end of the benchmark era.
Speaker 1:It feels like when these models are released, it's much better to solve a novel math problem or do something else. The pelican on the pelican on the bicycle is still still one of my favorites. Yeah. I love that one.
Speaker 2:But... Yeah. It changes everything.
Speaker 1:It does. But the benchmarks, you know, they've been accusations of bench hacking, odd, hard to interpret. I just think many of them...
Speaker 2:Very, very low trust in benchmarks.
Speaker 1:They do.
Speaker 2:And at this point, everyone has had enough experience using various models. Yep. They have their own sort of internal benchmark.
Speaker 1:Yeah. And so the the the demos of like, built this game, I did this thing with it, And then also just the trusted voices of people who, you know, use a bunch of these models and they kinda give you the breakdown of what they like, what they don't. That has been where people lean a lot more. But the artificial intelligence... The artificial analysis intelligence index, this bar chart that you see, has been a good way to kinda compress down a bunch of benchmarks into one meta benchmark.
Speaker 1:So Fable 5.1 get the highest result ever on the index. The score is also ahead of Opus five, which got 63, and Fable, which got a 62. Anthropic says Fable 5.1 is also cheaper and more efficient, made possible by an improved caching system, should make ordinary workloads 25% cheaper and long horizon agentic jobs 45% cheaper, the company says. That's good news. And as Ben Thompson pointed out, Anthropic is also sort of dropping its no zero data retention policy, which there was a whole news cycle around a few weeks ago.
Speaker 1:People were saying, you know, why is Fable not taking off in adoption? It's a really great model. And there were a bunch of different explanations. One of them was companies demand zero data retention. They don't want closed source AI labs to be hoovering up their private information.
Speaker 2:Yeah. And I think they said that they're testing functionality that will allow for...
Speaker 1:Data retention, but it's on servers and infrastructure that the company owns. Yeah. So you do keep some of the data. You still are monitored for hostile usage, but it's not going straight into Anthropix databases. So Alex Karp and Satya Nadella both warned against this idea that companies should have data sovereignty.
Speaker 1:So the policy is going to be replaced. The no zero data retention, no ZDR is going to be replaced with something called EFS, enterprise frontier safeguards, and that may have attributed... May may have contributed to lower Fable adoption among enterprises, and it sounds like it was direct response to user feedback. So good news that, you know, the people spoke and the companies listened. So over in Google World, Gemini 3.8 Flash is the company's third Flash release in six weeks.
Speaker 1:They are flashing out these Flash releases and scored 73.7% on DeepSUI, just behind Opus five and competitive with models that cost several times more. Independent testing gave it a 59 intelligence score, which isn't the absolute frontier, but it's a great result for a model generating roughly 300 tokens per second. So very quick, very cheap, and very good at coding, at least on this particular benchmark, DeepSui. We'll see what adoption looks like and and where enterprise spend goes. Ara Karazian has some very interesting data from the Ramp Economics Lab.
Speaker 1:You can go check out. He also has a new post that's very interesting that we can talk about in a second. But last model, MetaMuseSpark 1.3. Did very well in benchmarks, scoring 75.4% higher than Gemini 3.8 Flash on DeepSweet, beating both Opus five and GPT 5.6 SOL. It didn't sweep the board.
Speaker 1:Opus still beats it on several professional work and computer use evaluations, but it got six to two on the intelligence index, which is only behind the newest Claude models. Tons of stuff to think about and discuss here. So the interesting post from our Aracrazian, and I don't know if we have it in the timeline, if we can pull it up, but he was saying that there's a lot of concentration in the enterprise AI revenues right now. OpenAI and Anthropic, 80% of their enterprise revenue comes from just 1% of the companies. And I was like, 1%?
Speaker 1:That seems crazy. And and and he notes that this is uncommon for software categories. Like, if you look at CRM, if you look at databases, if you look at all sorts of different software spend, typically, you don't see as much concentration. You don't see 1% driving 80% of the spend. And I was wondering about this, and so I started looking up like what where else do we see this type of inequality, if you can call it that, this distribution, this power law?
Speaker 1:Power laws are everywhere, but where else does this exist? And you might go to hiring. Like, is AI a drop in replacement for hiring? Is it gonna be proportional to hiring? And, in fact, 1%...
Speaker 1:The top 1% of biggest companies in America, they do hire a ton of people. The top 1% of American businesses employ 65 of the total workspace... Work workforce. Not 80%. But interestingly...
Speaker 2:Concentration risk there, John. 65% of jobs Yeah. Are tied to just 1% of companies.
Speaker 1:There is. Yeah. There is. And and and, I mean, yeah, you you definitely see that. Although the top 1% companies tend to be pretty lindy, you're talking about
Speaker 2:No, no, I know.
Speaker 1:I'm joking. Yeah, government and whatnot. The interesting 1%, 80% correlation comes from sales. So the top 1% of American companies by sales generate 80% of total revenue. And so there's this weird dynamic where I don't know exactly how correlated it is, how causal it is, but there is an interesting dynamic there where it feels like if you look at the total AI spend, it's around $150,000,000,000 a year or something like that.
Speaker 1:And then you look at total revenue for all U. S. Businesses, AI is roughly a quarter of a percent of total U. S. Business revenue.
Speaker 1:And it tracks fairly closely to the revenues of those individual firms. So you see that that 1% of the top businesses generate 80 percent of the revenue. They also spend 80%. They also generate 80% of the AI revenue. And so there's this interesting dynamic where, because enterprise AI particularly, you're not going to be on the $20 plan, you're not going be on the $200 plan, you're going to be consumption based, and you're going to look at it a lot more like a marketing line item that's proportional to your revenue, potentially.
Speaker 1:That's at least one interpretation of this. Another fellow over at Ramp said that this is a Rorschach test for how you feel about AI. Either you look at this and you're like, it's great. Or you look at this like, it's over. But fun fun fun fun chart to dig into.
Speaker 1:Anything else on this? You guys? Reading this? Nah. Let me tell you about...
Speaker 2:I just was realizing
Speaker 1:Codex is a powerful workspace for getting work done with AI agents. Whether you're writing code, analyzing data, creating content, or automating business workflows, Codex helps you move projects forward from start to finish.
Speaker 2:Yeah. We forgot to cover. Pablo Torre.
Speaker 1:Yeah. Pablo Torre.
Speaker 2:At 11:30. 11:30. To talk about the Clippers. Yeah. And Kawhi Leonard.
Speaker 1:So these are people that go, they watch live streams and they clip them and they put them
Speaker 2:out on
Speaker 1:social media?
Speaker 2:Yeah. Why it's... I think that's what the team... Why they named the team that.
Speaker 1:Yeah. Yeah. Kind of an homage. Because there's a lot of clipping that happens in LA. TikTok clips, Instagram clips, they call them...
Speaker 2:This whole Balmer Kawhi Leonard thing. Yeah. You had an interesting pronunciation of Kawhi's name earlier because I don't think you'd ever heard of
Speaker 1:him. Well, was calling him Steve Baumet. I I I dropped the r because I thought it was a friend. Oh, nice. No?
Speaker 1:Yeah. So, the other big story in the news, the open source community is stronger than ever. You saw, three, four, who knows, five closed source releases. All the all the big labs are duking it out. Meanwhile, Nvidia is going even bigger on open source with the $13,000,000,000 acquisition for Hugging Face.
Speaker 1:All over the timeline. This was leaked a couple weeks ago, I feel like, rumored. What are you laughing at?
Speaker 2:No. Just every single day, I would see a headline about Nvidia hugging face Yeah. And think, okay. Now it's official.
Speaker 1:Yeah. No. It's official. Today is the day they can talk about it. They can explain it.
Speaker 1:And there's a lot that makes sense. There's not too many questions. It's just a great outcome generally. But it's an interesting story because it's a true ten year overnight success. And and they're having fun with the acquisition price.
Speaker 1:Nvidia agreed to pay $12,930,300,000 for Hugging Face, which just happens to be the exact decimal code for the Hugging Face emoji, which, of course, is the icon used by the company. And then also, you take that number and you turn it into a color code, I think you get a green that sort of hints at Nvidia. So they're having fun both ways. Like, yeah, this was always in the plan symbolism. It's just funny to be having fun with a price this big.
Speaker 1:I remember the Instagram acquisition and the idea of a billion dollar outcome as being insane during the social media boom. And now, we're seeing like, you decacorn liquidity events every couple weeks. Yeah. I'm of course thinking of OpenRouter
Speaker 2:and... It's honestly an incredible time to be investing in AI seven years ago.
Speaker 1:Yes. Yes. Best data plan
Speaker 2:of... Three to seven years ago was an amazing time to be investing in AI.
Speaker 1:It was. And so obviously, this is... This shouldn't come as that much of a surprise because Jensen has been the... Probably the loudest voice on open source. He put out that open letter that everyone signed on to.
Speaker 1:And he wants to maintain Nvidia's dominant position in the ecosystem, both selling chips to closed source labs, who might wind up making their own chips as well. And there's a whole tug of war there. But for open source AI development, he wants to be the place where developers companies go to pick their models and then hopefully rack them on Jeep... On Nvidia GPUs. The simple distillation is just Hugging Face is the GitHub of AI.
Speaker 1:It's a little more complicated than than the Microsoft GitHub deal, but it still makes a lot of sense. So Hugging Face doesn't own the smartest models. They don't even try to build them, and there's some interesting financial dynamics there about how capital efficient they were because of that decision. But they created this nexus for people to upload, discover, test, modify models. And the numbers are good.
Speaker 1:They have over 18,000,000 developers, 200,000 companies using the product, 3,000,000 models, and over half a million data sets. And so they have certainly created this vortex of activity that's really valuable. How strong is the network effect? It's certainly cooking and it's certainly driving a lot of value here. So the interesting thing about Hugging Face is that it did not start as an AI GitHub for AI.
Speaker 1:It started as a completely different idea. The founder worked at a French computer vision startup called Moodstocks that was eventually acquired by Google. And in 2016, he teamed up with two cofounders, one who was a mathematician and the other one who was a scientist who had worked in patent law, apparently. And they started building an AI that could basically talk about everything. So this was post Siri, post Alexa.
Speaker 1:But instead of focusing on like, tell me the weather and be as... Be an assistant, know, set a timer, he wanted to just to be able to talk to you.
Speaker 2:Solved by the way.
Speaker 1:Wait. Which one? Siri. Yeah. Yeah.
Speaker 1:Yeah. So, yes. It's pretty good at setting timers. But, so the goal was to build something like a Tamagotchi, something very cute, hence the Hugging Face icon. A funny, emotional, digital friend targeted at teenagers.
Speaker 1:The app let their name... Let users name the bot, text it, send selfies, trade emojis. It was explicitly marketed as an AI best friend for bored teenagers. And they scaled it. I think this is pretty significant.
Speaker 1:It was doing a million messages a day. They had more than a 100,000,000 messages in total by 2018. That seems pretty significant. That doesn't seem like, you know, you're languishing in the App Store with no downloads because how many messages a day can a bored teenager possibly put up with a... With an AI agent?
Speaker 1:Even if it's like a thousand, you still have, I guess, a thousand users, power users. I don't know. Probably the average user is doing 20 messages a day. So you're you're you're seeing pretty significant adoption. And so they were able to raise a series of financing rounds.
Speaker 1:The big one that grabbed headlines was Kevin Durant. Durant for three. Go. 2018. This is pre GPT-three, and it didn't become a durable consumer business.
Speaker 1:So in 2018, Google released BERT, which was sort of the first language model, very primitive, but people were really excited about it. But it wasn't delivered just as weights that you could download on the Internet. It was delivered as a paper from Google. And the paper was implemented in Google's TensorFlow framework. And people liked PyTorch.
Speaker 1:So the Hugging Face team converted BERT from TensorFlow to PyTorch and released the conversion for free. And so developers really liked that and that became sort of like the initial go to market flywheel for developer adoption. And eventually, they added more and more models, eventually thousands. Now, I think they have millions of models, which is sort of crazy, but when you think about all the forks and fine tunes, it makes sense. Eventually, the team stopped trying to build this one application and focused on building tools, became the picks and shovels trade.
Speaker 1:So instead of trying to pick a winner, you just host every model. They became the Switzerland of AI to some degree. And so the flywheel started compounding. More models, more developers, more model creators, more companies. And it was the same basic network effect as GitHub.
Speaker 1:The, you know, GitHub was the default home for open source software. Hugging Face very quickly became the default home for AI models. Over time, Hugging Face grew from a code library to a place where developers could publish models, version them, attach data sets, discuss changes, and they even allowed them to build demos. Hugging Face eventually launched a Spaces product where you could demo these... The different models.
Speaker 1:Companies can maintain private repositories, same GitHub strategy. So 2019, Lux comes in with $15,000,000. Woah. Then, yeah, Lux got in early. Series 15 mil.
Speaker 1:They also came back for the series c in 2022. That was a 100,000,000 at a $2,000,000,000 valuation. Sequoia and Cotu were in that round. There was also a series b in 2021. That was 40 mil.
Speaker 1:And then the big step up was in August 2023. Hugging Face raised 235,000,000 at a $4,500,000,000 valuation. And it's a murderer's row of potential acquirers. You got Salesforce, Google, Amazon, Nvidia, AMD, Intel, Qualcomm, and IBM. So you're...
Speaker 1:You know, it's not like they were doing a roadshow to sell the company, but it's very much like we want to be the Switzerland of AI. We want good partnerships with everything. We're gonna be chip agnostic. So, yes, we have Nvidia on our cap table, we also have AMD and Intel and Qualcomm. So, you know, you you can count on Hugging Face as being like an independent place.
Speaker 1:We're not we're not purely Nvidia backed, which is maybe one of the things they'll have to deal with now, but they're not purely Nvidia backed at that time. So it's very much like, oh, yeah, we'll host a model that runs well on AMD. We'll host a model that run... Runs well on Nvidia. We'll host models that are from Google or from Amazon, etcetera.
Speaker 1:And so it looked like this, like, peace treaty moment from the major AI infrastructure companies. You get everyone around the table. Everyone's aligned with the mission. And Hugging Face becomes this neutral territory where it supported competing clouds, chips, frameworks, models no simple no single company can control the platform. And so even though they did a number of rounds, Hugging Face, I'm going say only raised under $400,000,000 which is a lot of money, but not at a $12,000,000,000 outcome.
Speaker 1:And it's pretty small considering the outcome. And it was very capital efficient because they weren't actually buying chips or serving models directly. And they had this flywheel that sort of spurred growth through the network effect naturally. So not a lot of cost in the business. They became profitable in 2025, still had half the money that they raised.
Speaker 1:So Nvidia came in to offer 500,000,000 late twenty twenty five at a 7,000,000,000 valuation, but they turnt it down. Woah. Turnt it down. Because they said, hey, if we're gonna go deeper with one particular area, it's gotta be the whole shebang. And so that's what what wound up happening.
Speaker 1:So, pretty high revenue multiple...
Speaker 2:My question my question is, I wonder what Jensen's vision for Hugging Face is. They they do offer model routing and and, you know, they they rank a bunch of inference providers. Yeah. Is this something that... Could we see Hugging Face and OpenRouter and RampsRouter competing more and more?
Speaker 2:Yeah.
Speaker 1:Sure. I think it's two things. I think I think one is closed source is already its own business line, sell chips to the labs, but the labs are building ASICs. They're doing a lot of stuff, and there's this whole back and forth tug of war there. But on the flip side, you have open source, which is continuing to grow.
Speaker 1:And if you can be sort of the front door to that and then say, hey, you found your best model, your framework on Hugging Face, and now you're ready to go and buy chips or buy inference or buy compute, and Nvidia is right there. That is a very logical flow. And anything that they can do to make open source powerful, exciting, a place where where you can build a career, have a great outcome. I think that's beneficial to Nvidia because a lot of people will see this and say, yeah. Like, you can go and build a great company and have a fantastic outcome.
Speaker 1:I mean, the retention packages are apparently a billion dollars for a pretty small team. And so if you're... If if Nvidia is just trying to send a massive signal to the world that you can make it in the open source world, that's a really good signal to send. It feels like it's landing loud and clear, especially today. What else is on your mind regarding Hugging Face and Nvidia?
Speaker 2:Are still waiting --... Benchmarks? For an official launch
Speaker 1:What time?
Speaker 2:From of of Astra. Lisan Al Gayebe is sharing some Astra benchmarks. ARC AGI three ninety eight point six. He had, they had previously said, are you ready for a nuke to hit ARC AGI three? So almost fully saturated frontier math, tier four v two gets a 97.6.
Speaker 2:Deep SWEE 74.1. X Plate Bench also saturated at a 100%. So Wow. Seems pretty good.
Speaker 1:Let's go through what else is in the timeline. What is John Palmer saying these days? He says, I actually think Snapchat for work might be a good idea in today's big companies. Work platforms, work communication platforms have always followed what teenager were doing ten years ago. I did use IRC when I was a teenager.
Speaker 1:Unc style. Tyler has to look it up. He doesn't know what What Internet relay chat is. Wow. Wrong.
Speaker 1:Wrong.
Speaker 2:For what it's worth, I didn't use IRC either.
Speaker 1:No. What... Did you use AOL? You didn't use AIM? AOL instant messenger?
Speaker 1:Wow. What what was your first communication platform on the Internet?
Speaker 2:Email. Email.... Being inundated with emails. I remember there was like a summer there was a summer...
Speaker 1:He's laughing at me.
Speaker 2:Wow. I I just Rude. I remember spending a summer as like Yeah. Not even a teenager yet, just being super stressed about my inbox. Yeah.
Speaker 2:Because like every kid had just started using email.
Speaker 1:Yeah.
Speaker 2:And so they were just sending these super long emails I would and I would be thinking... I'd just be like playing outside in the grass and thinking, man, I gotta I gotta check my email. Yeah.
Speaker 1:Do you think there's anything actually to this? It sounds like he's being serious. In the age of AI slop, the most efficient form of communication is just short videos of yourself speaking. What do you think?
Speaker 3:I would love to use Snapchat for work.
Speaker 1:So if we if we said, hey,
Speaker 2:as a
Speaker 1:as a team, we're going to communicate through short selfie videos.
Speaker 3:Yeah. With the filters?
Speaker 1:With the... You gotta have the dog filter or whatever, the hot dog filter, whatever those were. He says he's serious. A two minute demo video. Yeah.
Speaker 1:I guess I guess in terms of actually just taking a video of your screen, showing people what you're working on, you know, there's a lot of different things you can do. It's official. Hot bot summer is over. Super Grock is saying goodbye to companions. I remember we were debating...
Speaker 1:You know, there's obviously a lot of people that had ethical concerns about about AI romantic companions, but we were more discussing just like, is there actually a business here? Once you break the seal of like, okay, I'm... We're we're doing it. He did it. Would it actually be successful?
Speaker 1:Because Replica has seemed to get to scale. There's been other products that have played in this world and and seemingly reached adoption and scale. But this is the same thing as the Sora discourse, where everyone was caught up in Sora is is either going to be the most powerful thing ever and we're not gonna be able to stop watching it or it's gonna be good and so fun and and it's gonna be amazing and dominant. But no one was counting just like, oh, it might just go away in six months. And it's the same thing here.
Speaker 2:I don't know. I Internet... I think I was saying it was Gonna go away. Away. Yeah.
Speaker 2:Just because I thought it was gonna be a tool.
Speaker 1:Yeah. And I think I think we benchmarked the market for this, like, if you if you look at other romantic stuff... Said,
Speaker 2:okay, I'm going all in on on adult entertainment. I did think it was a potential path for for Grock to get into the single digit billions. I probably overestimated Mhmm. Overestimated the market there. But it felt like one of the plays that he had to get back in the game at the time.
Speaker 2:But again, this was last year. And back then, if you could get into the single digit billions, you were doing pretty well. Yeah. And then the game obviously, you know Yeah. Really shifted down.
Speaker 1:Yeah. It makes sense to wind it down, focus on enterprise software. That was what was in the SpaceX s one was the the... What what was it? $13,000,000,000,000 market he was going after?
Speaker 1:Something like that? It was in the... It it was a shocking, shocking number. Maybe 20,000,000,000,000 or something. Absolutely huge, huge numbers, and it makes sense.
Speaker 1:There's a lot of value in the enterprise, much less in this controversial topic. Well, fish dot audio ran a billboard campaign. We love out of home. This one sort of confused people. Hashi Brody says $100 for anyone that can explain what this company does without looking it up.
Speaker 1:This is on the New York City subway today. And we'll read it to you, and you can take a guess. Fish dot audio says, we put voice AI on a silent sign. You see the problem. Is this one of those jokes though?
Speaker 1:Because...
Speaker 2:Think it's a joke. I think this is an Eleven Labs competitor.
Speaker 1:We even know. But but is this a real ad from a real company? Or is this a prankster making fun of tech ads? Think it's you
Speaker 2:think. Really makes you think. It made me go to the website.
Speaker 1:Okay.
Speaker 2:Text to speech, speech to text, audio separation, voice changers, translation.
Speaker 1:Mhmm.
Speaker 2:And they partner with global innovators, John. They're working with Okay. Retail.
Speaker 1:Real company.
Speaker 2:Yeah. A bunch of games companies, it looks like. Clout Kitchen. You ever heard of Clout Kitchen, John? No.
Speaker 2:You ever been in the kitchen cooking clout?
Speaker 3:Okay. So isn't it like they put voice AI on a silent sign?
Speaker 1:Mhmm.
Speaker 3:You can't like, if you have a billboard, you can't listen to it. But their product is audio.
Speaker 1:Oh, so if you get...... The billboard. If you get...
Speaker 3:That's the problem.
Speaker 1:Yeah. Right? I guess I would expect...
Speaker 2:Your explanation is a problem.
Speaker 1:GPT-six Astra has landed. The blog post is up. You can go check it out on openai.com.
Speaker 2:To the stars.
Speaker 1:To the stars. And the headline, I believe, good performance on a bunch of things. But the RKGI three number is crazy. The score is 99.9. So it feels like they just beat that.
Speaker 1:They just beat RKGI v three. So that's the video games, the ones that Tyler was briefly
Speaker 3:Globally ranked.
Speaker 1:Globally ranked. You're out of a job, Tyler. Say say goodbye. Hang up your Arc AGI v three hat. Don't worry.
Speaker 1:The team over at Arc AGI is working on v four. They're... We're gonna move the bat... We're gonna move the goal post soon. We're gonna move the goalpost soon.
Speaker 1:But this is very impressive if you've played around with Arc AGI v three. It it it's... It it requires, yeah, like, real creative thinking to actually learn how the games work, how to be efficient. What else is sticking out? Has anyone been able to to monitor the timeline at all?
Speaker 2:Very chaotic launch, so
Speaker 1:Okay.
Speaker 2:I I think it's hard to get a real reaction yet. Tyler, what are you seeing?
Speaker 3:Yeah. I mean, there's still no actual post on OpenAI's, like, x account. It's just a blog post that's
Speaker 2:now live.
Speaker 1:Okay.
Speaker 3:But, I mean, there's a bunch of benchmarks in there that are
Speaker 1:Yeah. All Terminal bench science. We're going over into the world of science. GPT-six Astra scored 64.6% on reasoning effort max, cost $26.26 bucks. I'm sure there will be a lot more.
Speaker 1:Also, the, exploit exploit Jim Honey Pot, lower is the better. GPT-six Astra, 0%. So good performance there. The world's best computer use model, I'm putting that to the test ASAP. Let's see if it can one v one me on Rust.
Speaker 1:Because if it's good at using the computer, it should be able to no scope. Right? That's the bar. That's where my goal posts are. Agents last exam, good performance.
Speaker 1:So you can go check it all out....
Speaker 2:Gonna be on Bloomberg Very fun. Just a few minutes over with our with our buddy at Ludlow.
Speaker 1:Very fun.
Speaker 2:It'll be fun.
Speaker 1:So we'll be digging into this and we have some special guests lined up to talk more once we've been able to digest, take it for a spin, and have a lot of fun with it.
Speaker 2:Mike over at ARC is Moving the goalpost. Moving the
Speaker 1:Moving the goalpost.
Speaker 2:Yep. They're moving. Thank you.
Speaker 1:Let's do it.
Speaker 2:Welcome to waiting next to move the goalpost either.
Speaker 1:What did he actually say?
Speaker 2:He said Astra is is the new state of the art on art AGI three. It's a qualitatively large leap towards AGI and the pace of progress is frankly surprising. That said, we lack evidence to call this AGI yet. AGI. Is is AGI?
Speaker 2:While we are still studying the human capability gaps, we believe open ended invention is unsolved and this will form the new basis for ARC AGI four. So now it's like Yep. You gotta invent new physics, new science. Yeah. You gotta go to that.
Speaker 2:You actually... Astra has to actually go to the stars.
Speaker 1:Yeah. He really said, what have you done for me lately? I love it. I love it. It's it's AGI when Mike says it's AGI.
Speaker 1:It's a good time. Anyway, very interesting. You can dig into his post. He gives a lot more context there about Arc AGI v three, benchmarks, Astra. You can go check it all out.
Speaker 1:And of course Yeah.
Speaker 2:We'll do.
Speaker 1:Discussing it tomorrow. We have a bunch of special guests. So have a great day. We'll see you tomorrow. Leave us five stars on Apple Podcast.
Speaker 1:It's got a nice time for newsletters at tbpn.com. And we will see you tomorrow.