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.
We're back. France. We miss you guys. We miss you guys a lot. Yeah.
Speaker 2:It was rough.
Speaker 1:It's genuinely brutal. We feel deeply unwell Yeah. By around 1PM if we haven't gone live
Speaker 2:Yeah.
Speaker 1:Regardless of the time zone.
Speaker 2:The sweating and shaking is crazy.
Speaker 1:Yeah.
Speaker 2:It is a full I mean,
Speaker 1:the yeah. And the withdrawals. I mean, maybe it would get better after a few weeks.
Speaker 2:Yeah.
Speaker 1:But, yeah, the symptoms brutal. Are Yeah. Brutal.
Speaker 2:Anyway, it was a high stakes trip. Of course, Jordi has battled with French president Emmanuel Macron in the past. And then
Speaker 1:And what and
Speaker 2:appeared to be a direct shot at me. France recently banned nicotine pouches fully. Very controversial.
Speaker 1:Pretty much all oral nicotine products Yeah. Banned. Yeah. So As of April.
Speaker 2:Canned Lions. It used to be mostly about creative advertisements giving awards for creativity and there's a big focus on the creator economy. It's not packed with a list celebrities, but it feels like everyone has a following of some sort. And that was a big point of discussion. Like, there was one level of conversation that was just like creators are the future.
Speaker 2:Everyone's heard this pitch a million times. But the bigger discussion was something that Andrew Ross Sorkin from Squawk Box, CNBC, DealBook mentioned, talking about the trade offs between independence and consolidation. You've talked about this before, just the idea of a roll up of independent sub stackers. Where is the right economy of scale in the creator economy? Maybe not everyone should be independent, maybe not everyone should be with a large organization.
Speaker 2:And this trade off is sort of interesting to dig into as we see both creators go bigger and produce more expensive shows, and then also the legacy media companies are getting better at social media type content. So we can kind of dig into this. I'm flying back from Cannes. Something I've been thinking about. I was taken aback by how some of what I thought were most success the most successful influencers say that they need to spend so much money to keep up, that they're not making the kind of headline popping numbers we often read about when he asks for thoughts.
Speaker 2:I think there's a few things happening. One is the big transition from independent creators pivoting from making content to making shows. And you can look at Subway takes in this in this way. The other show is keep the meter running much more high much more produced, much more edited Yeah. Much more costly to to produce.
Speaker 2:But you're still faced with some of the awkwardness around how do you actually monetize that. There isn't a logical ad break, although I think there should be in a three minute subway takes video, a three minute episode. I think that it would be fine to have a five to ten second mid roll in the middle of that, but that has not been done. Like, that floodgate hasn't opened yet. So most creators are doing so either it's either a really polished vertical video about their own content, and then they'll do a promoted post that's another thirty second, sixty second, multi minute vertical video in their same style, but purely sponsored content.
Speaker 2:And a lot of those videos don't go as viral. And so you have this awkward trade off where the content that's great, that people want, can't be monetized, so you have to do one and one. It'd be a little bit tricky. Like, part of the way our show works is that we're just talking about whatever, and then we'll tell you about Cisco, critical infrastructure for the AI era, unlock seamless real time experiences, and new value with Cisco. It would be different if we had to do every other every other episode is just purely about Cisco, and then we go back to about just the news and talking.
Speaker 2:And so that's that's something that's like starting to evolve in those higher cost structures. I think that they still have more enduring
Speaker 1:Four team out with their top creators list and Yeah. Earnings. Mister Beast was at the top with 300,000,000. Huge. I would imagine.
Speaker 1:But that's revenue. That yeah. Like Yeah. Earnings is is not the right word here. It should just be gross revenue.
Speaker 2:Yeah. Yeah. Yeah.
Speaker 1:I I have no idea what what Jimmy's margins actually look like on that 300. It it would appear he's spending like
Speaker 2:But he's obsessed with reinvesting for the
Speaker 1:Yeah. That's always been a part of
Speaker 2:his to Joe Rogan who has these big deals and he seems completely content to be in a tiny Yeah. Yeah, people are still breaking through, but they're breaking through with new more polished products that, of course, have higher costs. Simultaneously, traditional media companies are starting to dial in content strategies for traditionally creator led platforms. I think New York Times has done a fantastic job here. You have the Ezra Klein show and Ross Douthit, who are sort of like the left and right perspectives.
Speaker 2:And their shows have become really successful on YouTube a lot due to the packaging. They understand how to get a good title and thumbnail together. The editing, the production values are really, really high quality. And so those podcasts do very well on YouTube in a way that other shows from legacy media companies, it's been hard for them to break through traditionally. But they certainly have and it's interesting that it's recent.
Speaker 1:The other thing that was interesting, France appeared to put Brian Johnson on Wait. What happened? A pack of cigarettes. Oh, yeah.
Speaker 2:You had
Speaker 1:Pull up this image. That does look like Brian Johnson. Really out of pocket You take a picture too. To take to take an asset.
Speaker 2:It works pretty well though. Don't die. Don't smoke.
Speaker 1:You know, it's his it's his message. I think he would approve this. But I saw I saw this and I had to take a picture because I was like, is it actually uncanny? He has like a lot of pictures of himself
Speaker 2:It does look
Speaker 1:like down like that, eyes closed.
Speaker 2:Also, this looks like a man being put in a body bag, but at the same time, it looks like just a man who got a getting a beautiful night's sleep getting tucked in after a nice couple of hot heaters. He ripped a couple darts, had a couple of beers, crushed a couple cold ones and now he's ready to get put to bed by someone.
Speaker 1:Yeah. This is actually much more tame than some of the others.
Speaker 2:Some them are gruesome. Like Yeah. So gruesome showing like the x rays of cancer patients and like black lungs and stuff. Like this is pretty mild. If you look like this guy after you're smoking a whole pack, like, you're doing pretty well.
Speaker 2:The end state of of this debate, it's it's interesting. There's a little bit of grass is always greener on the other side going on with independent creators thinking, oh, maybe it'd be better if I was with a larger organization, and then vice versa with a lot of the folks inside big organizations saying, ah, if I left, I could probably make so much more money. I don't think it's a one size fits all approach here. I think it's more of a sorting process. Some creators will find elegant business models outside of large organizations.
Speaker 2:Like, those 80% to 90% EBITDA margins still exist for certain categories,
Speaker 1:certain Yeah. Look at obsession. Right? It's like a low margin, highly competitive, but it's hits driven. So you can have one breakout hit That's
Speaker 2:great example.
Speaker 1:And, you know Yeah. Have insane margins.
Speaker 2:Yeah. So I think it's more of this like sorting process. Some creators will find elegant business models outside of traditional organizations. Others will probably wind up leveraging the possibility of going independent into more meaningful contracts within traditional media organizations. But it's it's increasingly clear that a one size fits all approach, perhaps with a tagline like independent creators of the future.
Speaker 2:I don't think that quite fits anymore to each their own. I think everyone sort of needs to find their own way and figure out what's right for them. Smart glasses are inevitable, but what or who are they for? Says Christopher Mims in The Wall Street Journal. This is on the back of the new Meta Glasses that launched.
Speaker 2:Chris Christopher Mims says, the deluge of smart glasses has begun. Has it? It feels like I don't know if it's a deluge. It feels more like a like a just a trickle or like a steady stream. But maybe I've been following it too closely.
Speaker 2:I think that's all of my problems with analyzing this category and we'll get into that. Yeah. My criticisms are like very odd technically.
Speaker 1:Yeah. And it's and it's and it's so Meta has been selling a lot of glasses. The products are the products are
Speaker 2:Dozens? Hundreds?
Speaker 1:No. They've been selling I believe they've already I believe they're already in the in the mill in the millions of units of not Millions of units. Not the
Speaker 2:Hundreds millions of dollars?
Speaker 1:I believe so.
Speaker 2:Okay. Ben. I think it's getting I I Work on facts. I agree with you. Think it's getting Work on facts.
Speaker 2:But
Speaker 1:what I was say is that it's for products that are functionally like a great pair of glasses with a camera attached. Yeah. Right? These are not like I I don't see a lot of people talking about like, oh, I'm replacing AirPods entirely Yeah. With these even if even if that is one
Speaker 2:Is that just because
Speaker 1:And that is also with having being able to sell through a one of one ad unit on the greatest distribution platform in history,
Speaker 2:But which is they're going up against stiff competition. Like, the Apple Vision Pro is so good that like, you know, it's like we are job's finished. Like, we have the Apple Vision Pro. Why
Speaker 1:so funny. So I never tried the Apple Vision Pro until this trip. And I asked John. I was like, hey, can I can I take him for a spin? And you were like, oh, it's like kind of a hassle to set up.
Speaker 2:It is.
Speaker 1:And and You have to like put it in mode. But John was like, no, I want Jordy to experience this. Yeah. And so we go through this whole like hassle of setting it up and I'm like twenty seconds into the dinosaur demo and the flight attendant comes up and is like tapping me on the Yeah. Like actually out of virtual reality.
Speaker 1:Yeah. It was like the audacity.
Speaker 2:And I think the question was
Speaker 1:just like, are those cool? Or something like It wasn't
Speaker 2:it wasn't like, sir, you need to put on your
Speaker 1:seat or something. It was like I was I was fully immersed
Speaker 2:Yeah.
Speaker 1:And it and it completely ruined it for me.
Speaker 2:And what's funny is like the whole pitch for the Apple Vision Pro is like, you don't need to take it off to talk to somebody because it will show the eyes. And now Tyler's on his way back.
Speaker 1:You went to France for raising
Speaker 2:Raising Cannes. Yes. Raising Cannes. Raising Cannes is good.
Speaker 1:Yes. Tyler got left at CDG. Yeah. Yeah. Shoddity.
Speaker 1:But I have the We Timu. Timu Tyler.
Speaker 3:The the I have the answer for the Meta Glasses.
Speaker 2:Okay. Yeah. How how we doing?
Speaker 3:So since 2025, there's been 7,000,000 plus sold across OKIE and Meta. $2,000,000 was just the Ray Bans.
Speaker 2:Okay.
Speaker 3:And
Speaker 1:since Wait. 2,000,000 units.
Speaker 3:2,000,000 units.
Speaker 2:Okay.
Speaker 3:And that's the estimated gross retail sales was 2.1 to 3,000,000,000
Speaker 1:That's pretty good. Yeah, John. You the way you were framing it, it was like, yeah, we've sold 7,000,000. And you're like, $7,000,000? And it's like, no, it's $77,000,000.
Speaker 1:Like, it's a lot of units. Yeah. It's lot of units.
Speaker 2:Yeah. It's getting there.
Speaker 1:And if it wasn't meta Yep. We'd be like, wow, this company has like insane product market fit. Yeah. Selling 7,000,000 units of a new Yeah.
Speaker 2:Consumer product. Whoop or something like that, we'd be very impressed. Yeah. I I I I think you're a 100% correct.
Speaker 1:And it and it would be like, it could probably be like a standalone Yeah. Public company.
Speaker 2:But people would still be slamming it for it for being like way worse than the Apple Vision Pro in terms of watching watching old films. I watched Master and Commander on it. It was great. Anyway, let's go back to Christopher Mims in the Wall Street Journal. He says, at the low end, there's a new line of Meta Glasses, which start at $2.99 and play well with vision benefit plans.
Speaker 2:Interesting. So you can buy them with insurance. For another $100, there's a pair aimed at fans of Kylie Jenner. We saw these in action, and they looked great. They they I and I think reviews are very high quality.
Speaker 1:If if I hadn't seen the launch, I wouldn't have noticed that they were matte glasses Yes. At all. Yes. I would have just seen has they were like
Speaker 2:this jewel. You can see it. It's that small white, like, glint right there. And I was so confused by this, but this is purely a me problem because I assumed that that was a recording light or or some sort of display. Because if you look at the Meta Ray Ban displays, you'll see that there's slight grooves in the glass for forming the beams to show you the display.
Speaker 2:And so I was like, oh, that's a that's a smart feature. But it's in fact not. It's purely aesthetic. And people who are in the market love this. I was confused by it, but I think I might be the only person that is confused by it.
Speaker 2:So I I don't think it's a problem whatsoever. But Metas is $800, and specs from Snap is $2,200. They are so chunky, they look like something you might see in a Prada ad and never in real life. Silicon Valley has been trying to make smart glasses happen for more than a decade. Remember Google Glass, Facebook and Instagram parent meta platforms has captured more than 80% of this market.
Speaker 2:Not bad. But in 2025, that meant selling just 7,000,000 pairs of glasses. By comparison, the world buys more than a 100,000,000 smart watches a year and more than
Speaker 1:say does about like the broad investor views on smart glasses overall? If Meta has 80% of this like emerging category Yeah. That's selling 7,000,000 units and they get like basically zero credit for
Speaker 2:it. When I see 7,000,000 pairs of glasses compared to a 100,000,000 smart watches, I feel like this is perfect trajectory. Like this is great.
Speaker 1:That Meta ever makes a free pair of glasses that's ad supported?
Speaker 2:I don't know. I think that an ads business would actually be much more attractive. When you when you say, oh, they're making a billion dollars in revenue, that's not a 100% margin. I think about, well, what is Threads revenue gonna be? And that's gonna be much higher margin.
Speaker 2:Like, Threads is probably more exciting to the core business and to the financial outcomes of
Speaker 1:I just think I I think if you
Speaker 2:I understand it's exciting. I like hardware. Like, it's cool. It's just like it's it's it's hard to underwrite to like some oh, it's moving the needle on cash flow.
Speaker 1:I would This is gonna pay
Speaker 2:for another data center.
Speaker 1:I would expect to see just like we saw the ad supported Kindle Mhmm. I think we'll see the ad supported Meta Glasses.
Speaker 2:Are the ads on the inside Yes. For where? Or on the outside everyone
Speaker 1:working the user. So you're walking on a street and and you walk past a restaurant, it pops up like, you know, get get a free matcha.
Speaker 2:I I don't like that. I like the other way. I wanna be wearing the glasses and I want it to be barking ads at everyone around me while noise cancelling those ads to me. So I don't get I don't have to hear them.
Speaker 1:You become a a loud billboard.
Speaker 2:Exactly. I pay for the no ads version but you didn't pay and you're sitting next to me.
Speaker 1:You dress up as a guy on the street Yeah. And you're pretending to do like what do you do for a living interviews. But then right when you get up, instead of saying what do you do for a living, you just activate the ads Yeah. And you just try to bark like Yeah. It's four or five
Speaker 2:ads. What is your intelligent cloud provider? Is it Railway? What which one is it? Let me tell you about Railway.
Speaker 1:Railway is the all in
Speaker 2:one intelligent cloud provider. Use your favorite agents to deploy web apps, servers, databases, and more while while Railway automatically takes care of scaling, monitoring, and security. That would be a great ad to bark out of your glasses. The idea of a face worn computer has garnered skepticism, even hostility, which is understandable. When most of us are trying to reduce our time in front of screens, it seems absurd to mount them right in front of our eyeballs.
Speaker 2:Why should we be okay with everyone we meet pointing Internet connected cameras at us? It feels like an assault on what little privacy we have left. So why do carry a flash bang around? Then there's the built in AI. Artificial intelligence let these glasses identify objects in our field of view.
Speaker 2:That is not a very compelling use case for me.
Speaker 1:Hot dog, not hot dog.
Speaker 2:It's just like I'm looking at a building. What is this building? It's a building. Like, I'm what is this tree? What what do you see?
Speaker 2:It's a tree. Like, I don't know. Like, do you care if it's a ficus tree or not? It's pretty rare. South African meerkat?
Speaker 2:How often are you running into an unidentified meerkat? Like, this is not that big of a problem. The bigger one is that you're just able to talk to an AI agent and have it look up things for you, have it do whatever you want in on your phone, in your world, on in your agent. But if it's very rare that you're seeing things in the real world, oh, my God. What can is this?
Speaker 2:What is this? I need to know. It's like, come on. I know what a Diet Coke is. And yet there will be smart glasses and you and I might even choose to wear them.
Speaker 2:They're technological inevitability, a new consumer tech arms race. For the last and best real estate on our bodies. I envision that someday all glasses should could be smart glasses, said a man who should know. Zaid, who heads the division of Qualcomm responsible for making the chips that go into the Meta the Meta, Snap, and Samsung smart glasses. Well, Meta is working on a bunch of other stuff.
Speaker 2:They're working on artificial intelligence. You may have heard. Zephyr says they seem to be pretty focused on coding right now. I'm pretty sure this is the largest coding training data generation effort in the world. They have the compute.
Speaker 2:30 to 50% of engineers and core teams have been forcibly forcefully reassigned to data labeling. Meta's radical plan to become a leader in AI will require vast amounts of capital, and her door was open to discuss innovative financings financing structures and partnerships. They might wind up raising billions more to fund the build out of MSL and Meta's AI efforts. One engineer told this reporter that the whole situation feels like the movie The Hunger Games when tributes are randomly selected and then removed from their environment to something completely different, except at Meta, many more folks are being affected with between three to five from a 10 person team going from building products used by hundreds of millions to giving human feedback on AI generated GitHub repos over and over. So a wider impact than in the Hunger Games with less drastic consequences.
Speaker 2:Yeah. Still air conditioned with free snacks which they are improving apparently, dude.
Speaker 1:So, Duck is we're gonna have Meta code.
Speaker 2:Yeah. I mean, it seems like they're trying to have a model that they would sell it on the API or something?
Speaker 1:I was expecting Meta to spend enough money to get the Meta AI, you know, chat, GBT, Gemini, Claude competitor Yeah. To the top of the charts and keep it there. Mhmm. Because I expected Zuck to think like, hey, if I get people asking questions all day long, I can just target them much better Oh, you mean charts. Yeah.
Speaker 1:So I expected them to go super hard on consumer. Yeah. They don't have background enterprise. They bought Manus but even Manus was like sort of unclear what they were gonna do over time. Maybe they would integrate it into Manager.
Speaker 1:But I expected them to just say like Zach say like, hey, I'm gonna spend a lot of money Yep. And I have billions of users already. I'm gonna do whatever it takes to get a consumer chat app to the top of the charts, keep it there, and over time, just focus on this one thing. Yeah. And the meta AI app sitting at 17 in the app store right now, you know, not not not nothing.
Speaker 1:Like, the market is not is clearly not like sending signals like, hey, we're really excited for you guys to, pivot into enterprise and, like
Speaker 2:Yeah.
Speaker 1:Spend hundreds of billions of dollars competing against
Speaker 2:Google,
Speaker 1:Microsoft, Google, OpenAI, Anthropic
Speaker 2:SpaceX.
Speaker 1:You know, all the Chinese labs, SpaceX. Like, this would be an absolutely incredible come from behind story. Yeah. But I think it's reasonable that the market is saying like, no. They're not pricing it in.
Speaker 2:Certainly. Well, there's other news. Zuckerberg directed Meta to build a prediction markets app, which is a wild shift. The experimental app internally called Arena would be independent of Facebook and Instagram. It could compete for attention with Polymarket and Kalshi, the biggest prediction markets, says Mike, Mike, Isaac.
Speaker 1:I think sentiment is at near rock bottom. Yeah. Sentiment on poly on on prediction markets is like, I don't know if it's at an all time low. I think users actually like generally Yeah. Like they they like
Speaker 2:It's way low when it was like, oh, we're gonna find out who the next president is. And now it's like, it's a gambling thing for sports. And it's like the vibe has very much shifted.
Speaker 1:Yeah. And I think I think a lot of people have had like have benefited from using prediction markets as a data source.
Speaker 2:Like A lot of people have benefited from sports betting. There's a story about a guy who won $6,000,000 on DraftKings and built a mansion with a bowling alley in it.
Speaker 1:So what I was saying is like prediction markets are are
Speaker 2:That's what you're
Speaker 1:saying. You know, revenue's exploding. Yeah. I think that's a sign that users like the product. Yeah.
Speaker 1:I appreciate them as a data source. I always have. Yeah. But sentiment around them broadly is Yeah. Like bad.
Speaker 2:Yeah.
Speaker 1:And then sentiment around meta
Speaker 2:Yeah.
Speaker 1:Is at almost an all time low Yeah. Least in the capital markets. People are, you know, very very unhappy. And so maybe Zuck is just saying like, f it. I don't think it can get any worse.
Speaker 2:But, feedback to the Kylie Jenner Meta Glasses has been way way better. Wow, Silicon Valley finally figured out who controls consumer spending. This is not finally. Meta's been partnering with great celebrities for a really long time. This does not come as a shock to me, but it is the right person to
Speaker 1:And it's so partner with. It's fascinating because we were trying to estimate like how much Meta spent on this like partnership campaign. Yeah. I had I was putting it at like somewhere around like 50. I have no information on it.
Speaker 1:Million? It's like 50,000,000. Yeah. Yeah. How much does it cost to get the top influencer
Speaker 2:Yeah.
Speaker 1:In the entire world to do a campaign like this Yeah. And launch a product with you Yeah. You know, maybe something like $50,000,000. Then Alex Heath Yep. Heathenator Yep.
Speaker 1:Who we're with this week.
Speaker 2:Scoop athlete.
Speaker 1:Scoop athlete Yeah. Comes out and scoops that Snap is working on a deal with Robert Downey Junior to spend a $100,000,000 Yeah. Potentially for a partnership there which is deeply confusing Okay. For like, you know, a lot of reasons like the Kylie Meta
Speaker 2:Yeah.
Speaker 1:Campaign makes a lot of sense. Kylie is like one of the like probably the or one of the top influencers on on Instagram. Yeah. Right? So it's a deep existing partnership.
Speaker 2:Don't know that this couldn't be this isn't deep. It could be a super deep integration. We could be looking at like Iron Man four, Snaps Back Spectacles. And the whole movie could revolve around Yeah. Iron Man taking off the suit and only using the Snap Spectacles to stop Thanos for the fifth time in a row.
Speaker 1:Previewing today, GPT 5.6 Soul. Next generation model.
Speaker 2:Named after Soul, bro?
Speaker 1:That's crazy. Yeah. I don't know about that one.
Speaker 2:Because I would assume Soul is the biggest, then Terra is the medium, and then Luna is the smallest in terms of the model size.
Speaker 1:Of course, by US government directive, access is limited to only 20 preapproved companies.
Speaker 2:20? That's so small.
Speaker 1:Which is brutal because our very own Dan Shipper will not be able to do a vibe check on it at least.
Speaker 2:He's gotta get access soon. But a lot of this comes on the back of the distilled gate that's happening in through OPEX as Alibaba has distilled Claude many, many times. And I saw a chart of the amount of tokens that are getting resold off the back of that API, and it is staggering. It's a true somebody was like, they they professionalized fraud Yeah. Which seems accurate.
Speaker 2:And so having really tight control, it makes sense in the distilling world. It's still very unfortunate.
Speaker 1:To go back to your other question. So SOL is the frontier model. Terra is a balanced model for efficient everyday work. And then Luna is a fast and affordable model for high volume
Speaker 2:Okay. So Cool. There's there's more data center backlash. Data centers, do they work? Theo Von says nobody wants a data center, dude.
Speaker 2:And the people that want them to me, they seem kinda evil. Nobody wants this. One of these companies is gonna own all this information. He had some very funny takes
Speaker 1:about He said emotional credit score.
Speaker 2:Emotional credit score would be weird. Yes. The data center issue needs to be messaged much more clearly and
Speaker 1:And and I think
Speaker 2:There's such a wide gap between the water and the power
Speaker 1:I'm assuming I'm assuming he starts this way but then he says, I'm I'm gonna be switching distributing my podcast on
Speaker 2:That's such a bad counter take because truly like the data centers that are required to power the YouTube and podcast stream, it's like so so small compared to what AI is doing in terms of data center construction and and and the power requirements and the build out requirements. I mean, like, you fire off one AI agent, you're using, like, u it's equivalent to YouTube for, like, ten years, probably. Anyway, we should move on to other stuff. OpenAI limits access to new models, citing government security concerns. Company says White House review of AI releases shouldn't become long term default.
Speaker 2:Ban on mythos model remains. Opening eyes said it's limiting access to its newest artificial intelligence models after discussions with the Trump administration, but warned that the recent trend of the White House restricting industry activity on national security grounds on a case by case basis shouldn't become the norm. The maker of ChatGPT yeah. I saw a lot of, like, safety people being like, this is the worst of both worlds. It's not some sort of, like, broad reaching apply you know, just like the rule exists and then everyone can easily comply.
Speaker 2:It's, like, the highest touch government intervention and it makes the government have much more power over AI, which is something the safety folks, I think, at least the ones that I saw talking about this, were not a fan of.
Speaker 1:But Yeah. It's also just like, you know, there's the new it's GLM 5.2 Yes. That a lot of people saying like, hey, I could use this as a daily And all
Speaker 2:the lab leaders have said open source is like six months behind. And so now we're like approaching six months past like the last jump in paradigm and like massive cyber capabilities. So those cyber capabilities are coming to fruition in the open source world. There's an interesting dynamic between the risk of, like, bio stuff and cyber stuff. So if you have a model that can hack system, it can usually patch that system for that hack in far less than six months.
Speaker 2:Like, probably in like an hour. Like, it can it can see, oh, there's a there's a there's a exploit here, and we ship the code, merged it in, and now that exploit is closed. And so the gap cyber between, like, you create the HackMachine 9,000, you can hack everything, but instead you use it for legal reasons, for economic reasons, to patch all the systems, that's great. The the problem that I'm worried about now is if there is a if there is a world where building the counter building the counter, like the good version of something, takes longer than six months. So the example I'll give is is in bio.
Speaker 2:So let's say that you have a next generation model. It's really good at finding viruses that, if created, would wipe out a ton of people. It's also really good at building early warning detection systems and counteractive, like take this pill and you'll no longer be susceptible to that deadly virus that is now possible. That's fine as long as you continue to roll out the biosecurity initiatives first. You should be good because by the time the open source hacker, nefarious actors, state actors get access to the bioweapon 9,000, you've already patched your society so that you have biorisks, security, early detection, all the different medicines that you need to fight back.
Speaker 2:But if manufacturing all of that and actually deploying the biosecurity, the biorest, like the good version takes a year, but the frontier is only six months behind, you could wind up in a situation where manufacturing the virus is much shorter than manufacturing the antivirus. Well, let's go back to the OpenAI news. OpenAI followed a similar process with versions of GPT 5.5, like Cyber, which demonstrated an ability to discover vulnerabilities in software that could be used in cyber attacks. The similar capabilities of Anthropix Mythos models prompted the White House to increase its oversight of the industry and overhaul its light touch approach to AI. There was, of course, that news that the NSA commented that Mythos was able to hack into a bunch of their systems.
Speaker 2:That was during a red teaming exercise. The unprecedented intervention led Anthropic to halt all access to comply with the new rule. Anthropic had previously worked with the administration on a limited rollout of an earlier version of Mythos. And OpenAI said it hoped to make 5.6 generally available in the coming weeks and that the government's current approval process represents a transition period while President Trump's recent executive order on model oversight is implemented. Quote, we don't believe this kind of government access process should become the long term default.
Speaker 2:It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them. We are taking the short term step because we believe it is the strong strongest path to broader availability, especially as the open source frontier advances. It feels less and less reasonable to keep closed source models locked away.
Speaker 1:Well, why is the open source model frontier advancing? It's because the models are widely available. If you make these models available Yeah. Through everyday consumer subscriptions, there will be groups that will create networks of, you know, thousands of accounts and distill the models. And then you're giving away this like advanced capability.
Speaker 1:Whether or not that causes like, you know, these sort of like doomsday Yeah. Scenarios, you know, not not entirely clear. You think Meta's stock would pop if Zuck could go get a preliminary ban Yeah. They on some of their new models.
Speaker 2:The aura farm of the ban, it does seem to be working. It's too powerful. There's other OpenAI news. I mean, the the chip Jalapeno launched designed with Broadcom. Jalapeno is purpose built for LLM workloads powering Chachibidi Codex API, future agentic products.
Speaker 2:Very exciting.
Speaker 1:Crazy thoughts. Crazy timeline here. What? Just how quickly they were able
Speaker 2:Oh, yeah. Yeah. A lot of people are saying, like, this is the first chip designed with AI agents in the loop, you can write the instructions set much quicker. Sort of unclear exactly.
Speaker 1:You gotta give a lot of credit to the humans on both sides.
Speaker 2:Yeah. But it is just cool because for a long time, it was like, oh, it'll take five years to launch a chip, and now companies are clearly doing it faster. And so the the the go to market cycle is tightening. Also, apparently, OpenAI made deals to purchase 40% of the global raw undiced DRAM wafer output until 2029. Millions of raw DRAM wafers that cannot be used until they're processed.
Speaker 2:Smart move, going deeper and deeper into the supply chain. Always always interesting moves from the deals guy at the top.
Speaker 1:Speaking of timelines
Speaker 2:Yes.
Speaker 1:People are comparing GTA six with the Burj Khalifa. Mhmm. Burj Khalifa was built for 1 and a half billion dollars and took six years. Wow. Apparently, GTA six is $2,000,000,000 and it's taking twelve years.
Speaker 1:But but at the same time, people really have said like it is the Burj Khalifa of of video games.
Speaker 2:It's got two Burj Khalifas. Opening others topic, Stripe and Bill Gates are putting $500,000,000 into funding a new organization called Intercept. Intercept's goal is to prevent the common cold and the flu and eventually eliminate all respiratory viruses completely. Incredibly cool project. Yeah.
Speaker 2:Just like an incredibly cool, like, tangible thing. I saw a stat that I saw this chart of what percentage of weeks are parents sick with something viral generally as a function of how many kids they have. And it's like a fast takeoff curve. So it's like one kid and you're sick. Like 25 of the year, you have like some virus in your system, a cold, a flu, something.
Speaker 2:And then as you get to like two, three, four kids, it ramps to like fifty percent of the year, you are sick. And people will feel that if they're like, yes, I'm getting less sick. It's working.
Speaker 1:Yeah. When when the podcast, you know, optimizers that struggle to have, you know, two drinks and continue podcasting, like, they are so woefully unprepared for having children. Yeah. Fortunately, most podcasters, they don't temp tend to go down that path of of of having of having families. But It's a it's a menacing But if they did Yes.
Speaker 1:We could potentially see the strategic podcast reserve meaningfully.
Speaker 2:Anyway, leave us five stars on Apple Podcasts and Spotify. Sign up for a newsletter at tbpn.com and we will see you Have
Speaker 1:a wonderful weekend. Goodbye.