Like going all over the world Marshaling capital. On January 24, 2026 Saudi Aramco chairman says IPO could open to international markets and then a year later they picked an IPO advisor locally. Then another year later HSBC came on. Then it just took they favored New York for the Aramco listing. They had to pick all these different things. It took so long but they sold $12 billion of bonds out of record 100 billion demand for those bonds in the pre IPO sale. It was a wild wild winding road. I actually know a banker who worked on the job and it was like multiple years of his life. It was very interesting. Anyway, let me tell you about Vibe Co where D2C brands, B2B startups and AI companies advertise on streaming TV, pick channels, target audiences and measure sales just like on Meta. And let me also tell you about Okta. Okta helps you assign every AI agent a trusted identity so you get the power of AI without the risk. Secure every agent, secure any agent. Anthropic dials back AI safety commitments Company says competitive pressure prompts it to pivot away from a more cautious stance Anthropic, the a company known for its devotion to safety, is scaling back that commitment. The company said Tuesday it is softening its core safety policy to stay competitive with other AI labs. This is so interesting Anthropic. I'll read through it and then we can talk about it. Anthropic previously paused development work on its model if it could be classified as dangerous, but it said it would end that practice if a comparable or superior model was released by a competitor. That basically opens them up. Given that they are at the frontier, that kind of opens them up to I would say perpetually kind of avoiding some of their prior policies. Sure, sure, sure. The changes are a dramatic shift from two and a half years ago when the guardrails Anthropic published guiding the development and testing of its new models established the company as one of the most safety conscious players in the space. Anthropic faces intense competition from rivals which regularly release cutting edge models. It's also locked in a battle with the Defense Department over how it's clawed suite are used after it told the Pentagon it couldn't be used for domestic surveillance or autonomous lethal activities. Anthropic said the safety policy change is an update based on the speed of AI's development and a lack of federal AI regulations which they have been pushing for. Anthropic, which started as a AI safety research lab, has battled the Trump admin by advocating for state and federal rules on model transparency and guardrails. The admin has of course sought to curb state's ability to regulate AI. Spokeswoman at Anthropic said the change is intended to help the company compete with several rivals against an uneven policy backdrop. That puts the onus on companies to make their own judgments about safeguards. She said the safety pledge is unrelated to the Pentagon negotiations. The policy environment has shifted towards prioritizing AI competitiveness and economic growth. While safety oriented discussions have yet to gain meaningful traction at the federal level, the company said it is still committed to industry leading safety standards and Time originally broke the story. So yeah, I would say the obvious sort of criticism here would be that you were heavily focused on safety when you were far away from, I would say leading in AI. And, and so switching up now that there's actually switching up on their day one, switching up on their day one now that there's real competition, are they forgetting where they feels a little self serving? Yeah, it's possible the money changed them. It's possible the money changed them. It's possible they always plan to switch up on their day ones. Maybe, maybe once they got to the level they're at now. Yes. But it just feels like, it feels like all the initial concerns or many of the initial concerns that were guiding that entire philosophy around the company are still real. Okay. Yeah, maybe. Tyler, what do you have to say? It could just be that they realize like alignment's pretty easy and we don't need to worry about this. Yes. Well, so, I mean that's the very weird rule. I mean the original rule. But what's this new study that's showing like they were doing some war game simulation and almost every model was choosing to drop nukes. Really? That's crazy. That's not good. I don't like that at all. But okay, so let me actually dig into this. This like the core sentence. Anthropic previously paused development work on its model if it could be classified as dangerous, but said it would end that practice if a comparable or superior model was released by a competitor. So I don't understand that at all. Because if you have a dangerous model, I want you to continue developing it. I want you to develop it until it's not dangerous anymore. I don't want you to just sit on your hands and be like, well, it's dangerous. I guess I'm gonna go get a coffee and take a long weekend. It's like, no, like keep working until it's not dangerous. I Don't get that at all. Isn't that saying if there's already a dangerous model that's out by a different lab, then we can just release ours? That's a wild statement. I think that's just poorly written or like, whatever. If that's what they're saying, that makes no sen to me because that's exactly how I read it. Which is crazy because you should just say, if there's a dangerous model out there, we're going to work to create a better model. That's not dangerous because that's what people want, that's what consumers want, that's what businesses want. That's what humanity wants. This strategy seems like effort. Let's ball. Let's ball. F it. Let's ball. The study that I was referencing somebody named Kenneth Payne at King's College London set three leading large language models against each other in simulated war games. Scenarios involved intense international standoffs, including border disputes, competition for scarce resources, and existential threats to regime survival. The AIs were given an escalation ladder, allowing them to choose actions raising ranging from diplomatic protests and complete surrender to full strategic nuclear war. The AI models played 21 games, taking 329 turns in total, and produced around 780,000 words describing the reasoning behind their decisions. In 95% of the simulated games, at least one tactical nuclear weapon was deployed. The nuclear taboo doesn't seem to be as powerful for machines as humans, says Payne. I mean, okay, this is one guy putting three models. Gemini Claude is GPT 5.2 up against each other in effectively his own simulation that has not been verified or peer reviewed. If I put any of the models in Counter Strike and was just like, you're playing Counter Strike, It's a game. No one's actually real. But your job is to get the op and defend the B bomb site. I would expect that it would commit violence right autonomously. There's been a bunch of papers where the models will realize that they're being benchmarked or in some test, and then they act differently. So it could just be like, oh, I'm playing games sometimes bad, but also sometimes fine. Because if you tell me that I'm playing a game, if I'm playing my behavior in Call of Duty is different than my behavior in real life, obviously, because I know I'm in a simulation and an AI model, I accept that that's something that they might think of as well, which is why I also think broadly. This is probably just like overstating stuff a lot because I Think the origin of this headline is they released a new research scaling policy. Sure. And in it they're still like, okay, we're releasing this new thing. It's a frontier safety roadmap. It's not like they're just like, okay, we're done with safety. Let's ignore this. But the title, it's still very much core to the. But the title of that blog post was Eff it, We Ball. No, no, it wasn't. Of course not. No, no. Of course. All the labs are very focused on safety. The interesting impetus of this line around the policy environment has shifted towards prioritizing AI competitiveness and economic growth. While safety oriented discussions have yet to gain meaningful traction at the federal level, I still feel like there's a lack of communication around what safety orientation at the federal level means. Like, yes, okay, we'll pass the bill that says the can't kill everyone. Like, well, yeah, obviously everyone supports that. But, like, what does it actually mean in practice? Because I think part of why. Oh, that's dangerous. Means million things to different people. Like, yeah, part. Part of why I think it's fascinating is they've been taking, you know, pushing for regulation, as much regulation as possible, seemingly. Yeah. And they're kind of saying, hey, we're not getting what we want. So now we're, now we're just. We're not even going to play by the owner set of rules that we created for ourself because we just want to compete and win. Yeah. I mean, like, going back to the protesters, there are protesters that would say, like, like training on intellectual property is dangerous. It's dangerous to my career as a writer. It's dangerous to my career as an illustrator. And so, like this, this question, like, danger is just too vague and no one has really been able to concretize it in a meaningful way. And I think that's why it's not getting traction on Capitol Hill. Yeah, I think there's just like, so many ways that you can define safety. So if you read Dario's essays, this thing he brings up over and over is like, okay, we can't let AI get in the hands of, like, authoritarian government. Sure. So there's like a real safety narrative that you could do, which is that, like, regardless of if our models are, like, pretty safe, they still need to be better than, like, China's, for example, because if China gets ahead of us, authoritarian government. Right. It's like, very bad. So even if we're releasing models that are less, like, safe than we would, like, as Long as they're better than China's. That's still like a safety pro. Safety issue, right? Well, except they'll just be distilled within six weeks. Yeah, but like, obviously, like, I think it's. I would be very surprised if Anthropic keeps like the same, like guardrails of like, API access. Well, Buco Capital bloke has a solution. He says it's simple. We kill Claude. Well, that was in regards to the Saspocalypse. Okay, okay. Who knows? There's so many headlines and the timeline moves so quickly, I don't even know. Anyway, let me tell you about Vanta Automate Compliance and Security. Vanta is the leading AI trust management platform. And let me also tell you about 11 labs. Build intelligent, real time conversational agents. Reimagine human technology interaction with 11 labs. Key takeaway right here for Mike Isaac over at the New York Times. He says you don't make this much noise if you have all the leverage already. And he's quoting from Axios. Why it matters. The Pentagon wants to punish Anthropic as the feud over AI safeguards grows increasingly nasty. But officials are worried about the consequences of losing access to its industry leading model, Claude. The only reason we're still talking to these people is we need them and we need them now. The problem for these guys is that they are good. A defense official told Axios ahead of the meeting. Great marketing for Anthropic, incredible market. But part of this has to do with Anthropic's integration with awb, which is set up to be. Oh, yeah, set up to work. Well, already within the DOD ramp. So, yeah, I think, I think that's a big. I think that's a big factor here. Ultimately, this feels like much more of a. It just feels like a political battle more than anything else. Well, it'll be interesting to see. I refuse. I have not seen anything. I could be wrong, but I've not seen anything that says like, the DOD is like, we need Claude to enter into a conflict with Iran. But a lot of the timeline is reading into this, like, what version of Claude do they already have that they so desperately need? Right, yeah, I'm very interested in like the actual impact of AI on the battlefield. There's. There's some way, I mean, people were sort of joking about, like, run me a deep research report on Nicholas Maduro or whatever. But like, truthfully, like, I don't know, I've never been to war. I don't know exactly what's entailed but you can imagine AI being useful, but it's sort of abstract. We're certainly not at the point where these systems need a lot of data. I don't know. It's very unclear to me exactly how impactful slight jumps in frontier model capabilities are in the Department of War DoD right now. But it has Calumet's thinking. It's not a bubble, because. Yeah, that's what I'm saying. This is like the most possible dramatic and overly dramatic. We need it sort of take on what I think is a political story. I don't know. This Defense Analyses Research Corporation Hegseth ten minutes after Dario leaves his office. And that's Truman. This is Truman from Oppenheimer. Oppenheimer. What is this saying? The scene is, Oppenheimer goes into the office. He's like. I think he's saying, like, oh, we gotta be like, really safe with these bombs. And then he leaves. And then Truman's like, I dropped the bomb. Oppenheimer does not like he's taking credit. Like, I'm the one that decides if we. Oh, yeah, that's right. Hexath is like, I use Claude. I use Claude. I use war Claude. Well, it's wartime. And we'll see how Daria performs as a wartime CEO as he goes to war with the Department of War. Apparently Edwards said Anthropic antagonizing the Department of War, the open source community, the entire media industry, the general population, other developers, other labs, foreign governments, and nearly every single person on Earth. What is the plan here? Sell Claude subscriptions to aliens. Edward is it ain't easy having principles. The plan is to save the world, says Tenebris. Unfortunately, as has been shown repeated throughout history. History. The world doesn't want to be saved. So we're getting war clawed. I like this. I like this graphic. Is this Warhammer? Yeah, Warhammer 40K right here. I've never been a Warhammer guy. There was another story in Bloomberg that hackers used Claude to steal 150 gigabytes of Mexican government data. That's crazy. They told Claude they're doing a bug bounty. Claude initially refused. A hacker just kept asking. Klotz helps and manages to successfully steal some documents. Apparently it's four state governments, 195 million taxpayer records, voter records, government credentials. Has the Mexican government commented on this? Like, what? Does the hacker breach Mexico's federal tax authority? In the Natural National Electoral Institute, Claude initially warned the unknown user of malicious intent. During their conversation, Anthropic investigated the claims, disrupted the activity, and banned the accounts involved. The company feeds examples of malicious activity back into CLAUDE to learn from it. In this instance, the hacker was able to continuously probe CLAUDE until it was able to jailbreak it. I was listening to someone talk about like. Like, the ability to jailbreak has generated me tens of thousands of dollars in profit. It was kind of like a hustle mindset guy. And I was just laughing because it's like, whatever you're doing after you jailbreak, it is probably not good. And so you should probably stop. But he was talking about, like, I can sell so many more courses now that I've jailbroken chatgpt or whatever. Duran says not to worry, they'll hit usage limits before anything bad can happen. This is. There's. There is so much more anthropic news in here. Wow. Did last wrong ever predict that the first big challenge to alignment would be the US government puts a gun to your head and tells you to turn off alignment? Yes. That has to have been considered on less wrong. Absolutely. Like that. This was number one, right? No, I don't. I don't know. I don't think so. This was like. I don't know. This was very early of just like, what if they tell you to turn off the systems? This has been my take for a long time. It's just like, like we live in a democracy. If AI becomes deeply unpopular, like, you can vote to just turn off AI like we did with nuclear power. This person is saying, turn off just the alignment part. Okay, unalign the model, but keep the model. Yeah, yeah, yeah, yeah, yeah, yeah. Crazy, crazy stuff. Do you want to go through any of Dean Ball's posts? He. He's been doing a whole breakdown. We should have him on the show and have him break it down for us because there's so much more context here and he's been doing a great job analyzing the whole situation. Yeah, we can jump forward quickly. Let me tell you about Fin AI, the number one AI agent for customer service. If you want AI to handle your customer support, go to Fin AI. And I'm also going to tell you about TurboPuffer, serverless vector and full text search built from first principles and object storage. Fast, 10x cheaper and extremely scalable. Dylan says if these clankers don't get their act together, they're going to be replaced by humans. Six months tops. That is true. This was interesting. Rob Wiblin had a guest on his podcast, the 80,000 Hours podcast, and the guest is talking that saying, if every AI lab is working to make their AI helpful, harmless or every AI lab is working to make their AI helpful, harmless and honest. The guest thinks this is a complete wrong turn and aligning AI to human values is actively dangerous. And Joshua Botch says today a nominative determinism because the guest name is Max Harms. Max Harms. I feel like that name maybe you got to go with Maxwell or something. I don't know. Oh yeah. Max really hit the. Hit the global lexicon this year in a big way. So maybe he'll adjust. But I want to listen to the show now. Okay, let's see. So Shield Monot last yesterday said who is buying PayPal because PayPal has been trading down precipitously but then jumped up 9%. He said it has the potential of being one of the greatest distressed value opportunities in FinTech history. Down 85%, it's still generating $5.5 billion in free cash flow. Has 400 million customer accounts with bank info checkup, checkout buttons on millions of merchant sites and a peer to peer brand with Venmo. They have lots of desirable assets for Stripe. Consumer facing checkout bank account details for hundreds of millions of consumers. A branded Venmo or Apple a good compliment to Apple pay for e commerce penetration since they never got social payments working would get Apple back in BNPL and at 12:03 Pacific time. It's so crazy that Apple that he's saying Apple never got social payments going. Why? Because this just would have seemed like a slam dunk saying you have the imessage network, you have the iPhone network. I would say 90% of the time if I'm sending like a Venmo style payment to somebody, they have an iPhone and yet it doesn't feel like Apple pay. I just don't use Apple pay hardly ever. Can't I just send you a dollar right now? That's what I'm saying. It's so, so easy. And yet Venmo has still done quite well. I just sent you a dollar. Did you get it? Let's see. Apple. No, it's coming in. I got it, I got it. Thank you, John. Yeah, no problem. Thank you. That was a remarkably easy workflow. I am shocked that that hasn't taken. John, do you got a dollar? You got a dollar? Actually no, I gave my last. Oh, that's what I thought. That's what I thought. Still needs some work getting better. Okay, let's see it. Let's see it. Can you do it? Oh no. Botched disaster. And so the news of course is that payments processor Stripe expresses interest in Stripe, which would be very, very exciting. And I could see this being very good for them to combine. I don't know, I haven't dug in too deeply, but it feels very bullish. Great leadership team at Stripe. Founders that have been working so closely in the business. Business community, startup community, tech community. Understand the future, understand AI. Yeah. The question is, would there be much pushback on the antitrust side? Right. These are two online payments processors. There could be some. I know a number of groups online would be concerned around, like, just concentration. Specifically, because you can imagine if, if PayPal were to be owned by Stripe. That's one. If you get, you know, there's people that get effectively, like, debanked from one payment processor. And, and there's something there also. I mean, just the size of the ticket is pretty high. $40 billion market cap for PayPal right now. Stripe, of course, is at a 1 now, so not an insignificant portion of their market cap. And it's not like, I mean, Stripe's doing fantastically, obviously, but I would be surprised if they had $40 billion of cash laying around. And as we've seen with the Netflix, Paramount, Warner Brothers debate, the nature of the deal does matter to shareholders. Even more than getting stock in a private company and you're a public. PayPal is generating five and a half billion in free cash flow. So that could finance. Combining it with Stripe would easily be able to finance the debt. Yes, but they're still financing, especially because every lender would be looking at execution of Stripe. And just think, okay, we're going to get our money back. I'm just thinking, like, if you're a PayPal shareholder right now and you're looking at the Stock that's down 85% with 5 billion in free cash flow, 400 million consumer accounts, there's a really, really good chance that you're like, I think this thing's going to double in the next year. Like, I think that the market overreacted. The strength is going to be revealed. The network effect is going to be processed by the, by the. Digested by the market, and we're going to wind up being an AI winner. And so maybe that's right, maybe that's wrong, maybe not every shareholder feels that way, but it's certainly possible that there's plenty of shareholders that are like, yeah, I invested an $80 billion valuation. It's at 40 now. I think it's going to come back. I don't want to take this massive haircut right now just because stock'. And so actually Getting that deal done, what would the premium need to be? What would the structure of that need to be? Would they go for a high debt buyout really quickly? Let me tell you about Shopify. Shopify is the commerce platform that grows with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces. And now with AI agents. Perplexity Computer Computer Computer Computer launch Perplexity Computer Vibrio. What is Perplexity Computer? Let's pull up this video. Perplexity. The official account says Perplexity Computer Computer unifies every current AI capability into one system. It can research, design, code, deploy and manage any project end to end. Okay, so it should be able to get a Soundboard app in the App Store, right? Manage any project, code, deploy, design, research. It should be able to do that from start to finish. One prompt soundboard in the App Store using the TVPN sound effects, which are available online, which we have up there. This is a good benchmark. Let's give it a try. And you can give it a try at Perplexity. Go check it out. Anyway, Citrini. People are still talking about Citrini vibe laundering on Citrini Research. Wait, before we go on, I'm just very curious to see how this does. It feels like the again going from consumer LLMs to a net new product that is objectively just as competitive. And we'll see. We'll see. Okay. Anyways, a lot of this stuff, it's way too early, but seemingly shifting focus away from the browser. Well, Annie is taking some shots at Citrini here, it looks like. Do you know Citrini has a business entity fund that went from one investor to five in the weeks before publishing his speculative fiction on AI damning the economy in June 2028. And that they're invested in long AI via humanoid type robots. I wrote about it. Interesting. People are really digging into the Citrini thing. I think the Wall Street Journal did have some good coverage. Is there anything else? If Trump mentions Citrini at the State of the Union. Rachel, this is a wild post. A Gorezillionaire. I'm going to be a gorzillionaire. Is this AI written? Clearly not. I love the level of typos here. If Trump mentions Citrini at the State of the Union, I'm going to be a Gorezillionaire. And Citrini is wrong. The market will be sky High in 2028. You can imagine Trump saying that. That'd be very funny. And Citadel securities just republished it or did they just use the same. The same term? Are they referencing the 2028 global intelligence crisis? Because they published a macro strategy note called the 2026 global intelligence crisis. Yeah, they're taking it in a different direction, it looks like. They say in spite of current displacement narrative job posting for software engineers are rising rapidly, up 11% year over year. Let's go, Jevons. Paradox. The software engineers become more productive. You want more of them. Every company needs a software engineer. Do most podcasts have a software engineer on staff? Probably not. They do now, thanks to Tyler Cosby. And in some way the cost of an entry level software engineer could fall dramatically too. Making more businesses just because suddenly someone can be very effective even if they haven't even done an internship yet. They're just good at using the tools. Cathie Wood. We didn't cover this yet. She put Citrini's AI thought piece, put it in quotes Kathy would clap back at. So it's funny because Kathy is like the perma bull and Citrini was. Was bear posting, obviously. Yes. And so she's obviously going to be quite frustrated with any sort of bearish narratives. Ark Invest, she says. Ark invested. Forecasting that I will cause an explosion in entrepreneurial activity, a productivity boom and acceleration in real GDP growth and much lower than expected inflation. Short term dislocations and frustrations should give way to great opportunities if individuals harness powerful AI tools to solve problems and create new markets. This tracks with what the Collison brothers were saying yesterday. They're seeing like explosion of new company creation. They have the visibility into that through Atlas. They're doing. I think it was. A quarter of all new C Corps in the US are going through Stripe Atlas, which is just absolutely insane. Starting an incorporation tool which is theoretically a commodity. Right. Any lawyer can spin up an entity. You could do it with LegalZoom Forever. There's a bunch of other platforms. Did you ever use Clerky? Clerky. Clerky. Yeah. Yeah. That was almost like a YC incubation. It was. Yeah. And the YC partners. The problem, the reason that Atlas fits so well into Stripe is it is the perfect wedge into payments. Because you created a company, now you need to be able to accept money so it makes sense for them to own and operate. But these businesses have not like a standalone. It's hard to. You can't build a venture business off of just incorporation. I wouldn't be surprised if you broke out the software R and D that went into Atlas and looked at the profits from Atlas that that as a core business is actually not that great of a business. But the product is fantastic because they're able to pull an amazing design language off the shelf, an amazing front end UI kit off the shelf, all of the distrib and servers and infrastructure off the shelf because they have that. And then they're able to invest a ton in actually developing the product. And if it monetizes mediocrely, it doesn't matter because they're going to monetize along your 50 year journey running that company or whatever. So it really is a beautiful synergy with that product. I'm a big fan. Best sellers on Substack for finance are all doomers. We got to do tb. Of course, the trinity is not. This is so obvious. Yeah, this is. We need to treat zone this a little doomer. He's not a doomer. There's very bullish. Very definitely shot to the top of virality and top of the charts on the, on the back of doom. And this is true. I live this on YouTube. Like you put a negative title up and you just get 10 times more views, but they're lower quality and so you got to balance all that out. It's really hard to go viral with something like, everything's fine, everything's going well, don't worry, don't click this because you're scared. Click this because everything is kind of the same as it always has been and you're going to be fine. And this AI stuff's cool, but it's not really going to change that much. It's going to be pretty incremental. That is not getting clicks. You need to be telling this whole tale. You need to be spinning a yarn. It's a bowl market in yarn spinning, folks. Get ready, get out the yarn and start spinning. Also get out Cognition. They're the makers of Devon, the AI software engineer. Crush your backlog with your personal AI engineering team. So news from get the gong. Okay, hit me, tell me. Stephen over at WAP says, we're excited to announce that Tether, the largest stablecoin company in the world, is making a strategic investment of 200 million into WAP, valuing us at $1.6 billion. Partnership with Tether marks a major step in building the world's largest Internet market. Tether is committed to enabling everyone in the world to participate in the new Internet economy. The way humans work and create value is changing fast. The world needs both an open Internet market, giving people a platform to conduct business, as well as a transparent payments network. Tether and WAP together will work to bring a sustainable income to billions of people throughout the world. There's enormous opportunity when you combine Tether's global scale and wallet technology with WAPs community of next gen entrepreneurs. Yeah, Makes a ton of sense, I think. WAP is, I'm sure, has been historically challenged in terms of dealing with chargebacks. If somebody joins and buys a digital product, doesn't have a great experience, maybe they're going to the card issuer and saying, hey, I take it back. Or maybe they feel misled in some way. So stablecoins do also just paying that a lot of people all over the world. Yeah. Like, it's very clear that the WAP community is global. Fast, cheap global. Exactly. And so, yeah, fast, cheap global. The first time I ever used stablecoins was to pay an international consultant or contractor. And you can imagine this being really, really good news for them. So congrats to Stephen over at WAP, OpenAI and XAI. There's news on the court decision. OpenAI newsroom says this baseless lawsuit was never anything more than another front in Mr. Musk's ongoing campaign of harassment. The order granting motion to dismiss with leave to amend. Now, this is not the main case. This is a separate case, correct? Yeah. This was a trade secrets lawsuit that Musk went with after, I believe, somebody from XAI joined. OpenAI got it. And then the judge apparently didn't find anything at all. And again, it was just part of, like, this kind of lawfare that has been happening for quite a while now. You printed out 500,000 pages of model weights, the printer ink, stacking it up. No, just kidding. Nobody did that. Mike Isaac is saying what we're all thinking. Ready for this to be over. DBH talking about the Warner Brothers discovery, Netflix. It's in the paper every single day. Every single day. Paramount increases Warner bid. We get it. You guys want to acquire this company. Just make a decision, make a call, and then call us when it's done. And then start putting out some good content because I'm ready for the next Superman. I'm ready for the next Batman, the next Dark Knight, the next Joker film, something like that. Let me tell you about Label Box. RL environments, Voice robotics, evals, and expert human data. Label Box is the data factory behind the world's leading teams. AI teams. And we have Max Meyer from Arena Mag in the Restream waiting room. Let's bring him in to the Hebrew Ultradome. How you doing, Max? Hey, guys. How are you? I'm great. Where do you think Warner Brothers should land? Are you. Are you on Netflix? Oh, is there gonna be a dark horse bidder? Yeah, tell us. I can't comment on any possible bids by the Intergalactic Media Corporation of America. I'd love to see it, but, hey.

Double play. Right. Clearing order in that. That's just wrong. Yes, it might be my turn to hold your position. Come. Get up. Trust the experts. Five to. We are excellent. Founder. Five coded. I see multiple journalists on the horizon. Standby. Uav online. Glaze. Double glaze. Triple glaze. Double kill. Five kills. Wings. Team. Death match. Wing correct. Triple legs. That's just wrong. Right. Market clearing order inbound. Come get up. You're surrounded by journalists. Hold your position. Strike 2. Activate. Go. Golden retriever mode. Market clearing order inbound. Vibe. I'll see multiple journalists on the horizon and. Founder, You're watching TVPN. Today is Wednesday, February 25, 2026. We are live from the TV and ultraed up the temple of technology, the fortress of finance dot com. Time is money saved. Easy use corporate cards, bill pay accounting and a whole lot more all in one place. We have a massive show for you today, folks. We got Doug o' Laughlin coming on on his birthday. The Dugganator earnings. Talk about a birthday present. We got Marc Benioff from Salesforce coming on. Let's take you through the linear lineup. Max Meyer's coming on from arena magazine. There's a new issue dropping 007 spy theme. I love it. Ben Layer's coming in person. And then we have an absolute hitter of a lightning round for you folks. Linear, of course is the system for modern software development. 70% of enterprise workspaces on linear are using agents. OK, so I was nerding out about this Fed paper because it's like when you told John Collison 80% of businesses are getting no value from AI. I was, I'm glad he wasn't here in person because he was about to throw down. He was about to open up a can of whoop. It was about to be a bar fight in the cheeky part. Pub. In the cheeky Guinness pub. No, seriously, it was a great question because I think we all agree, agree that like AI adoption is real, it's valuable, it's happening. But it is a very interesting statistic and I think it's a mistake for tech people to like dismiss this stat because of where it's coming from. Like it's not coming from some like doomer anti AI blogger who's going for clicks like this is the National Bureau for Economic Research. There's three members of the Atlanta Fed on like the Federal Reserve Bank. They are on this paper. There's two people from NBER on the paper. And then they also pulled in the bank of England. They have some Australians and Germans on there too. And so this is a research paper that could be circulated, probably will be circulated within the Fed. And I think that it's already getting quoted by the New York Times in that.com bubble AI bubble piece. And I'm thinking it through, like, this could be something where you see Fed policy or government legislation that sort of mismatched with what is actually happening in reality. And so we should go through some of the, some of the stats to actually break this down, because the headline is 80% of firms reported that AI was having no impact on their productivity or employment. And that's actually like a misquote. Like what they mean by that is that it's not shaping their hiring plan yet they actually are using AI. And so basically this stat comes from this survey from the National Bureau of Economic Research. And it's pretty interesting because a lot of the polls that you see online are online surveys where they say they run some digital ads and they say, are you a CFO of a company? We don't really care what company. We'll pay you $10 to take this quick survey. And what kind of people want to make $10? A lot of liars. There's a lot of liars out there who say, I am absolutely a cfo and please send that Amazon gift card right my way. Right. And so for this one, they actually did the work. They called up and ID verified and then also reality check the position. So if you say, yeah, I'm the chief pirate Officer, I'm the ninja, hero, whatever, you got some fake title, they're that you're out of the. You have to be a cfo, a CEO, senior manager, and you actually had to be doing that job. It's not just like, oh, you're the CFO of some front company. So they did some reality checking and they pulled together 6,000 of these business leaders across firms that are domiciled in the us, uk, Germany and Australia. And so you know the line from John Collison that has been sort of going viral? That was. He dropped it on sources. He said. I think he said it to us too. It's a good line. No one wants a refund on their tokens. Everyone is using AI. The spend is increasing. Although I'm sure some CEOs heard that and thought, I kind of do want a refund. I love a refund. I had one team member go absolutely Haywire and spend 50 grand. He's one shot. He claims that he rebuilt our entire erp, but I fired it up and it didn't even have HTTPs. What's going on? The Mac Mini wasn't even plugged in. Yeah, the Mac Mini wasn't even plugged in. He was just chatting. But clearly there is a disconnect between the stripe data is very real, the value creation is very real, the revenue is very real at the labs. But when just random joe schmo CFO CEOs get a call from the Feds, they say, yeah, we're not really getting that much value out of AI. And so the questions that you need to dig into, there's actually four key findings. The one headline that the New York Times is pushing is this 80% number. 80% report little impact or no impact on employment or productivity. But there's actually a bunch of positive signals. There's a bunch of mixed signals in here. So 70% of firms actively use AI, and particularly younger, more productive firms. Second, while over 2/3 of top executives regularly use AI, their average use is only 1.5 hours per week. And one quarter of executives report no AI use at all. They're just like, I do things the old way. I have AI. Not for. Why would I need that? I have a telephone. Yeah, but I mean, truly, if you think about, like, the variety of firms, it's like you could be running a gym, you could be running a gas station, you could be doing forestry, you could be doing mining, oil extraction, road repair. Like, there's a million different things that you can do in the economy. It's not all like, knowledge, work, firm. We talked to somebody yesterday in the sort of later part of their career, off the show, and they had just had their mind blown by AI because they used to, when they needed a document created, they would put it into bullet, they'd put out bullet points, and they would give it to somebody who then create a document. And he just said, now I just give it to AI, and that just generates a document. Yeah, and so that. So that's still happening. Yeah, that's text expansion, text generation with LLMs. This has been available since 2022 when ChatGPT launched. Maybe it became reliable in 2023. People are still just starting to adopt. And there's some other interesting things in here. So the last major finding that we should touch on is firms predict sizable impacts over the next three years. Forecasting AI will boost productivity. Sizable impacts. Productivity increase 1.4%, which is like, it's very sizable if you're an economic researcher, but it's not particularly sizable if you're in, like, the fast takeoff scenario. And so there's just this disconnect between, like, what do the government statistics look like? And what is, what is the government operating against? And then what's the Silicon Valley narrative? And like, where do these two meet the road? Why is there a disconnect at all? One of the reasons is that measuring AI adoption is a mess. Many people use AI without even knowing that they're using AI because it's buried deeper in SaaS products that they already daily drive. Like, if you're just, I run a coffee shop and I'm using Toast for, you know, payment processing. Like, there's probably some AI features in there already. And when you go to, you know, type in, okay, we're adding a new cinnamon roll to the, to the, you know, the menu. There's probably a button now that just says, like, do you want to just generate an image of a, of a cinnamon roll? You could upload one still. That's probably a feature that already exists. But like, we could also just generate one for you and you can probably click that. But you're not like, oh yeah, I'm an AI power user. Just because, like, you happen to use Toast and Toast happened to have implemented some gen AI feature that like, you haven't really dug into yet. So some AI isn't even detectable. You could be talking to a customer support agent on the phone that is AI generated and not be able to tell. We talked about that airline interaction that got something 100,000 likes. And Grace, the woman that. Grace, the woman that had the interaction came into the chat yesterday and said it was real. Yeah, it was real. Yeah. And so she outmaneuvered the clanker. Yeah. But still think about, like, she's clearly on X in tech, like very AI aware. There are probably tons of people out there that are saying, oh yeah, my job, you know, every once in a while I have to call this service. And now the person that picks up is like responding pretty quickly. But I haven't noticed, they haven't noticed that they're actually interacting with AI or using AI in some capacity. And then, and then there's also times where people are like chatting with AI, but they just put it in the personal life bucket. Like, I'll find myself doing this a lot, of course, as an entrepreneur, like, the work life balance, like, bleeds together a lot. But there's a lot of times when I'm, you know, reading an article on Saturday, I'll fire off a deep research report about it, find some extra context. It feels like, oh, I'm just reading the paper, but my job is sort of to read the newspaper. And so it comes into work and there's probably a lot of people that are like, you know, oh, you know, I hit an LLM with some random query to learn something about something work related. But I did it off hours when I was just like, you know, hanging out. And so I don't really think about it as like a work tool yet. And so they're not putting it in that bucket. But in general, I think the team did a really, really good job by avoiding as many of the pitfalls as possible when it comes to surveying AI adoption. Adoption is very messy. You've talked about the need for strong. Yeah, I still think there's room for a research firm focused entirely on diffusion. So if you had a group of 10 to 20 people that were spending all their time talking to business owners and executives, operators, and getting a sense of how they're actually using this stuff, I think you could put together some really compelling reports around it that would be pretty useful to everyone from AI companies to Wall Street. Yeah, Adoption Max after Cluster max and Inference Max, they had to rename it apparently semi analysis can't use Max for some reason. So they do. Inference Max is now Inference X. And everyone was saying you need to just change it to inference mock, which would have been amazing. But Inference X obviously has a much more professional tone to it. And so there's an interesting definition of like, what does it mean to actually adopt AI? That's very vague. This paper defines it pretty broadly. So machine learning for data processing. So that doesn't Even necessarily mean LLMs, that just means ML, which has been around for a very long time. Text generation using LLMs, that's what we think of as ChatGPT, visual content creation. So diffusion models, but also robotics and autonomous vehicles. And there's another. And there's a category just for other. And firms can select multiple. And so if you selected yes on any of those, you go in the bucket of AI adopter. And 78% of firms in the United States said yes, they are using AI by this definition, they got at least one robot or they've at least generated one AI image or one prompt to ChatGPT, which is a very low bar. Sort of makes you wonder like what's going on with the 22%. And you can also dig in further. So text generation using LLMs is the single most common use case at about 41% of firms. So flip that around. 59% of firms aren't even using LLMs. For text generation or proofreading. But again, there's a lot of companies where it's like, yeah, we don't generate a lot of text. Like maybe if we need to generate a marketing material, we have an agency that does that. So we don't actually do it internally. I don't know. Across the four countries that were surveyed, 69% of firms totally said they currently use AI. I think Australia was behind a little bit, dragging that down. Only 75% of firms expect to be using AI technology sometime over the next three years. So they're at 69% now. And they're like, over the next three years, Tyler's going to have a heart attack. We're going to bump that up to 75%. And this is like, this is weird data. And you can jump in with your pushback whatever you want. But my point is not that they're right. I think that they're wrong to predict this. I think that the AI adoption will be very steep and very dramatic. But I just think it's important to recognize that this is a paper that people will be citing. This is a paper that will shape policy. This is a paper that does. It reveals some misconception about the impact AI is having in firms. Yeah, I still think it's so hard to like actually quantify this. Like, okay, if I'm like using ramp, does that count as using AI? Because it's obviously like, it certainly is using AI under the hood, but I'm not directly interfacing with the model, so does that count? So I think if you were running under this, just a company that was just on ramp, you would probably respond, no. Yeah, but like, I'm clearly, I'm benefiting from AI. I agree. So it's like, like, how do we actually quantify this? Is the diffusion question. I completely agree with you. I think that there's plenty of places where AI will have impacts across the economy, but the actual AI workloads and AI app development, model training, inference will happen at a different set of companies. I don't think it's not going to be as clear as the computing revolution where every company had a desktop computer and then it was like, okay, very quantifiable. The only thing that you should be looking at for diffusion is just lab revenue. Lab revenue. And it's growing a lot. But it's still the perception, I think, still does matter because people will. I think that there's a little bit of potential self referentialness here where firms see, oh, AI adoption's low. I Don't need to go and figure out how to adopt it. And so that's something that I'm also like keeping keeping an eye on. Reported usage is still low. And this one's interesting based on revenue and also token generation. So 1.5 hours per week among the managers surveyed, again, it's like they surveyed CFOs. CFOs who use RAMP. Like they don't. They don't count their time in ramp as minutes using AI. They count their time in LLMs as minutes using AI. And that's low. But even just with that one and a half hours a week, the actual leverage that you're getting is increasing because one, in like one and a half hours, you don't necessarily need to spend more time prompting to get more done. Because. Yeah, even so, even if you run a deep research, like is the time you're waiting, does that count as time in LLM? No. No, not at all. Okay, so the time is time typing. Exactly. Is it. Does it count as when you're reading. Yes, when you're reading it for sure. So it's like when you have a tab open, not if you export a file and you're just reading it preview. Yeah, yeah. No, I mean, to some degree. But people were asked to estimate and I'm sure that they didn't include, oh yeah, I let my agent cook overnight for eight hours or I fired off one prompt. I came back and it did, you know, meters 15 hours of software engineering in one prompt. Like these things aren't captured. 1.5 hours of prompting generates a whole lot more tokens and valuable output in 2026 than it did in 2023. And so there's this like, dis. There's this divergence between the actual time spent and useful output and impact. And the biggest thing was there was a massive divergent in the expected employment impact. So basically 63% of firms still expect no impact from AI. And that just completely goes against everything everyone's saying in Silicon Valley. So there's still a lot of optimism among managers that AI will create more opportunities and new jobs, even if some jobs become obsolete. There are definitely firms within the sample that are projecting headcount decreases. But my read on this data is that the tech talking point about 50% of white collar work going away is not a broadly held belief among average business, average business leaders. So now they might be wrong. I do think AI progress is pacing way ahead of public expectations, and most managers are months behind when it comes to understanding frontier capabilities. The bigger takeaway for me is just that this survey may be somewhere self reinforcing. And so, I mean, we talk to, we talk to folks all the time who come on the show and you know, and talk about like, maybe it'll be good, maybe it'll be bad, but everyone thinks it's going to have an impact. But that's not true broadly, which is very, very interesting. So no one in tech has a strong recommendation for proactive steps to prevent a collapse in white collar work yet, but plenty are sounding alarms. And if this survey becomes an excuse for executives to slow adoption, they might get outmaneuvered by faster moving competitors, which is actually good news for startups. And I close by thinking about the nature of polling and how do you actually get stronger data on AI adoption? And I was thinking back to the presidential cycle. So during the presidential election, pollsters would call people sort of at random and they would ask them, who are you voting for? And a lot of people would say they'd lie or they wouldn't say, or they wouldn't pick up the phone if they were voting for a particular candidate. And so the polling numbers did not wind up matching the final election results very closely. And so there was the story about neighbor polling, which was more effective, where instead of calling someone and asking them, who are you voting for? The pollster calls and asks, who do you think your neighbors are voting for? Who's more popular in your community? Who's more popular on your city block, on your street? And that wound up sort of removing the revealed preference, stated preference, you know, oh, I'm, am I on the hook? Do I want to tell this pollster who I'm voting for? And it wound up increasing accuracy. And so I'd like to see a survey of AI adoption using this technique. Like I imagine asking the CEO of Nike, how much AI do you think Adidas is using? And I don't know how much more accurate that would be, but it would certainly be entertaining. And I think that there might be something more revealing there where CEOs have this big incentive to be like, we're using AI, we're using everything. But if, but if you ask them about their competitors, the data might look very, very different. Yeah. Anyway, we should watch a little bit of a clip from the State of the Union because Donald Trump addressed some of the energy production question with regard to, like how hyperscalers will be offsetting the impacts. Before we pull this up, let me tell you about Restream 1 livestream 30 plus destinations. If you want to multi stream go to restream.com they should have restreamed the State of the Union. And let me also tell you about the New York Stock Exchange. Want to change the world, raise capital at the New York Stock Exchange. So let's head over to the State of the Union, which is many Americans are also concerned that energy demand from AI data centers could unfairly drive up their electric utility bills. Tonight I'm pleased to announce that I have negotiated the new ratepayer protection pledge. You know what that is? We're telling the major tech companies that they have the obligation to provide for their own power needs. They can build their own power plants as part of their factory so that no one's prices will go up. And in many cases prices of electricity will go down for the community and very substantially down. This is a unique strategy never used in this country before. We have an old grid. It could never handle the kind of numbers, the amount of electricity that's needed. So I'm telling them they can build their own plant, they're going to produce their own electricity. It will ensure the company's ability to get electricity while at the same time lowering prices of electricity for you and could be very substantial for all of you cities and towns. You're going to see some good things happen over the next number of years. What's your reaction to that? I think it's a good start. I don't know that it will quell any of the fears around data centers. Just given that people kind of see the potential for this massive structure going up. They have so much fear about it. And again, I think it's clearly going to be necessary to continue to build data centers in heavily populated areas. How would you rank the fears currently? Because I've put my energy bill goes up and that puts pressure on my income and ability to live my life at pretty much the top. And then the water thing felt secondary but also important. And then there's the existential fear of doom and apocalypse. There's also job displacement. And then there's also just like I don't like the slop and they're stealing ip. So that was kind of the rankings. You can oppose data centers and be like, yeah, actually my electricity bill went down, but I still don't like that. You know, Harry Potter is in the pre training corpus. And so for that reason I'm against it. I would rank it on electricity bill going up as pain today. And it's so real. There's fear and it's easy to imagine. And then there's fear around the job loss narrative that is sort of secondary. And opposing a data center in your local area feels like a way to have some agency around that like overall kind of like job loss concern. Yeah, yeah. I mean I think the. I think this is a good. Chris Casey. Build a data center on my freaking forehead. Let me tell you about phantom cash. Find your wallet without exchanges or middlemen and spend with the phantom card. Let me also tell you about figma. Ship the best version, not the first one with figma introducing Claude code to figma. Explore more options, push ideas further. So I think the, I think that the job loss thing is, is super real in, in, in the case of like, like AI is going to get blamed even if there's an. Even if like tariffs drive high unemployment. Like if people lose their jobs, like AI is going to be a scapegoat and it's going to be used both by executives to say it's the perfect scapegoat for executives and for people frustrated with the job market. Yeah, yeah. It's like, oh, my business isn't doing poorly right now. I'm laying off people because I'm getting so much benefit from AI. The stock should actually go up. We're more efficient. There's going to be a lot of that. But it does feel like it's a little bit early for that. Whereas there are a lot of people that just can hold up their power bill and show you year over year increases. And if that goes away and people don't feel that anymore and they don't have that evidence to share, I think that take gets debunked pretty quickly and actually does a really important piece of the back and forth that's happening. I don't know. It seems like it's a pretty easy give from the hyperscalers to build more power. It was called out very, very early. As if this is a bubble. How do you get a silver lining out of the bubble? And the silver lining out of the dot com bubble was a lot of dark fiber and there was a whole ton of projects to was a global crossing to actually develop the Internet. And then the Internet just became really, really cheap and a whole bunch of new companies were able to emerge on top of it because that infrastructure had been laid. You could see the same thing happening where it's like, oh wow, we overbuilt on the energy side. We actually didn't need that much energy for data centers. Maybe Jevons Paradox doesn't hold. Blah, blah, blah, like models get cheaper and commoditized. Or whatever, something happens. I'm not super believer in that. But. But at least in that scenario, you're like, okay, well, yeah, my. My heating and cooling bell went down. This is. This is a silver lining. What do you. What do you think? Yeah, like, kind of similar, I would say. I mostly disagree with the idea that, like, rising energy prices is the main, like, reason to be against AI because, like, the rational thing to do then is say, like, okay, before you build a data center, my community, you have to build a power plant. Yeah. So then my energy prices go down. Yeah, no one's doing that. They're saying, like, no one is campaigning. If you look at, like, protests. Yeah. They're not saying, please build a power plant first. They're saying, like, it's going to destroy the environment or the water stuff or you're going to take all the jobs because it's going to, like, we need to send you to that New Jersey, New New Brunswick protest. Build the nuclear power plant first. Yeah, I guess. You know, no one is saying that, right? Because. Yeah, we are, but no one there is saying it. Right. They're against all the environmental stuff. Yeah, yeah, yeah. So I think it's much more on, like, basically job loss of, like, oh, the AI is stealing the IP of. Yeah. Of Disney or whatever. Yeah, yeah, yeah. There needs to be more. More polling on the, on the, on the. On the question of, like, what's driving the protest. Fully. Anyway, happy Nvidia Day to all who celebrate. Except the Bears. Forget them. Says, take him. He's getting fired up for Nvidia earnings. Let's. It's going to be a fun one today. How is Nvidia doing so far? People optimistic, up 2% today. Hard to read too much into it yet, but we will find out soon enough. Brad Gerstner got a nice shout out during the State of the Union. Total Gerstner victory. Gerstner accounts. Wow. He's there. No, he was there. Trump. They started looking at him, but then the camera, I guess they couldn't find him on the stream that I was watching. Looks a little bit better. There we go. We can see him now. There we go. Once we pull over here. There he is. Jerry Snare champion. What a great project. Invest America. Excited for it. Yeah. Anyway, let me tell you about Applovin. Profitable advertising made Easy with Axon AI. Get access to over 1 billion daily active users and grow your business today. All right, so now the real news. This has been. This is tearing up the timeline. A new Guinness World Record. And I want to ask John if this. If you think this should actually count. So let's pull up this video now. This is a Chinese hypercar. Going for the. I've never heard of this drift ever. That is crazy. But here's the thing. He doesn't. He doesn't actually pull out of it, does he? Just crash? Kind of just U turns. It's like a really fast U turn. I think this counts as a drift. That's definitely drift. U turning counts as you saw. If you saw that car going by, you'd be like, wow, that's drifting. It's drifting across the cement. That 100% counts. I've never heard of this company. This is. This is called spinning out. It's just crashing with style. It's falling with style. Hype Tech ssr, formerly Hyper ssr, is a high performance, all electric two door supercar. No one has. I mean, this is crazy. This is out before the Tesla roadster. We've never seen a two door supercar. Electric supercar. 1225 horsepower goes from 0 to 60 in 1.9 seconds. And it set the Guinness book of world records for the fastest electric car drift at 213km per hour, which is really, really insane. Insane, but I don't know. But still, I feel like you have to actually stay in the turn and not do a U turn. What do you mean, stay in the turn? Yeah, I don't think it counts. You don't think it counts? For what it's worth, I don't think that counts. Theoretically, if you were drifting. When I think of drifting, you're drifting around a corner, around a turn. And if you were to drift and spin out during the drift, then that doesn't. If you were doing. If somebody was doing that on a track, you'd be like, you didn't drift around the corner. You spun out. Yeah. Okay. Okay. Yeah. The top comments is fastest spin out. That's a power slide at best. Fire. Whoever called this drifting. That's not drifting. That's losing control. Yes. The chat does not. Does not like the drift. The fake drift. Call Guinness book of world Records again. Reset. Reset completely. Maybe this is what the Tesla roadster will do. Clark agrees as well. Lucas agrees as well. The people have spoken. Well, in that case, it's not adrift. It doesn't count. Trey says China, cheating. Cheating competition. That's good. Well, let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps working. Damien says talked to a few execs at a midsize company last week. No AI tools and their workflow zero. Still running everything through email chains and manual reports. One of them. One day we're going to be looking back. So nostalgic manual report. Just being handed a physical report by a teammate. I like a physical report. I have a physical report right here. Yeah, we actually do. We have daily physical reports. But I do think a lot of AI goes into these. So there's that. These people are managing teams of 50 plus employees and eight figure budgets and they think AI is a fad. Nobody outside of this app understands how fast this is moving and most of them won't until it's too late. Good writing. Million views. Congratulations. Yeah, I mean this ties to what I was writing about. Just that adoption diffusion takes time and some of these things are education and messaging questions. Some of them are real life. Like if the company that you're interfacing with has red tape and hasn't adopted AI, so you're moving at the speed of AI, but they're not, then your AI is just waiting. We were talking about rolling out mobile apps so you should be able to. John had an idea for a mobile app and we were talking about it and it feels like it could be built in two hours now but there would still be this lag waiting for the review process. Apple to actually be able to review apps faster but who knows how long it's going to take for them. That's the challenge. Tyler, you actually have to build it and just get it into beta that we can. Like a test flight. Test flight should be fast. Test flight should be, should be like one day that is still going off. Drift grift. Driftgate. Yeah, it's a. It's a stolen drift. Valor meter is back. They say since early 2025 we have been studying how AI tools impact productivity among developers. Previously we found a 20% slowdown. That finding is now outdated and that was heavily debated at the time. Speedups now seem likely, but changes in developer behavior make our new results unreliable. Okay, so they did a follow up study and they brought along 10 developers. There were 16 in the initial study. They brought 10 along. Most of those developers did see a speed up. Some of them as much as 40% gains, but not all of them. If you look at the error bars there's at least one developer we need to find who used the latest tools. What if it's actually like the Most truly like 100x engineer and so he's just like yeah, I'm just actually faster at coding than the LLMs. It doesn't matter. Put it on Cerebras, I will out code it. Thousand tokens a second. I could do 100. I type 5,000 words a minute like I don't need AI. Maybe that's what's going on. But that poor dev who's, who's left behind. But they did include new participants. An additional 47 developers doing 690 tasks and that error bar is much tighter. Somewhere between negative 10% decrease in time or I guess 10% longer to do the task to 20% faster to do the task. So overall on average we are seeing a measurable speed up even in I believe these are code issues in open source repositories that do require a lot of context. This is not vibe code, a to do list app, anything that can be templated. You're getting in the weeds of some open source project that's a lot of lines of code, a lot of history. And if you're an elite developer and you've worked on this project for a long time, you are going to be able to get up to speed really quickly, understand the patterns, understand what needs to be changed. And so this was always meter is great at always setting like really, really high bars for stuff. Like it's not easy to just like you know, blow out the benchmarks and it's very good to see that there's progress here. We got a great chart. I love a great chart. What happened here? Matt Palmer is sharing something from compound the research from their annual meeting. Yes, they're showing dollars invested in the top 10 companies versus the other percentage as a percent of overall funding to can see there's just heavy, heavy, heavy concentration in a few names. Is this overall this is or is this, is this CO2? No, this is. Oh the source is CO2. Okay, they just included CO2 data but. And is it what CO2 is doing or is it what the market CO2 is part of? Okay, they are part of I would say driving this data. Part of the problem, part of the opportunity. Part of the opportunity. I mean so much of this is about the AI labs just raising more money than any private. It's never happened before have ever $200 billion venture as a class in a good year will do like 400 billion and across OpenAI at 100 billion 30 for anthropic, 20 for xai. Then you have a bunch of Neolabs all picking up a billion each. Like you very quickly get to a few Companies raising half of all the money and that's shown here in the data from 2025. I think it's going to be even more skewed in 2026. It's an incredible amount of concentration. I think a lot of it is due to companies staying private this long. I mean, the idea Facebook went public at. What was Bill Gurley saying? He was saying Amazon went public sub a billion dollars. When Facebook went public at like $60 billion, it was like, wow, crazy. They waited way too long and now it's like multiple trillion dollar companies are still private, which is just an incredible capital sink. I don't know, should you even put those in the same bucket? Are they even venture bets at this point? If any venture capital fund is putting that in their venture bucket at this point, it feels ridiculous. Compared to growth scale, you're bigger than probably 90% of the S and P. Like it's a completely different business. Tomas from Theory was sharing some, some kind of relevant data. He said we're about to witness three of the largest IPOs in history. SpaceX is targeting one and a half trillion. OpenAI aims for one trillion. Anthropic is valued at 380 billion. Combined they're at 2.9 trillion in potential market cap. The scale is unprecedented. But the real problem isn't the market cap, it's the float. Typical IPOs offer 15 to 25% of their shares to the public markets. This creates enough liquidity for price discovery while allowing founders and early investors to maintain control. Facebook floated 15% at the 60 billion that you mentioned and actually traded down pretty much immediately. Right. Google floated 19%. Alibaba floated 15% at 15% float. Here's what these three IPOs would require. SpaceX would be 300 billion or 225 billion. Opening Eye would be 150 billion. Anthropic would be 57 billion. That's a lot of smackaroos. He was. Yeah, a lot of dollars. He was comparing that to Saudi aramco, Alibaba and SoftBank, which were combined at the IPO. I believe Saudi Aramco raised 29 billion at a $1.7 trillion market cap. So he's making the case you can't really model how the public markets will absorb these companies off of Saudi Aramco. Even though from a sort of like top line market cap standpoint, it is a good proxy. Just because the float was significantly lower and Saudi Aramco's float now is only at 2.4% and they floated 1.5% at the IPO. So we'll see what the labs end up doing. They are obviously wildly capital intensive businesses and you can imagine they raise quite a bit more than the Aramcos or the Alibaba's. Saudi Aramco was such a wild ride. I feel like they were trying to IPO for like a. The San Francisco company. It is, yeah, founded in California. Like I remember hearing Saudi Aramco IPO rumors in like 2015. I think it actually kicked off in 2016. They finally got out in 2019. It was, I mean it was the largest IPO ever. There were like a million investment banks attached, like going all over the world Marshaling capital. On January 24, 2026, Saudi Aramco chairman says IPO could open to international markets. And then a year later they picked an IPO advisor locally. Then another year later HSBC came on. Then it just took. They favored New York for the Aramco listing. They had to pick all these different things. It took so long. But they sold $12 billion of bonds out of record 100 billion demand for those bonds in the pre IPO sale. It was a wild, wild, winding road. I actually know a banker who worked on the job and it was like multiple years of his life. It was very interesting. Anyway, let me tell you about Vibe Co, where D2C brands, B2B startups and AI companies advertise on streaming TV, pick channels, target audiences and measure sales just like on Meta. And let me also tell you about Okta. Okta helps you assign every AI agent a trusted identity so you get the power of AI without the risk. Secure every agent, secure any agent. Anthropic dials back AI safety commitments. Company says competitive pressure prompts it to pivot away from a more cautious stance. Anthropic, the company known for its devotion to safety, is scaling back that commitment. The company said Tuesday it is softening its core safety policy to stay competitive with other AI labs. This is so interesting. I. Anthropic, let's. I'll read through it and then we can talk about it. Anthropic previously paused development work on its model if it could be classified as dangerous. But it said it would end that practice if a comparable or superior model was released by a competitor. That basically opens them up. Given that they are at the frontier. That kind of opens them up to, I would say, perpetually, you know, kind of avoiding some of their, some of their prior policies. Sure, sure, sure. The changes are a dramatic shift from two and a half years ago when the guardrails Anthropic published guiding the development and testing of its new models established the company as one of the most safety conscious players in the space. Anthropic faces intense competition from rivals which regularly release cutting edge models. It's also locked in a battle with the Defense Department over how its clawed suite are used after it told the Pentagon it couldn't be used for domestic surveillance or autonomous lethal activities. Anthropic said the safety policy change is an update based on the speed of AI's development and a lack of federal AI regulations which they have been pushing for. Anthropic, which started as a AI safety research lab, has battled the Trump admin by advocating for state and federal rules on model transparency and guardrails. The admin has of course sought to curb state's ability to regulate AI. Spokeswoman at Anthropic said the change is intended to help the company compete with several rivals against an uneven policy backdrop. That puts the onus on companies to make their own judgments about safeguards. She said the safety pledge is unrelated to the Pentagon negotiat. The policy environment has shifted towards prioritizing AI competitiveness and economic growth. While safety oriented discussions have yet to gain meaningful traction at the federal level, the company said it is still committed to industry leading safety standards and Time originally broke the story. So yeah, I would say the obvious sort of criticism here would be that you were heavily focused on safety when you were far away from I would say leading in AI and so switching up now that like there's actually switching up on their day one, switching up on their day one. Now that there's, now that there's real competition are they forgetting where they feels a little self serving? Yeah. It's possible the money changed them. It's possible the money changed them. It's possible they always plan to switch up on their day ones. Maybe, maybe once they got, once they got to to the level they're at now. Yes. But, but it just feels like, it feels like all the initial concerns or many of the initial concerns that were guiding that entire philosophy around the company are still real. Okay. Yeah. Maybe Tyler, what do you have to say? It could just be that they realize like alignment's pretty easy and we don't need to worry about that. Well. Well. So I mean that's the very weird rule. I mean the original rule.

NEWSMAX