Limitless Podcast - THIS WEEK IN AI: Tension at World AI Summit, Snapchat Glasses, Meta's Crisis

Episode Date: June 19, 2026

We discuss Snap’s new AR glasses, AI governance at the G7, and Noam Shazeer’s move from Google to OpenAI. We also cover Anthropic’s Fable 5 being shut down, Microsoft’s use of a Chinese open-source model in Copilot, and Meta’s latest AI push.------🌌 LIMITLESS HQ ⬇️NEWSLETTER:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------TIMESTAMPS0:00 Ridiculous AI Glasses7:27 AI Summit Tensions11:46 Noam Shazir Joins OpenAI15:48 Anthropic’s Fable Shutdown18:52 Microsoft Embraces Open Source24:36 Meta’s AI Reorg Crisis29:15 China’s New Coding Model32:06 MidJourney Enters Medicine------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

You either die a hero or live long enough to see yourself become a villain. Well, today, there's a new version of that, which is you either exit Silicon Valley or spend enough time there for no one to tell you that these glasses look absolutely ridiculous. What we're showing on screen right now is the Snap Spex. This is the new version of Snapchat's hardware offering that is marketed as this spectacle of a device. It is this remarkably intriguing and compelling and innovative device that sits on your

face, that has all these capabilities. but the reality is that, I mean, dude, these glasses on Evan Spiegel's face, the CEO of Snapchat, not only do they look ridiculous, but they cost $2,200. I mean, what was the thinking? What was the company thinking? Who let this get released? And how are they so confident as he's walking around the venue yesterday as they released these specs? Listen, this is a serious show, and I want to have a serious conversation about this,

but this is ridiculous. Do you know when you buy sunglasses for a little child? and like, or they're like, Mommy, can I, like, wear your glasses, please? And it just looks oversized and ridiculous on their head. This is what this looks like. I just want to point out, for those of you who are looking at our stream right now, the back half of the actual glasses, the handles, barely kind of broach his ear.

They kind of, like, sit on top of his ear. That's got to be so uncomfortable. Wait, and for those of you who say, oh, this is just an AI rendering, no, look at it. This is live, like in the mix of things. As he turns, you can see, they kind of extend beyond the back of his head so that it's like oversized fit for him. Now, that dude's ears are fighting for their lives. They're fighting for their life.

So this is Snapchat, who hasn't really done anything major for the last 12 years. In fact, for the last 12 years, they've been investing a heck ton of money to the tune of $3.5 billion to figure out the AR glasses game. You know, you had meta that's released a bunch of models since then. You've got Apple that's released the Apple Vision 4 for $3,500. And finally, Snap has released their flagship model. You would think after all that R&D, they'd come out with something cheap, slick, and aesthetic. And unfortunately, it's been the complete opposite. This is $2,200 to achieve or get these specs that you're seeing on your screen now.

It's a standalone device. It has around 51 degrees field of vision, which is similar to Mehta's Raybank glasses. But of course, this thing is like 3x larger than anything that we've ever seen. You don't need to connect it to your phone. And basically, it runs its own software operating system. on the lens itself. Now, I have seen demos of what the software itself looks like, Josh. And believe it or not, it's not entirely the worst thing. You're seeing someone on the screen here where someone's playing golf. You're seeing someone basically ordering off a menu in a language that they don't

understand. You're seeing someone taking directions. You're seeing someone watching a movie on a plane. You're seeing someone design the architecture of a building. This all looks good, but these are all curated demos, just to be clear. I don't know whether it actually looks this good, but these demos are already better than meta rayband glasses. But no one's going to buy these things and it's such a shame because it looks terrible. You don't want to look like you're wearing old school 3D glasses watching a movie that no one's even there to see. It's just, it's so dumb. Well, forgive me for my PTSD, but I think I'm going to need to see it to believe it. These seem like a lot of kind of structured demos that are not out in the wild. I give meta a lot of credit

when they released their glasses. Their meta rayband displays, I think they were called. I forget because no one actually bought them. They did do some demos on stage, and they just completely flopped, and they failed. So I give them credit for trying. Snapchat, I don't think did that. This is mostly a curated view of demos. The form factor, well said, there's some nice things to say about this thing. Like, the fact that they're trying for one.

I give them a lot of credit for actually trying for throwing their hat in the ring. It looks terrible, but the technology is on its way. You mentioned the 51 degree field of view. That actually is an improvement. what that means is it's about equivalent to a 24-inch monitor up close. So if you have a 24-inch display that you use your computer on, that's roughly what it looks like. Or if you step back to 10 feet, it's equivalent to about 115-inch TV with seemingly pretty good fidelity. I was listening to one of the interviews with Heaven Spiegel, the CEO, and the interviewer asked him, the very obvious

questions is like, how are you planning to justify this? It's $2,200 for this, like, very clunky pair of glasses. And the way that he justified it is like, Hey, people spend that money for a computer all the time. In fact, they spend a lot more money on that for a computer. And what you're buying with these is essentially a computer. This is a standalone display. It does have two Qualcomm chips built into it. It has its own vision suite of sensors. It has two full-color cameras. It has two infrared cameras. It has two infrared cameras. It has a microphone array. It has hand tracking, voice control. A lot of the features that you expect inside of a computer, it's on these pair of glasses. Now, the pair of glasses, the battery life is decent. It's four hours.

of mixed use, and with the charging case that it comes with, you get about 20 hours. The old spectacles managed about 45 minutes. This is a pretty considerable jump over previous versions. It's just like the Vision Pro, there's just not really a market for this. I can't imagine there's going to be a world in which, like, I love this type of technology. I am version one beta buyer of all the hardware, but even this is like, I don't want to wear these things. They're too bulky. They're too clunky. I give them a lot of credit for trying to do this for innovating. And it's certainly correct, like directionally at least, where the hardware of glasses as a form factor is great. It's just the actual hardware required to make them a great

product. It doesn't seem to be there, at least in the case of Snap for now. But it's also like the worst time to release a premium product. Like it's not cheap. This is a high-end laptop at $2,200. At that level, you are competing with Apple Vision Pro, which hasn't taken off either, right? Apple, yesterday announced that they're going to increase the cost of all their products because the cost of memory and everything is going up. So it's the worst time to, when people are kind of like pinching their wallets and not really spending much because, you know, their premium gadgets that they want to buy from Apple, they can't purchase themselves. What makes Snap think that, you know, they're going to be the replacement there? Now, one, like, the points you make around the software thing and that they're trying, I agree, is good. The one thing that I saw when I was looking into the dev kit for this particular device is,

you can hook it up to Claude or chatGBT, which I think is pretty cool for creating new kinds of AI apps. Typically, with meta, they restrict it to meta AI specifically. I know with Google glasses that is coming up, I believe it's still tied to Gemini. The fact that they've kind of like open source this to kind of like create and run your own agent could be pretty cool. It's like the closest thing that maybe we've come to Jarvis from Ironman, except this is like discount store Jarvis. This is like we've got Java at home, you know, like you can't buy the real Jarvis. I'm hoping that this is the worst it's ever going to be. I like the attempt on goal that they have for creating the software suite around this.

And I also like that it's a standalone product that I don't need to hook up to my phone. It kind of like runs on its own. But it remains to be seen. Like I want to see someone in the wild that isn't from Snapchat actually trailing this thing. But to spend $3.5 billion, they spent $1.9 of that, by the way, in 2023 alone. And it's taken them like, what, three years to launch this thing? it remains to be seen, but I think we should probably move on to the next topic. There was a very important meeting amongst world leaders in Evian, France, which is where

the famous Evian Water comes from, by the way. It's the G7 summit where you had the top seven nations of the world come together to discuss an array of political affairs. One of the main ones, however, was around AI and governing and regulating AI. And we had the main heads of AI were there, including Demis Sissivis from Google, Dario Modi from Anthropic, and Sam Altman from Open AI. Now, the main call, the main plea from Darius and Sam was for a US-led coalition to determine the global standards and rules for AI in a closed-door meeting that happened in the morning. So it was kind of like a lunch thing that happened that the G7 Trump was present, as well as all the leaders of Europe and a bunch of other countries. And the main idea was, let's start a conversation

around how we regulate this thing, because, as we know, from recent news, the models have finally gotten good enough where they pose themselves as a national threat if it's in the hands of an adversary. And so that was what that discussion ensued. Yeah, this is hopefully a productive meeting. I mean, the working theme of this lunch was ensuring a safe, rapid, and effective deployment of AI. And I think this is effectively what everybody wants. There's just a lot of varying opinions on how you actually get there. And I am hopeful that the intention of this meeting was to find some sort of convergence to find some sort of middle ground on how they could consider moving forward with these frontier models. We very clearly see now that we are experiencing this takeoff scenario

in which the models are becoming better and better. They are improving themselves. The accelerated rate at which we are getting high-quality models is like a little staggering. It's like very fast. It's overwhelming. And I think a lot of times there is this clash between the government, which traditionally moves a little bit slower, and these AI labs who are like, hey guys, this is not going away. In fact, this is only speeding up. And finding some sort of middle ground is productive so that we don't have any more issues like Fable. I mean, not being able to use Fable or being able to use it for three days than having it removed, it sucks. It hurts the users. It hurts the company. It seems like it mostly hurts everyone. So hopefully the outcome of this will be a little

bit more of a formalized process or at least a mutual understanding of where things lie. So when companies release the model, they feel very confident in the model that they're releasing. They feel confident that the government supports them and that there is not going to be a situation in which something gets released. That is a frontier model and then gets pulled back. And I mean, fingers crossed it go as well. Everyone who should be in that room seems like they are in that room. And hopefully that leads to a pretty productive conversation. Well, being in the room is one thing. Where they were sat was another thing. I don't know if you saw the seating plan for this thing. So I was watching clips of this actual discussion.

because it was filmed. And I noticed one thing. So if you look on the right side of the image here, on either side of Trump, you had Demis Hesovis, head of Google, Deep Minds, so Google's AI division. And then you had Sam Altman to the right of him. And then that guy, Darya Modi,

yep, seated exactly opposite. So nowhere near Trump. And so the tension was pretty obvious. The kind of rivalry, or rather the grudge that was going on between the government right now and anthropic, as they're discussing kind of like re-release a fable and how they kind of regulate and govern that is obviously a big tension between the two parties. When Dario spoke up and he was the last to speak of the three,

he spoke about warning of regulation and saying we need to over-regulate this kind of things and we cannot get these models in the hands of China, which is nice to hear because that's, I think, him and the government kind of a line on that. The second main discussion that was had between the three of them was the idea of making sure we govern hardware as well as software. Now, typically the US has taken a very proactive approach to making sure that China and adversaries don't get access to cutting-edge chips, primarily from Nvidia,

but that has also pushed China to kind of like create their own chips, which have now almost caught up with the US in terms of like inference and training. Dario has been warning like, hey, we need to restrict the hardware even more, and we need to be careful with the safeguards that we set up for our models, which is ironic, of course, because that has led to the ban of fable so far. Sam was the first to speak up. So he was given, if you're talking about the Trump's favorites or the G7 favorites, it was Sam, it was Demis, and then it was Darien.

So a lot of political things going on. Okay, there's another big update that I want to talk about. And this one feels pretty noeworthy, and it's around this man by the name of Noam Shazir. And Noam is interesting because he is very much a legend in the AI space. For those who are familiar with the word GPT, perhaps you've heard it in the context of chat GPT, the letter T stands for Transformer. And Transformer is an engineering concept that was actually pioneered in part by Noam Shazir, who wrote the canonical paper, attention is all you need, about how a transformer actually works all the way back in 2017,

which in AI years, that is an eternity ago. So Noam is this foundational character in the world of AI who has really been there through all of the changes, all the evolutions of it. And he's had this kind of rough, bumpy road with Google. he's been working at Google, or he joined Google initially in 2000, I think. That's 26 years ago he's been there. And for a long time, he was an engineer, he was building these AI models, they figured out the transformer architecture, and then he left. And he left to go start his own thing. He went off for a couple of years. And then in 2024, Google missed him so much, they reportedly paid

$2.7 billion to license his company, character AI's technology, to bring Noam Shazir back, making him a VP and one of the people in the small group that are steering how Gemini gets towards AI. So from 2024 to 26, he's been working there after a pretty hefty payday. Well, yesterday we got the news that Noam is leaving the company. And not only is he leaving, he's going to open AI. So that's the backstory. And this seems like a pretty big deal. Because when you think about the Avengers of AI, who are the people who are really making a

difference, Noam is one of those guys. And now he has changed houses, I guess, in this game of throne. from the Google House to the House of OpenAI. And he's going to be leading the architecture team at OpenAI, which is arguably the priority team that will design the model, future iterations of the model, and may need determine who leads in the race. If you want to think of who his counterpart is at Anthropic,

it's Andre Carpathy, who also recently joined Anthropic. So if you want to kind of think of the titans of the game, the leaders, the geniuses, it's going to be Noam Shazia now versus Andre Carpathy. And Noam, as you mentioned, father of the transformer, I can't believe he left. When he left Google initially, when you mentioned, I think he was there for like 26, and then he left in 2021, he started character AI, Josh, the AI that like kind of like creates

kind of like different personable characters, the original AI agent, if you want. And then getting acquired for $3 billion was just massively impressive. It's also a huge knock on Google. I don't know how I can view this, aside from being bearish on Google. Everyone knows that for the last year, they've struggled to keep Gemini models up with the frontier. They had a massive lead at the end of last year when they released Gemini 3. Recently, 3.5 and 3.5 Flash has kind of caught up in some kind of generalized LLM qualities, but they've missed out on the major quality, which is coding.

It just hasn't been able to catch up. We know that a lot of Google employees have actually been using called code and codex itself to try and distill what the model is, but they haven't been able to achieve that. So Gemini has struggled and the fact that their chief architect, he's the VP of engineering. He is basically Demis de Sivis's right man. Demis de Sivis is like the front man, the CEO of Google Deep Mine. And this guy is the actual technical guy. The fact that he's left to build architecture at one of the biggest competitors ever tells me that he was either being restrained at Google

or that he wasn't able to kind of build a thing that he wanted to. And now he will be. Sam in his announcement tweet for the hire basically said, I've been trying to work with Noam for 10 years and it's finally happened. And I can't wait. So, yeah, let's see. Congratulations, Sam. Yeah, we'll see what happens. It feels, I can't help but notice that it feels like all of the top talent is starting to aggregate

itself between only two houses. We had Andre Carpathy recently joined Anthropic. Now we have Noam joining Open AI. It feels like a lot of the top talent is kind of converging on these two places. It seems like these two places are kind of running away with it. Like we have Google Gemini. And every time we try to say, like, this is a new frontier model. Every time they have a big release, it doesn't quite do it.

GROC and the XAI team and I mean, SpaceX, XAS, space, the whole, that universe, the Elon universe, they're trying. They haven't quite figured out how to get to the frontier. And right now, there's very clearly between these two companies. And with this engineering talent piling on, I don't see a world in which other companies are really able to catch up as quickly as they would like to. So we'll see. We'll continue to monitor that as we go. And speaking of these leaders, one of two of them had quite an update this week that came in the form of a copy of a letter from Commerce Secretary Howard Lutnik, who wrote to Dario Amode. Do you have any light on this?

I haven't really looked much into it. Yeah. So just to catch everyone up with some context, it's been a rocky week for Anthropic, to say the least. They started on such a high last week when they released their brand new frontier model, Fable 5. It's their Mitos grade model, and it smashed every single benchmark. And everyone that used it, including you and I, said that this is the best model ever.

And then three days later, the government yanked it. They sent a letter from Howard Lutnik to, Dario Modi basically saying, hey, we can't have this model getting into the hands of foreign adversaries. So either you KYC and ID verify everyone that uses this model or you shut it down. Dario doesn't have the infrastructure to KYC and ID everyone. So he shut it down for everyone globally around the world. He can't control it. So currently, if you'll listen to this right now, you don't have access to Fable 5.

Now, the one thing that was missing was we didn't know what this letter said. Finally, we have the details. There's a lot of legalese in this, Josh. that's the TLDR. There's a lot of kind of structured responses basically saying, hey, listen, according to regulation, blah, blah, blah, you can't put certain tools and certain products in the hands of foreign adversaries. If this is the case, then basically we have the right to restrict this. And then there was a lot of back and forth between Dario and this letter, basically saying, hey, this isn't a physical tool. This is a software product. Your regulation

basically doesn't cover this. And it was just stuff like that. My kind of whole from this is that Howard Lutnik and Trump or the government, whoever is responsible for this, basically wanted to take down this model, and they were going to go to any lens to be able to do that. And it was triggered by a warning that was issued from Amazon CEO Andy Jassy, basically saying that their security team had discovered a hole or an exploit that could be basically accessed by China. Now, in the interest of the letter, they described what this attack looks like, Josh. And basically, it's an attack that can be exposed through current marks.

that OpenAI is running right now and GROC and Google Gemini. So it isn't really a strong argument, but they're just picking on Anthropics specifically and on this model specifically because they believe that it could be dangerous in the wrong hats. Yeah, it's a bummer. I hope it gets worked out. Fable's an incredible model. It would be awesome for the world to be able to continue to benefit and use that model.

So hopefully all sorted out soon. This next part of the episode is all about open source. It's been a big week for open source, like shockingly big week. actually, starting with Microsoft, believe it or not. So, yeah, Microsoft still exists. They're still in the world of AI or AI, Jason, whatever you want to call it. And they have this product called copilot. You may not have used it unless you are forced to through your employer, but copilot is actually used by like quite a few people. And they have made a large shift this week. If I understand correctly, they're switching from usage-based pricing with open AIs, chat GPT, to fully open source

through the deepseek model, the Chinese open source model. They're running a Chinese open source model on Microsoft servers. Is that right? Yeah. So Microsoft's flagship AI assistant, as you said, is co-pilot. Since the law, I think over the last three months, they've released various versions of copilot, which does different things. One of them uses Anthropics model exclusively, and now this new one uses a Chinese open source model, which they're fine-tuned. Now, their reasoning for using this model, their main reasoning is it's super cheap. And Microsoft does undergone internally a shift from using these expensive coding models such as Claude Code

to cutting down their budgets and using cheaper models. This is their first consumer-facing or enterprise-facing product where you now get access to a Chinese open source model. Now, they received a lot of feedback about this, which is like, you know, is this safe? Is this cool to use? Their claim is that yes, it is, but it remains to be seen. The wider story about this, Josh, is Microsoft's the biggest investor in Open AI. Their stake alone is worth upwards of $135 billion, and they signed an exclusive partnership with Open AI to get access to all their IP. Right now, as we speak, they have access to all of Open Air's model weights for all their models. So they could theoretically train their own foundational model using Open AI's blueprint

and create even their own version. But you would think that they would use Open AI's chat GPT models because it's the frontier. Unfortunately, they've fallen out. Recently, Sam Altman announced that chatjibD will be available through AWS, which is Microsoft Azure's main kind of competitor when it comes to cloud services. Now, before that, chatchipd was only being served through Microsoft Azure. They were the exclusive partner, but they've since had a falling out over the numbers, over the for-profit structuring system and all that kind of stuff. So now they don't really talk. Satya Nadella, the CEO of Microsoft, has been having frequent meetings with Dario Modi. The point is, like him and Sad don't really get along right now, and he's taking a model agnostic approach.

and Deepseek is the first one. And Deepseek just had a crazy week as well. They raised around, I think, $8 billion at a $50 billion valuation. The founder himself put in, I think, $3 billion of his own money to maintain kind of like ownership class shares of this entire company. And he said he is committing to keeping all his models open source. So it's a huge win for China. While you're saying Microsoft has access to their open weights of Chash EBT,

and they're still not doing anything about it, it reminds me of this paper that I wrote. I forget who the author was, but I saw it on Twitter earlier this week. And I think the title was that the window is closed. And there is this like, yes, yes. And it was really interesting because it talked about how the opportunity for a company to become a frontier AI lab has kind of ended. It's almost impossible to catch up.

And this is kind of the trend that we're seeing. And I just want to make some time to talk about it for a sec. Because we saw XAI and Elon and Grock. Like they were the newest entrant. they were the newest competitor, and Elon is among the best in the world at aggregating resources and solving difficult problems, and they got close. X-AI and GROC, they got very close, and they're still close, but they haven't actually been able to reach the frontier.

And that's after building these data centers in a record amount of time by like a factor of five. It's like he did everything you could possibly do and did it all right and still wasn't quite able to catch up. For a company like Microsoft who isn't as hardcore, for a company like Google, who maybe just moves a little bit slower. That's such a disadvantage in a world where this rapid iteration is happening between, I mean, open AI and anthropic. And it feels like there is this kind of pulling away. You see it with the talent. You see it with now the benchmarks. You can feel it when you use fable. You're like, oh, this is a novel thing that I've never experienced before. And it's just,

it's a shift that I just wanted to highlight for a second because it really feels like the companies that are building the frontier AI today are going to be the ones that are going to continue to be leading the pack because of that like self-recurring improvement loop and it's just like the self-fulfilling thing where the better it gets the faster they can iterate the faster it gets it's just this unbelievable thing so we'll probably continue to see that I don't want to get too sidetracked because we have another open source thing to talk about I actually think it's it's worth talking about because I think few of our listeners actually kind of like can see the wider picture and I think it's important to like to like paint it like recursive self-improvement basically like the model like

building itself we've reached a point where these models are now good enough to do it and previously it was a capital race. Spend as much money as you can hire the best talent and then design the best thing. It looks like Anthropic and OpenA.R. are walking away with it. I will say that I think Elon and XAI have a good shot because they just acquired Cursor for $60 billion. Typically, GROC has been terrible. No one has really kind of used it. Their GPUs aren't being used. But if they adopt cursor into the model, you can now jump and kind of catch up at the frontier for coding all that kind of stuff, remains to be seen. Google just lost one half of their entire architecture team,

who's the most important member, arguably. So now I'm kind of like hesitant, but I'm willing to say that maybe they've lost the race now, and they're going to run away with it. So it's pretty crazy to say. Actually, talking about another company that might have lost the race entirely and have actually probably spent the most money

out of all of these companies, META. Meta's had like a crazy couple of weeks, weeks, but a crazy week in particular. So to kind of catch you all up, META over the last year and a half has spent, and I'm not exaggerating, upwards of $35 billion to acquire or hire around, I want to say, 150 people for their AI team. I don't even want to think about what the average amount they paid for that person is. That's crazy. I don't even want to think about it either. But the craziest part is they already had an AI team, and they were already, they weren't at the frontier a year and a half ago, but they were near the frontier. And they had the best

open source model. It was called Lama. And the guy that headed their team, Jan Lacoon, was, you know, he wasn't that bad. They since fired him and re-a-a-a-a-a-a-act the entire team. Every month and a half, now I saw a tweet that someone had measured out the cadence of this. They fire around 4 to 8,000 people from their team. And they had layoffs recently about three weeks ago. So they've re-aulked their entire structure. And this week, Andrew Osworth, who is kind of like Zuck's right-hand man in terms of like laying out the foundational structure for Metis AI efforts, basically said, yeah, the reorg was a disaster. We shouldn't have done this. And it was a terrible move on our side. They're also making some really weird

decisions now. So apparently, there's a rumor going around that 30 to 50% of their software engineers on core teams have now become data labelers. So what this means is they're being tracked and monitored for all the work that they do. And it's being fed into a data system, then use to train AI agents and AI models to do their jobs. So it's effectively kind of like reverse engineering, engineering what they do, and then eventually the AI models will be capable enough to actually do their job. And I guess they're not needed anymore. I think the narrative that's being given is that they'll be up-leveled

and kind of like hired at like a team leader for all these different AI agents. But I just don't see that happening. Meta has completely blown the lead that they had if they had any right now. And I just don't think that they can catch up. It's pretty sad. Yeah, it's disappointing because when, you spend that much money, you expect to get some sort of outcome that's interesting and unique and, like, noteworthy. And the only unique and noteworthy thing we've mentioned on the show as it

relates to things that they've released is their, what is it, the Muse Spark model that was serving as customer service that was then used to exploit all the accounts on Instagram and Facebook. It's just been a total disaster. And when it comes to headcount, I mean, I think at the beginning of this year or the end of last year, they were just around 80,000 employees. That's a tremendous amount of people to be working on like a problem. And I have to imagine that it's, it's very difficult to turn that group to be productive and to move as fast as they need to in order to make actual progress. But at the same time, he spent all this money on building a team. You have to imagine they exist in somewhat of a silo with all the resources they need. And still

they haven't been able to deliver anything. So it's this weird place where, again, I say this a lot, but I think it's true that like meta is really still just feeling like Facebook plus plus. It's like they tried to pivot to meta, to the metaverse. They fail that. Now their name is meta, but they have no metaverse to show for it. They're trying with the glasses. They're in the same boat as Evan Spiegel and Snap that we mentioned earlier in this episode, where the technology is promising, but it's still horrible.

No one's actually going to want to buy it or use it. So that's still too early. They have yet to prove that they can make hardware. They have yet to prove that they can make a compelling AI product. because everything that they've released so far has been a complete and total flop. And it just leaves me in this really weird spot where I want to believe in Zuck and this founder-led company. But nothing that they have produced over the last, I don't know, five, six, seven plus years has actually been compelling. So it's a difficult time for meta and I really hope they figure it out.

Yeah, if there was one thing to kind of like summarize meta in like the last couple of months, they created this internal token leaderboard. Do you remember this? Yeah. They ranked employees on the amount of tokens that they spent AI tokens to be productive, and it was a measurement of productivity. Obviously, for any sensible individual, the number of tokens you spend doesn't necessarily correlate to how smart you are, how intelligently you are using tokens.

They have now used that exact same board to figure out who they fire for every single layoff. So if you're a major token spender, they're like, oh, you're misusing your tokens. We're letting you go. That's what they did in the recent layoffs. Sad, crazy, I hope META figures it out, but talking of lean teams that are spending much, much less, probably like one-fiftieth of what META spent over the last couple years, but are still not only maintaining the frontier, but in some cases beating it, we have a new Chinese open source model from the Chippoo, ZH-I-P-U, I believe, Chinese Lab, GLM 5.2,

open weights, MIT license open weights. So, you know, I think that means after you get to a certain number of users, you have to like credit the company or something like that. But I just want everyone to focus on this chart for a second. It's on agentic coding performance. Now, if you're wondering, why isn't it just looking at regular AI coding performance? Why not? Why is it looking at agentic? It's because a lot of these agents are the new way of working with AI tools going forward. It's not just going to be you in a single session or in a single chat window talking to one specific LLM. You're going to speak into multiple instances of the same LLM and it's going to be independently talking to each other and doing a bunch of work. That's what makes you exponentially more

productive. And so agenic coding is a strong measure of whether the model is good enough. Now, GLM 5.2, when compared to Claude Opus 4.8, pretty much matches it on its score. Now, notably, this doesn't include Fable 5, which is the restricted model that we can't get access to right now. but it's pretty impressive to have open source models still catch up with the frontier labs. I'm still trying to figure out how they're able to pull all of this off. They don't have any of the money. They don't have any of the chips.

I'm kind of encouraged to think that this is probably a data play so that Chinese have access to a lot of data that they can use to smartly train models. But it's just impressive nonetheless. And it has a really cool ability to create front-end UX design, which is typically something that's hard for models to achieve. visual interpretation, visual intelligence, but what you're seeing on the screen right now

is, I believe, like, a personal website that was created in one shot in five minutes. It's pretty cool to see. Yeah, very cool model. It's interesting to see the benchmarks. They always have to be taken with a grain of salt because when you see them being compared to other models, I'm not quite sure.

It's close. And for that, it's respectable. It's like, okay, even if you've come close, if you're a little bit above, a little bit below, just being in the ballpark is respectable. I find that a lot of times now with the models, you just got to try it out and play around with it really get a feel for it because a lot of that, the kind of nuanced feel that you get when you

play with the model is the thing that's actually meaningful versus the benchmark. So that's kind of how I've been doing it. I'm looking to try and play around with this thing and see if it could do some cool front-end design for us. Maybe rebuild our website. That would be kind of cool. But it's really exciting to see China still kind of hanging on. They're still pushing forward to the frontier. Every time we count them out, they come back and they clap back with something else. So an exciting release from Jipu or however you pronounce them. I'm not entirely sure. But yeah, it's very cool.

And then also another ex-open source company named Mid Journey. You might remember Mid Journey because they did the image generation model. They were one of the very earliest people to do. Yeah, exactly. They were one of the very earliest people to do image visual generation with AI. Well, they've pivoted in a pretty meaningful way. Like, I'm not even sure this is the same company. When I saw this news this morning, I had to click on the pro.

profile and make sure it was the same mid-Journey that I was thinking of. Because it does. The AI company that you know and love as the Image Generation Company has created a physical hardware device and a new division of the company that is known as Mid-Journey Medical. What is Mid-Journey Medical's first device? Well, if you're seeing on screen, there's a person being submerged underwater. This is not AI. This is a real person, I think, at least. And the person is being submitted or submerged underwater to get a full-body MRI scan, which is. shocking but also remarkably impressive I mean this is something that would traditionally take 60 or so plus minutes and now they're doing it in a minute like

how does an image company go from that to this this is a pretty huge leap I think founder context is important here the founders parents one of them is a dentist and I believe his mother is the tria MRI physician so he's probably inspired in some way then you know he's in the image game Josh it's not too far to I guess leap to imaging right yeah imaging and and so Honestly, what I thought was that this device used AI to kind of, you know, figure out what the insides of someone might look like. And I was getting super excited. Unfortunately, there's no AI involvement here.

But nevertheless, this is a massive, massive scientific breakthrough. And I'll tell you why. You mentioned one thing, which is typically an MRI scan takes 60 minutes to occur. You need to go to like a regulated, like, medical institute to be able to access this. I got a prenuvo scan recently. It took me 60 minutes. It's big, scary device system ages to set it up. This is a device that you step onto.

It submerges you into, and I'm quoting here, warm water. So you're not going to cold. It's warm water. It's non-invasive, so there's no radiation at all. And it takes 60 seconds. So the time it takes for a coffee to be made, you'll have a full-body MRI scan. The other interesting part is they're not putting these devices in hospitals. They're putting it in potential spas.

Or they're setting up coffee shop and spa type situations where you can grab a coffee, and just get an MRI scan if you bring your swimsuit along, I guess. And this is really cool because this is going to bring the cost of MRI scans down massively. MRI scan imaging is just amazing in general for figuring out the health of your body internally. But it's often priced a lot of people out who don't have health insurance or who don't want to spend several thousands of bucks to get access to it and then book an appointment and then wait months to do this. Now you can just walk into anywhere that has one of these machines. I don't know when they're running out and get access to this.

And the plan for the future is to use AI to look at these different images and help you with the diagnosis part of things, which is going to be pretty cool. Oh, I love this. It's so cool seeing a company who has a lot of money. They're clearly very well-resourced. I actually use it for something that's applicable that can make people's lives way better. And I'm sure there's going to be a lot of people who have a lot of critiques on how the technology works, how they haven't actually built any hardware. But just moving the puck in this direction, just coming up with the concept and figuring out how to actually turn this into reality, in a way that's going to improve a lot of people's lives, I think it's awesome. And it's following this trend that we have of this like health and wellness and like bioscience's trend that I think will probably continue to see a lot of, particularly as these AI models get much better at understanding

the human body, at understanding what makes us tick, how to improve things. And being able to capture data like an MRI machine is the first part of improving the human body. We first need to understand how it works, what it's made up of, where the problems are. And this will hopefully help to do that for a lot more people. So in the case of Mid Journey, this is awesome. Congratulations. Very exciting work to see. I love the pivot. More people should pivot from ImageGen AI slop to saving people's lives and health care. That seems pretty tight to me. But that is everything. That's our whole week rounded up. If you've made it this far, thank you. We had a few pretty great episodes this week. One of them is going crazy with Leopold Aschenbrenner. And the reason why he is short and video with a

massive amount of his portfolio. We had another one covering the SpaceX IPO and whether or not you should buy it or not. So if you haven't, go get yourself caught up on that. But if you've made it this far, that's it. You're good. You can go into your weekend feeling fully up to date on all things, AI frontier technology and I guess healthcare now too. If I were to leave all of you listeners with a final prompt, it's been a while since we've asked what the latest, coolest thing you guys have been doing with not just an AMOR, about any AI feature. I would love to hear from you guys. If there's anything that we should be talking about that we haven't cubbyer. Are there any use cases that you guys, the weirder, the better, by the way, I just want

to emphasize this. Like, don't feel like it needs to be anything impressive that makes you a million dollars. Like, if you're using this to design your grocery list and you think it's a pretty unique or novel thing, tell us. I want to hear about it because we're always looking for use cases where these AI models basically apply to the wider, like 99% of people who haven't gone past chat GPT at this point. So if you're using it for something cool, let us know in the comments, and wherever you're listening to us, whether it's on YouTube, whether it's on, Spotify, whether it's on Apple Music, wherever, please subscribe and give us a rating where you can turn on notifications if you're on YouTube. You have no idea how helpful it is for us.

Share with your friends. Share it with your friends. Share it with your friends. It's very important because it helps us get a wider reach. Josh and I try to grow this channel into the biggest AI and Frontier Tech channel that we can. And we're excited to get that. We recently, I think we got like 1,100 new subscribers on YouTube over the last less than 24 hours. So hello, to all of you. It's lovely to mean you. We come up with Frontier AI episodes, 20 minute chunks four times a week. Well, they're a little bit longer now. We're working on 20 minutes. A little bit longer. Some of the roundups are a little bit. All right. And a views less up, which goes out to 100,000 of you every single week, twice a week. So we have an essay and we have the

five weekly highlights if you don't want to watch all the videos and you just want to catch up through text. But aside from that, that is it. That's it. Go enjoy your weekend. Have fun. We'll see you next week. Goodbye everyone. We'll see. Ha ha.