Loading summary
Transcript3 lines
- [00:01]
A
Today on the AI Daily Brief, the way we use AI is changing. Before that in the headlines, more White House level discussion about the government taking a stake in the big AI labs. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright friends, quick announcements before we dive in. First of all, thank you to today's sponsors KPMG Scott Scrunch section and Outsystems. To get an ad free version of the show go to patreon.com aidaily brief or you can subscribe on Apple Podcasts. And if you want to learn more about sponsoring the show or really find out anything about the show at all, head on over to aidaily Brief AI and for sponsorship specifically you can email SponsorsIDaily Brief AI welcome back to the AI Daily Brief Headlines Edition. All the Daily AI news you need in around five minutes. Last week was a pretty interesting one when it comes to AI policy, specifically in the fact that you had people as far apart as Bernie Sanders and Donald Trump talking about frankly not totally dissimilar proposals for the government's relationship with big AI companies. Now heading into this week, President Trump has confirmed reports that the government is looking to take an equity stake in major AI labs. Notice first reported on the topic late last week. And while the reporting was well sourced and from a very well known Washington insider type of reporter, the it was unclear how far along the plans were, certainly seeming a lot more like a concept of a plan than a full executable plan itself. When reporters asked about the plan on Friday, however, Trump responded, there's a concept out there where pieces could be given to the American public. There's something very interesting about it where the American public essentially becomes a partner with the companies. I've spoken to all of them. We're talking about it where the American people can benefit from the success of AI and by doing that, they're going to like it better. Trump added that he is potentially meeting with, quote, all the big ones at the White House this week. Now some reporters even straight up asked him about Bernie Sanders call to tax 50% of AI company equity to form a sovereign wealth fund. Trump said, as far as the economics is concerned, we have certain things that aren't that far apart. People are surprised. Now OpenAI appears to be actively pushing the concept in Washington. CNBC reports that Sam Altman met with Bernie Sanders on Wednesday to discuss the idea with sources saying that OpenAI is pitching the idea of donating equity to the US government to cede a public wealth fund. OpenAI views this as a way for the public to benefit in the upside of AI growth, possibly through dividend distribution from the fund. They've also suggested that the fund could be allocated to individuals, such as through the new Trump accounts for children. Now, for anyone who's been watching this president closely, his interest in this probably isn't all that surprising. First of all, there's been talk of a sovereign wealth fund since early in the term, and the government has taken stakes in multiple companies, including intel, over the past year. He also understands the PR power of cutting a check to the American people, with his Friday comments suggesting that part of the plan is AI dividends directly attributable to OpenAI and others participating in the fund. Now, as you might imagine, there are a lot of cynical responses to this. Former Microsoft employee and tech commentator Daryl Bossinger writes, the groundwork is already being laid for a government bailout of OpenAI. Now, I don't want to dismiss a cynical take out of hand, and maybe some of you will find this even more cynical. But at this point, if you don't think the government already views our leading AI labs as too big to fail, you're not paying attention. In other areas, though, there is a bit more of a nuanced debate forming around the concept. Grappling with the issues former AIs are David Sachs wrote, while I'm no fan of socialism or arbitrary confiscations of wealth, I can see why Bernie Sanders proposal for the government to take a 50% stake in AI companies resonates, including with many on the right. Sachs argued that as public benefit companies, OpenAI and Anthropic could maximize their benefit by using a portion of their profits to pay down the national debt. Alternatively, he said he could almost support the Sanders proposal as a stupidity tax for the AI job apocalypse narrative stoked by AI leaders over recent years. There's just one problem, Sachs continued nationalization of AI will accelerate the corporate government fusion we're already sliding towards. Conservatives rightly fear a central bank digital currency. They ought to be even more concerned about central government AI, a system with even more totalistic power over information, decision making and human behavior. America won't win the AI race if we beat China, but end up with a CCP style social credit system in the US and that is a danger as the government becomes more deeply involved in AI development and assumes direct ownership and control. Now to be clear, Sachs was commenting on the Bernie proposal, not the later news about Trump. So make of that what you will. Investor Brad Gertzner took both sides of this, arguing that it's all about the mechanism, he wrote. When government owns or controls the means of production, it is socialism. I don't trust shares in the hands of some future politicians that can coerce or liquidate and spend on whatever their political beliefs. If purchased or donated, the shares should be directly held by American citizens through their Trump accounts or in a pool to trust or Trump account that will be divided among citizens in the future. In a later tweet, he said, I'm against government taking or nationalizing AI labs or other AI business, terrible slippery slope, crony capitalism, etc. But I'm encouraging founders and companies to donate shares for the direct benefit of all citizens through a pooled private account or ideally in their Trump accounts. Look, this is just the headline, so we're not going to go much deeper today, but this is a topic where the Overton window has become a flapping open Overton door and it is going to get even weirder before it resolves. The optimistic take, as Rasrx put it, is this is how you make the AI revolution something the whole country can support. Let Americans share in the wealth of the most important technology boom in human history. Next up, Elon's role as Earl of compute gets its second major customer as Google signs a three year deal with SpaceX. In an SEC filing, SpaceX disclosed that Google had agreed to pay $920 million a month to rent Compute. The deal will run from October of this year through June of 2029 and grants access to at least 110,000 Nvidia GPUs. The deal is structured in the same way as the landmark Anthropic deal last month, which granted access to the entire Colossus 1 supercluster. SpaceX will ramp up delivery over the summer at a reduced fee, but Google has the right to terminate the deal in October if SpaceX fails to deliver on the full capacity. Both parties also have the right to terminate the deal early on 90 days notice. The filing didn't specify which SpaceX facilities would be used to fulfill the deal. A Google Cloud spokesperson said, this is a short term timely agreement to ensure we have bridge capacity to meet surging customer demand for our agent platform Gemini Enterprise, which has been even higher than we expected. Now, as always with Elon deals, this one is amounting to a Rorschach test for one's personal opinion of Elon Musk. Certainly, heading into the SpaceX IPO later this week, both parties have a strong incentive to puff up the company For Google's part, they own a 6% stake in SpaceX that could be valued at $100 billion if the IPO hits its target. To some, the early termination clauses suggest the deal is all about boosting the stock over the short term. Prominent short seller Jim Chanos posted, this nine month contract has more easy outs than a kid's T ball game. The other interpretation, however, is that Elon's pivot to cloud kingmaker is succeeding. Boring Business wrote this is absolutely insane. Elon Musk's Xai reportedly spent $40 billion to build their data centers. Based on public disclosure of the Anthropic and Google deal, XAI will get paid 26 billion per year to license the compute from these data centers. That's a payback period of 18 months for all the data center spend from just two customers. And you still think AI infrastructure CapEx is a bubble now to the extent that you're thinking this is Elon's cloud strategy playing out according to plan? It's fairly unclear if there actually was a plan. In September of last year, when Xai was in the middle of scaling Colossus 2, Elon posted step 1 buy a sh tload of GPUs step 2 question mark step 3 profit and it now appears that step 2 was simply to have GPUs available during a compute crunch. Maybe that was Elon's plan all along, but I think people are wildly discounting how much. Just in advance of this IPO, Elon figured out how to make SpaceX make sense. Yu Chen Jin remarked, SpaceX has accidentally become the largest Neo Cloud on Earth. 550K GPUs more than double core weave. Starlink is doing 15 billion ARR, so GPU rentals is SpaceX's biggest business. Elon may not need XAI to beat OpenAI. Speaking of compute crunch, Jensen Huang has secured Nvidia's memory supply in a new multi year deal with SK Hynix. The two year deal will deepen ties between the two companies. SK Hynix will continue to be Nvidia's largest memory supplier as we head deeper into the shortage and in addition, Nvidia will join as a design partner on new memory chips for physical AI, personal AI and AI infrastructure. The deal also secures Nvidia's supply of high bandwidth memory as they begin to ramp production of next generation Vera Rubin chips. Announcing the deal at a press conference in Seoul, Huang said, we procure and buy from SK Hynix already billions and billions of dollars each year and it's going to grow substantially now. A big part of Nvidia's story this year has been Jensen traveling the globe to secure his supply chain, taking a very face to face approach to critical dealmaking. Last month he was in Taipei to shore up his fab allocation with tsmc, and Reuters wrote that in his trip to Seoul you could find him dining on grilled pork belly and local spirit soju with the company's top corporate bosses, throw a baseball pitch and meet with a well known gamer. Last week a video went viral of Jensen sitting down at the Nvidia booth during a Taipei conference, sipping a beer with executives. And while of course there's an element of marketing to the candid moments, it is also very clearly a critical part of Jensen's supply chain strategy. This new deal was reportedly sealed over a chicken and beer meeting with SK Group Chairman Che Taiwan at a local restaurant. And while the dining is casual, the stakes of the moment for Nvidia are very high. As he exited the restaurant on Sunday, Huang told the press, demand is enormous. Everything in the entire industry supply chain, from wafers to silicon photonics to cable connectors, is in a state of supply shortage. Now. If a lot of the headlines today were about the infrastructure side of AI, the main episode is all about some shifts in how we actually use AI. And that is what we turn to now. One of the most important AI questions right now isn't who's using AI, it's who's using it? Well, KPMG and the University of Texas at Austin just analyzed 1.4 million real workplace AI interactions and found something surprising the highest impact Users aren't better prompt engineers. They treat AI like a reasoning partner. They frame problems, guide thinking, iterate, and push for better answers. And the good news? These behaviors are teachable at scale. If you're trying to move from AI access to real capability, KPMG's research on sophisticated AI collaboration is worth your time. Learn more at kpmg.com us sophisticated. That's kpmg.com us sophisticated quick question when was the last time you actually visited a website to research something? If you're like me, AI pretty much. Does that work for you now? That, of course raises a new question for brands. If AI is doing the discovering, researching and deciding who or what is your website really for that shift in user behavior, the rise of AI bots becoming your most important new visitors is what my sponsor Scrunch, is taking head on. Scrunch is the AI customer experience platform that helps marketing teams understand how AI agents experience their site, where they show up in AI answers, where they don't, and what's preventing them from being retrieved, trusted or recommended. And it's not just visibility. Scrunch shows you the content gaps, citation gaps, and technical blockers that matter and helps you fix them so your brand is found and chosen in AI Answers. Now for our listeners, Scrunch is providing a free website audit that uncovers how AI sees your site, where there's gaps, and how you're showing up in AI versus the competition. Run your site through it at scrunch.com aidaily Here's a harsh truth. Your company is probably spending thousands or millions of dollars on AI tools that are being massively underutilized. Half of companies have AI tools, but only 12% use them for business value. Most employees are still using AI to summarize meeting notes if you're the one responsible for AI adoption at your company, you need section Section is a platform that helps you manage AI transformation across your entire organization. It coaches employees on real use cases, tracks who's using AI for business impact, and shows you exactly where AI is and isn't creating value. The result? You go from rolling out tools to driving measurable AI value. Your employees move from meeting summaries to solving actual business problems, and you can prove the roi. Stop guessing. If your AI investment is working, check out section@sectionai.com, that's s e c T-I-O-Nai.com this episode of the AI Daily Brief is brought to you by Outsystems, a leading agentic systems platform built for the enterprise. Organizations all over the world are building, orchestrating and governing agentic systems on the Outsystems platform and with good reason. Outsystems open and unified platform allows teams to architect, deliver and scale governed agentix systems. With agility. Teams of any size and technical depth can use Outsystems to build, deploy and manage AI apps and agents quickly and cost effectively without compromising reliability and security. With Outsystems, you can rapidly launch ideas from concept to completion. It's the leading agentic systems platform that is unified, agile and enterprise proven, allowing you to accelerate growth, reduce operational friction and deliver real Enterprise impact with AI. OutSystems build your agentic future. Welcome back to the AI Daily Brief. There was a story this weekend about some planned updates to the ChatGPT app that I think many are completely misreading and which the accurate read of tells us a lot about the future direction of AI user experiences. Now for those of you who are watching. Just to be clear, the mockup that is on your screen right now was generated by the person who was tweeting it, just as a way to have a visual alongside their tweet. We haven't seen any mockups or leaks of a potential OpenAI super app. What we did get was this article in The Financial Times OpenAI plot's biggest ChatGPT overhaul since launch and certainly the thrust of the article was that a super app is coming. FT writes the company intends to transform the Chatbot into a quote unquote super. Apparently that combines coding tools and AI agents adding products that executives believe will generate more revenue. Now, none of this is a surprise or exactly secret. OpenAI employees, leadership, etc. Have used the super app term extensively on Twitter X and last week during the Codex events for Enterprise. We even heard more About Codex and ChatGPT blending, although exactly what that meant remains unclear. Now for the Financial Times, this is about business strategy, they write. The changes are part of a broader reorganization at OpenAI, as the San Francisco company shifts resources into trying to win lucrative business customers and compete more fiercely with rivalanthropic. The role of ChatGPT in this ecosystem looks like it might be a little bit different again, FT writes. OpenAI executives increasingly view ChatGPT, which has attracted nearly a billion users since its launch, as a gateway to introduce users to higher value products. The majority of consumers use the Chatbot for free. FT says the overhaul is going to begin rolling out in the coming weeks and will initially appear as changes to the ChatGPT website and mobile apps, which encourage customers towards using coding, image generation and apps from external partners. Now for FT, this is all about the IPO. The changes they write underline how OpenAI's strategy is moving closer to that of Anthropic, whose focus on developing products for businesses has stoked its blistering growth and will be at the heart of its pitch to investors in an IPO this year. They quote Leonis Capital partner Jenny Hsiao, who writes approximately a year ago, OpenAI's strategy was swing for the fences, whereas anthropic strategy is make money first. Now the two are converging because both of them are trying to aim for an ipo, and investors care more about money than dreams. As evidence, FT points to the shutting down of Sora as an example of their commitment to this new business focus, which is obviously something that we've talked about here as well. Now, not everyone is sold on the idea of a super apparent David Georgiegan writes Super App usually means we couldn't find the next big thing, so we're bundling everything. We have hedgy markets, which you can guess their focus from. Their handle, certainly thinks this is about the IPO as well. The overhaul they write shifts resources towards enterprise clients. With 2 million businesses already at 40% of revenue and expected to hit 50% by year end, Altman said last year that apps would become obsolete because of AI, and now he wants to build a super app. 900 million people use ChatGPT every week, 50 million pay for it, and OpenAI still loses $14 billion a year. The super app is supposed to change that. OpenAI has shifted hard towards enterprise, with 2 million businesses already at 40% of revenue and a target of 50% by December. But those are the same enterprise customers who discovered last week what AI tools cost and whether they produce anything. Putting the slightly cynical take even more bluntly, nobody builds a super app because users ask for one. They build it because a chatbot is hard to put a multiple on. This is a feature for the S1, not for you. Now, one other take, which I don't think is exactly right but does deserve some discussion, is summed up by Yoshik, who writes, one thing I've learned from the best technology doesn't always win. The company that owns the user usually does. That's why OpenAI is trying to turn ChatGPT into a super app before the IPO. Every model eventually gets copied. Getting millions of people to open your app every day is the hard part. Although on the flip side, Anand thinks that consumers might be in for a shock with these changes. They write, it's going to be interesting to watch how casual ChatGPT users react to this. Most people have only used it as a chatbot. A full redesign into Super App is going to feel like a completely different product to them. So is this all just about money and the ipo? The answer is yes, but. And I think it's worth pausing to note how frequently the investor class cannot imagine that anything that any company does is not specifically and primarily about impressing them the investor class. What's actually going on here is the embodiment of a much bigger trend and the instantiation and extension of what we have discovered are the most valuable categories of use cases for AI, which are, simply put, not about chat. Now, what's very clear is that there is a major difference between power users of ChatGPT and regular users of ChatGPT. In a recent interview, OpenAI CFO Sarah Fryer said Our free users do about 7 turns or 7 questions a day. Our first paid tier does double that, about 15. Our real paid tier plus which is $20 is about 3x and pro is about 11x over a free user. In other words, the power users are using AI more, but they're not just using AI more, they're using it differently. It's clear at this point that OpenAI views Codex as their most successful product, at least their most successful product with the type of success that they want. And anyone who's living on the AI side of X can attest to the fact that there has been a major vibe shift towards Codex over the last few months. Developer Ben Holmes recently did a poll on Twitter asking how people use coding agents. Right now with 51.1% of nearly 2,100 votes going to Codex app, with the next highest 30.9% being CLIS in the terminal. We are in the midst of a widening AI advantage gap. The gap between the value that power users are getting out of AI and and that casual users are getting out of AI is increasing fairly dramatically now. For most of the early history of post chatgpt AI, while there was a differential between the value that power users were getting versus casual users, I'd argue that the space between them was relatively consistent over time. Casual users got more value and power users got more value as they learned better and more use cases. But then agents actually became a viable thing and specifically people figured out that coding tools were weren't just for software engineers, but for any knowledge worker who could use code and bespoke applications to solve their problems and create opportunities, which is all knowledge workers. This inflection point, which really happened around the end of last year and the beginning of this year, basically between November 25 and January 26, shift the advantage gap into overdrive. The people using agents are seeing compounding value while the people using regular chat continue to see linear gains. Only now, when it comes to the business model side, it is absolutely the case that the people who are using agents are spending far more money than those who are using regular chat. The difference between seat based pricing and usage based pricing is the difference between the $3 billion run rate that anthropic had last year and the 47 billion run rate that they're currently on. If you've been listening to this show at all over the last few weeks, the number one most dominant and most important theme has been the shift from the token subsidy era to the token scarcity era where the business models are all shifting to to sell people the tokens that they're actually consuming with lots and lots of consequent changes. What I think, though, is a mistake in just assuming that this is a business model question and an IPO question is to think the thing that primarily these companies care about is the revenue scoreboard. That obviously matters. But the reason that I think you're going to see a major change in the interfaces and user experiences that OpenAI and Anthropic put in front of their customers is a recognition of the fact primarily that the people using agents are getting more value and a desire to use interfaces and user experiences to bring more of that to everyone else. If you are watching closely, there is even a gap between the power users and the power power users, reflecting just how quickly user experience patterns are evolving. OpenClock creator and now OpenAI employee Peter Steinberger wrote here's your monthly reminder that you shouldn't be prompting coding agents anymore. You should be designing loops that prompt your agents. On CNBC recently, Claude Code creator Boris Czerny told the host that about six months ago he shifted from writing code by hand to prompting Claude to write all of his code. But it turns out that even that isn't the limit of how things are changing. Here's a snippet of a conversation with Boris from just last week.
- [21:30]
B
At that point I was running, you know, maybe five, ten quads in parallel, and my coding was prompting Quad to write code. Now it's actually leveled up, I think, again to the next wave of abstraction where I don't prompt quad anymore. I have loops that are running. They're the ones that are prompting quad and kind of figuring out what to do. My job is to write loops, and this is this kind of next transition that I think we're going to see in the next few months and maybe through the rest of the year.
- [21:55]
A
Now, this idea of loops isn't totally new. Last year and into the beginning of this year, we had a lot of discussion around the Ralph Wiggum loop, which is basically a way to set up your coding agents so they would continuously try and try again without you having to sit there and prompt and interact every time they came up against the challenge. Then in March, we had Andrej Karpathy talking about Auto Research, which was a specific type of loop designed to improve an actual AI model. And now we've got the Goal Primitive, which is a way that both OpenAI and Anthropic have embedded the idea of loops into the core experience of their major coding tools, Claude Code and Codex. The goal in each of these cases is to require less human intervention and get the AI to run for longer and longer, being able to fix its own mistakes and improve its own results, and ultimately accomplish much more comprehensive and complex tasks. And so the point is, if you've already got people who are using agents opening up a significant advantage gap compared to people who are just using regular chat, and then you've got the vanguard of people at the labs who are even going a step farther in terms of how they're getting the most value, it strikes me as obvious that one of the places that the labs can try to democratize those experiences to more users is via the user experiences of their core app interfaces. And that, I think is the key point of the ChatGPT overhaul. Now, yes, like I said, this does come with financial implications, because the people who are running loops are burning way more tokens than the people who are just casually popping in to ask a question that they might have asked Google before. I just think it's reductive to think that that is the primary motivation for the labs, as opposed to getting more people to experience the insane power and opportunity that comes with actually being able to run agents at scale. And there is a lot of work to do. Railways just Jake retweeted a post about loops and said, the problem with this and why I think people are frustrated nobody has taught folks how to do this. It feels both evidently the future and also somehow gatekept. Just to be clear, he clarified, I believe it's being expressed at the fastest rate it can be. It's just both evolving rapidly and so dense. Like pulling a neutron star out of a magic hat. Adding evidence to that, Shawnu Matthew had a post go viral with over 300,000 views that retweeted Peter Steinberger's call to use loops and said, non technical idiot guy here. What does this mean for the non coder audience? I truly want to feel what Boris and Peter are talking about. Anyone got resources to help bridge me there? Exponential views as Zemus R responded and said, it's interesting, really tricky to get it to work with deep analytical problems that aren't coding loops make it worse. The error rate of each loop magnifies. Meanwhile, Dane Ketch, the CTO of Cloudflare, had about 250 responses when he asked how are people using loops? So in a world where the advantage gap is compounding, where there's new, even more advanced behavior patterns evolving, and where most people don't have any idea how to do this, how do you get everyone up to speed. One option is learning materials, and OpenAI just published a list of codex use cases which get into some of this, so that's clearly part of the strategy. But the other approach is to just lead people to how they want them to use the tools by changing the interfaces through which they use the tools. That, I think is why OpenAI is plotting the biggest ChatGPT overhaul since launch, and that's why the way we use AI is changing. There will be lots more to talk about on this front as the changes become clearer, but for now that's gonna do it for the AI. Daily Brief Appreciate you listening or watching as always. And until next time. Peace. Sam.