The AI Daily Brief: Artificial Intelligence News and Analysis - Who Thinks There's an AI Bubble?

Starting point is 00:00:00 This podcast is supported by Google. Hey, y'all. I'm Brian J. Salva, product lead and maintainer for Gemini CLI, a lightweight utility that gives you bare metal access to Gemini for software development, research, and tons of creative use cases. We're building Gemini CLI as an open source project, where community members like you have contributed thousands of ideas and bug fixes. Check out Gemini CLI on GitHub to get started. Today on the AI Daily Brief, the changing conventional wisdom around AI

Starting point is 00:00:33 and a potential AI bubble. And before that in the headlines, NVIDIA gets to play in China, but not without paying a fee. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Hello, friends, quick announcements before we dive in. First of all, thank you to today's sponsors,

Starting point is 00:00:55 Gemini, KPMG, Blitzy, and Superintelligent. And to get an ad-free version of the show, go to patreon.com slash AI Daily Brief. We kick off today with a super weird one. Although maybe it's not weird in the context of just how out of the ordinary everything is right now, but the TLDR is that NVIDIA and AMD have agreed to give the U.S. government 15% of the revenue for Chinese chip sales in exchange for export licenses.

Starting point is 00:01:19 Basically, there's been this back and forth between NVIDIA, specifically Jensen Huang and President Trump around whether or not they were going to be able to sell their H20 chips into China. First, it looked like Trump had been charmed by Jensen and his promise to invest a half trillion dollars in the U.S., but then Elon came in and the White House surprised NVIDIA by blocking the sales, and now things have gone around. Trump and Wang have built a relationship. Right as they were finishing the deal, apparently, Trump asked Huang for Invidia to give the U.S. government directly 20% of their sales to China. He countered with 15, and then they took that to

Starting point is 00:01:51 AMD as well. The Financial Times writes, the quid pro quo arrangement is unprecedented. According to export control experts, no U.S. company has ever agreed to pay a portion of their revenues to obtain export licenses. But as they point out, as out of sync with historical norms as it is, quote, the deal fits a pattern in the Trump administration where the president urges companies to take measures such as domestic investments to prevent the imposition of tariffs in an effort to bring in jobs in revenue to America. Now, some back of the envelope math suggests that NVIDIA could have sold around 23 billion worth of chips into China this year without the heightened export controls, making the U.S. government stake around $3.5 billion annually.

Starting point is 00:02:28 NVIDIA for its part did not deny the reports, but it was playing pretty coy. A spokesperson said, we follow rules the U.S. government sets for our participation in worldwide markets. They continued, while we haven't shipped H20 to China for months, we hope export control rules will let America compete in China and worldwide. America's AI tech stack can be the world standard if we race. For China watchers, it's a bizarre decision. Liza Tobin, a China expert who served on the National Security Council during the first Trump administration said, Beijing must be gloating to see Washington turn export licenses into revenue streams. What's next, letting Lockheed Martin sell F to China for a 15% commission? Still this rationale from China Hawks is

Starting point is 00:03:06 been roundly rejected by the administration. After months of hemming and hawing, they basically came around to Nvidia's view of the situation, namely that China will have advanced AI chips one way or another, so it's preferable that Nvidia dominates the market rather than Huawei. The 15% kicker to seal the deal had a lot of people gobsmacked. Historian Aaron Astor writes, The Nvidia arrangement is corporatism at its worst. Each U.S. company must pay an extortion fee to do business overseas. If they aren't big enough, they are subject to a massive tariff. the effect will be to destroy smaller businesses. Political commentator Armand Domaluski said, I feel like everyone is missing the point here that we restricted chip sales to China

Starting point is 00:03:42 because China having access to these is a national security threat. The problem isn't that NVIDIA got shaken down, it's that they can now pay a bribe to endanger U.S. national security. Now, if those opinions were a little bit too politically left for you, on the right, we've got Eric Erickson who said, I don't think NVIDIA and AMD giving the government money in order to sell the China is a good idea. And everyone saying it is good will be livid when the Democrats built upon the precedent. Carnegie Endowment Fellow Peter Harrell writes, The Chinese would pay a lot for F-35s and advanced U.S. military technology too. Regardless of whether you think Nvidia should be able to sell H-20s in China,

Starting point is 00:04:15 charging a fee in exchange for relaxing national security export controls is a terrible precedent. In the finance industry, meanwhile, some are suggesting that it's starting to look like capitalism with mob characteristics. Scott Phillips, the CIO of Motley Fool Australia, wrote, Tell you what, we usually mention African nations when talking about sovereign risks, but this is a mafia-style sovereign risk shakedown. Still others on Wall Street are adapting just fine. Futurum Equity's chief market strategist Shea Bula wrote,

Starting point is 00:04:40 A predictable, modelable cost that replaces the binary risk of a ban. Investors should prefer this to the constant headline roulette we've had since the first export restrictions. If this becomes the export control playbook, the China AI market is no longer a question mark. It's a modelable growth stream Wall Street can actually price in. Now, interestingly, in the wake of this decision, China is now officially urging its firms to avoid using the NVIDIA-H-20 chips. Bloomberg writes, over the past few weeks, Chinese authorities have sent notices to a range of firms discouraging use of the less advanced semiconductors. The guidance was particularly strong against the use of H20s for any government or national security-related work by state enterprises or private companies.

Starting point is 00:05:18 Like I said, on the one hand, this is clearly something very new, but at the same time, right now, basically everything is new, so at this point it's hard to be surprised by just about anything. Now, on the product side, Invidia keeps humming along merrily and working on a whole set of things that aren't just their chips. The company has, for example, unveiled a new set of world models designed for embodied AI. Expanding their existing Cosmos family of world models, invidia has released Cosmos Reason, a 7 billion parameter reasoning vision language model. Nvidia says the model is tuned to perform data curation tasks, robot planning, and video analytics. One of the things we don't talk about all that much on this show is embodied AI, or AI in robotics

Starting point is 00:05:56 and physical devices. But even as we pay most attention to the latest LLM and the latest agentic tool, there is just a ton of work going on on this stuff in the background. And when it comes to big transformative opportunity, a lot of that work is likely to have as big an impact as anything happening over in vibe coding land. Even if you're not paying all that much attention, it's a good reminder that Nvidia is very much working on things far beyond just their chips. Now, speaking of vibe coding, one AI-related coding platform that's going through some big changes is GitHub. When Microsoft acquired GitHub back in 2018, the division was set up to operate independently. Microsoft was largely interested in acquiring its user base of developers,

Starting point is 00:06:36 effectively operating GitHub as a loss leader to drive sales to their cloud division. However, with the advent of AI coding, GitHub transformed from a value-added platform for Azure into a revenue driver in its own right. GitHub co-pilot was one of the first AI coding tools going live in late 2021, a full year before the release of ChatGBT. Now, it is a little difficult to pin down just how popular GitHub co-pilot actually is. Microsoft boasts 20 million all-time users but doesn't say anything about active users. Still, you have to assume that their raw distribution is just the biggest

Starting point is 00:07:06 of any coding platform out there. It's in that context that GitHub's CEO, Thomas Domke, is stepping down after nearly four years in the role in a decade at Microsoft. In a memo to GitHub employees, he said, after all this time, my startup routes have begun tugging on me, and I've decided to leave GitHub to become a founder again. Interestingly, he announced that his role would not be replaced. Instead, GitHub will be folded into Microsoft's core AI team. This is the new AI engineering team led by former meta-executive, Jay Perrick. The team is separate, however, from the consumer-focused AI team led by Mustafa Sullyman. The implication, obvious I'm sure to you AI Daily Brief listeners, is that Microsoft is

Starting point is 00:07:43 recognizing that GitHub is no longer as just a hub for developer relations, but rather a core tool in the AI competition. AI coding has become the most important vertical in enterprise AI, and GitHub co-pilot is Microsoft's main play in that field. Bringing it closer suggests that Microsoft is appreciating the competitive dynamics, although, of course, what it means for GitHub users remains to be seen. One person responded to Thomas's post on X saying, please tell me you're going to reinvent GitHub, to which he responded only with a winky face emoji. Now, one other interesting note, sort of to do with GitHub, certainly to do with the AI coding competition overall, has to do with Anthropics revenue. A few weeks ago, we learned that Anthropic

Starting point is 00:08:21 had reached $5 billion in ARR, driven largely by the release of Cloud 4 and its heavy use in AI coding. Unpacking the numbers, 3.1 billion of the 5 billion comes from API revenue, which was at least at the time, higher than OpenAI's API revenue. However, half of that number comes from just two customers, Cursor and GitHub co-pilot. Peter Gostov, the head of AI at Moonpig posted, OpenAI and Anthropic both are showing pretty spectacular growth in 2025, with OpenAI doubling ARR in the last six months, from $6 billion to $12 billion, and Anthropic increasing 5x from 1 billion to 5 billion in seven months. If we compare the sources of revenue, the picture is quite interesting. OpenAI dominance consumer and business subscription revenue. Anthropic just exceeds on

Starting point is 00:09:03 API 3.1 billion versus 2.9 billion. Anthropics' API revenue is dominated by coding with two top customers, cursor and GitHub copilot, generating 1.4 billion alone. Plus, Anthropic is already making 400 million ARR from Claude double from just a few weeks ago. My sense he writes is that Anthropics' growth is extremely dependent on their dominance in coding. Pretty much every single coding assistant is defaulting to Claude 4 Sonnet. If GP5 challenges that with, for example, cursor and GitHub copilot switching to OpenAI, we might see some reversal in the market. And this is why, of course, that one of the most important parts of the GPT5 story was the extremely

Starting point is 00:09:38 competitive pricing. At this stage, studies have found that most AI coders are more focused on having the best model than they are on price, and so far at least five days or whatever it is into the WPT5 adventure, it doesn't seem to me like Claude's coding dominance is under major threat. At the same time, it's clear that the pricing model is directly pointed at Anthropic, so this battle just got a lot more interesting. For now, however, that's going to do it for today's AI Daily Brief Headlines edition. Next up, the main episode.

Starting point is 00:10:10 This podcast is supported by Google. Hey, everyone, David here, one of the product leads for Google Gemini. If you dream it and describe it, V-O-3 and Gemini can have. help you bring it to life as a video. Now with incredible sound effects, background noise, and even dialogue. Try it with a Google AI Pro Plan or get the highest access with the Ultra Plan. Sign up at Gemini.com Google to get started and show us what you create. AI agents are the buzzword that everyone's talking about, but do you truly understand their significance? KPMG's agent framework demystifies the concept, offering practical steps to

Starting point is 00:10:52 unlock AI agent's immense potential. Think of it. as your GPS for AI strategy. KPMG partners with clients to harness the benefits of of AI agents, guiding you from strategy to execution with a secure architecture and a plan for workforce. Check out their comprehensive insights on scaling agent power within your enterprise. This isn't just about tech. It's a leadership imperative. Go to www.kpmg.us slash agents to learn more, that's www.kpmg.us. This episode is brought to you by Blitzy, the Enterprise Autonomous Software Development Platform with Infinite Code Context. Blitzy uses thousands of specialized AI agents that think for hours

Starting point is 00:11:35 to understand enterprise-scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80% plus of the development work autonomously while providing a guide for the final 20% of human development work required to complete the sprint. Public companies are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool,

Starting point is 00:12:03 pairing it with their coding co-pilot of choice to bring an active STLC into their org. Blitzy is providing a limited time, 30-day free proof of concept for qualifying enterprises. The team will provide a 5x velocity increase on a real development project in your org. Visit blitzie.com and press book demo to learn how Blitsey transforms your STLC from AI-assisted to AI Native. That's BLITZY.com.

Starting point is 00:12:28 If you are a regular listener, you will have heard about Super Intelligence Agent Readiness Audits at this point. But I wanted to tell you today about the full suite of Agent Readiness products that go beyond just the initial readiness report. Over the last six months, Super Intelligence has built out an entire Agent Planning suite. We help you move from discovery to planning to implementation. After you've completed your agent readiness audits, we help you double-click on your most important use cases with what we call our use case planning reports. These reports are going to help you understand

Starting point is 00:12:59 what sort of technical preparation you need to do to be ready for a use case, what challenges you might face in implementation, and whether you should be thinking about building, buying, partnering, or some combination. After that, you can even get a spec document in what we call our technical blueprint that gives either your developers or the developers of the partner you work with, what they need to build exactly the agent that you're looking for.

Starting point is 00:13:19 If you want to learn more about superintelligence agent planning suite, we built a custom GBT to answer your questions. Just go to bit.ly slash super super super agent. That's bit.l.combe, all one word. And if you have any questions, the agent can even help you book an appointment with our team. Welcome back to the AI Daily Brief. Today we are talking about something really interesting. We're talking about the changing conventional wisdom around many parts of AI, but specifically in this case an AI bubble. As we discussed yesterday extensively, the launch of GPT-5 has been extraordinarily significant, but not necessarily for the reasons that we thought. TLDR, it wasn't significant because it heralded AGI, but because it actually reset some people's expectations. You're seeing a lot

Starting point is 00:14:10 of posts like this one from David Sacks that are basically arguing that the rapid takeoff folks have been wrong, that these models are incredibly impressive, but that the development is happening in a way that's competitive and, yes, fast, but not likely to wipe out entire fields and job categories all at once. Adam Butler of Resolve asset management made this same point as well, basically saying, look, yes, AI is getting better, but it's not a straight line up in terms of pure model capacity, and that means what it's time to do is the hard work of actually integrating these models into, as he put it, the 80% of the economy that still runs on Excel and email. An interesting question is how is this impacting how the market thinks about AI?

Starting point is 00:14:50 Well, one interesting context to understand that is a story in the Wall Street Journal this week about a 23-year-old former Open AI researcher who has very quickly raised $1.5 billion for a hedge fund that is wildly outperforming markets currently. And to understand a little bit where this came from, you have to go back just a bit. It has been extremely painful to be on the wrong side of the AI trade. People have been calling Nvidia a bubble for years at this point, Analyst so far, they've been wrong the entire time. You might remember last summer when market analysts got a little bored,

Starting point is 00:15:20 and Goldman Sachs dropped the report called Gen AI, too much spend, too little benefit. You have to think that this thing is pinned up on notice boards across Wall Street as a cautionary tale. Because at the same time Wall Street was buying into that narrative, which if you were a listener to this show at that time last year, I'm sure you would not have fallen prey to, that same month, a 22-year-old OpenAI.

Starting point is 00:15:40 Researcher called Leopold Ashenbrenner, published a 165-page blog post called situational awareness. It explained where we were in the summer of 2024, where we were going, and what it was going to take to get there. In short, it was effectively a roadmap for AI investment. One of the opening paragraphs read, The AGI race has begun. We are building machines that can think and reason. By 25, 26, these machines will outpace many college graduates.

Starting point is 00:16:07 By the end of the decade, they will be smarter than URI. We will have superintelligence in the true sense of the word. Everyone is now talking about AI, but few have the faintest glimmer of what is about to hit them. InVIDIA analysts still think 2024 might be close to the peak. Mainstream pundits are still on the willful blindness of, it's just predicting the next word. They see only hype in business as usual. At most, they entertain another internet-scale technological change. Before long, the world will wake up, but right now there are perhaps a few hundred people,

Starting point is 00:16:36 most of them in San Francisco and the AI labs, that have situational awareness. So, my friends, what do you do if you have that situational awareness, if you have incredible conviction of what you are aware of, and you feel that other people don't? Well, of course, the answer is start a hedge fund and start taking investments. According to the journal, the fund now manages more than $1.5 billion. Some of the investors include Patrick and John Collison, the founders of Stripe, Daniel Gross and Nat Friedman, who are some of the most prolific AI investors and who recently joined meta-superintelligence efforts. And increasingly, it appears, the firm has money from traditional Wall Street investors as well. The strategy is pretty simple.

Starting point is 00:17:15 By semiconductor, infrastructure, and power companies that stand to benefit from the AI buildout, add a few startup investments, including a position in Anthropic, and chuck in some short bets on the industries that will be left behind in the AI revolution. This being Wall Street, though, of course, what ultimately talks is money, and at least so far, situational awareness has the gift of gab. The company was up 47% after fees in the first half of the year, compared to the S&P 500, 6% in the same time, and an index of tech hedge funds who are up about 7%. Despite having no professional money management experience, Leopold was highly convicted even all the way back last year. Speaking on the Duar Keshe podcast, he said, we're going to have way more situational awareness

Starting point is 00:17:54 than any of the people who manage money in New York. We're definitely going to do great on investing. Now, Leopold is far from the only person to build a dedicated AI fund. The journal highlighted several other AI-focused hedge funds who were up big over the past year, although none to match the performance of Leopold's fund. Wrote the WSJ, It's no surprise that thematic funds are springing up to capitalize on the AI frenzy. In years past, hedge funds that specialized in the transition to clean energy and investing with an environmental social and corporate governance lens proliferated in response to client demand. Identifying a winning theme isn't the same as trading it well. Investors' tastes can be fickle.

Starting point is 00:18:28 Many prominent ESG hedge funds have either shrunk or gone out of business. And yet, I think that the comparison to ESG, while some might see as an arrow in the heart of this whole AI fund, movement actually serves to explain how different AI is. In short, recent trends like ESG have been at least a little bit fabricated, by which I do not mean that climate change isn't real or that renewable energy isn't a valuable investment, but many of these ESG funds focused on arbitrage and carbon markets are buying into companies that were riding high government subsidies. Basically, the general problem was that the demand for ESG investments far exceeded the good ideas that could be funded by them. That led obviously to terrible performance and many of those

Starting point is 00:19:04 funds are now shuttered. In contrast, the AI theme is A, much larger, and B, has been starved of capital for the scale being attempted. You can see the issue in startup fundraising. Anthropics' concession to having to take sovereign wealth fund money from the Middle East was not because U.S. venture capital wasn't interested in funding them. It was because it didn't have the scale to fund them anymore. Now, I will say that to the extent that you're looking for a challenge for Wall Street investors, the ways to play the AI theme are fairly limited right now. A lot of its base, on the infrastructure buildout, which is understandable. I recently shared on Exa chart from Bloomberg showing how data center construction had almost matched in sheer dollar terms the amount being spent

Starting point is 00:19:44 on general office construction. And in that world, obviously investing in infrastructure makes sense. But still, there are so many themes in and around AI that Wall Street simply doesn't have access to. In fact, in the absence of that exposure, traders are getting creative. Instead of just investing in AI companies, they're also trying to get out of stocks that they perceive as being under threat from AI. Bloomberg writes, investors have started placing bets on just where the disruption will occur next. Ditching shares in companies some strategists expect will see falloffs and demand as AI applications become more widely adopted. This, as we heard just a minute ago, is a part of Leopold's strategy, but it's also just a general way to play the second-order implications of the AI

Starting point is 00:20:23 revolution. Bank of America recently published a list of 26 companies they think will be the most at risk from AI disruption. And interestingly, it's probably not exactly who you would think. The list is not stocked with outdated legacy companies doing boring things. Instead, it's mostly about tech companies that excelled in the last decade who are suddenly finding that their entire business can be replaced by AI. Among them are web development firm Wix, stock image company shutterstock, and even software conglomerate Adobe. This group has underperformed the S&P 500 by 22 percentage points since mid-May, after more or less keeping pace since the release of ChatsyPT in late 2022. Gardner is another prime example of a company that's facing severe disruption from

Starting point is 00:21:02 AI, with the market starting to take note. There are, of course, a research and advisory firm that makes heavy use of proprietary data in-depth analysis, and while it would be going too far to say that their business model was made obsolete by deep research, that's certainly the concern on many investors' minds. When Gardner cut revenue forecasts at their last earnings call, the stock fell 30% over the next five days, their largest single week drop on record. So far this year, the stock has been cut in half. Morgan Stanley also said that the results, quote, added fuel to the AI disruption case. Said Adam Sarhan, CEO of 50-park investments, there are a lot of pockets of the market that could be basically annihilated by AI, or at least the industry will see extreme

Starting point is 00:21:40 disruption, and companies will be rendered irrelevant. Any company where you're paying someone to do something that AI can do faster and cheaper will be wiped out. Now, we are treading in new territory here. While the PC boom and the internet did have a ton of disruption, the pace of change was much slower. We haven't really had an example of one industry threatening to take out large chunks of the market, basically ever. Said Phil First, CEO of HFS research, Wall Street clearly has the jitters. This is going to be a tough, unforgiving market. Now, one prediction that I have is that in this context, where AI just keeps outperforming, people are going to be desperate to find indications of a bubble. The latest example of this, I think, shows why it's so challenging to try to apply

Starting point is 00:22:22 historical analogies to the AI space. We've talked recently about reports that the coding platforms are spending more serving their customers than they are making in revenue, even though they're some of the fastest revenue-growing companies in history. Surrounding that, you're starting to see discourse like this one from Antonio Garcia-Martinez, who I like quite a bit and think is very smart, but happen to disagree with vociferously in this particular case. Following the reports of the negative gross margins of these companies, Antonio wrote, Every tech bubble is initially pumped by some extraneous source of liquidity poured into unsustainable growth. Web2 consumer used VC for paid ads growth to pump MAUs. Crypto used retail tokens to pump user rewards to inflate usage, and hence the token.

Starting point is 00:23:02 AI is using VC and inflated equity to subsidize compute costs and inflate consumer usage. This all also happened with railroads in the Telegraph, too. It all blows up at some point and the real businesses crawl out of the wreckage and eventually become Titans. Nothing new under the tech sun. Now, of course, the nuance to Antonio's point is that he's not one of these doom profits who's saying that AI is BS. He's talking specifically about the structure of the current market surrounding AI. an important distinction. And yet I don't think that these analogies hold, or at least not entirely. The argument that he's baking is really two bundled in one. The first has to do with a period in any

Starting point is 00:23:37 new technologies lifecycle when investment capital subsidizes the cost in a way that is unsustainable. Interestingly, he didn't point to one of the most accepted versions of this, which was Uber and the whole on-demand revolution in San Francisco in the 2010s. I live there, and my lifestyle was very much subsidized by VCs for a half decade or more at prices that no one has. seen since 2016 or 2017. And so on that part, I don't disagree that AI may be being priced too low right now relative to what it will eventually be because of the subsidy of external capital. Where I disagree is with the second part of the argument, which is basically in short that it is only because of that subsidy that usage is as high as it is. In the examples that he

Starting point is 00:24:19 gives, there was no there there on the other side. Monthly active users on a Web 2.0 platform do not guarantee revenue. You have to build a thriving ad model around that, which introduces all sorts of complications and was far from a guaranteed proposition. Crypto is a whole different thing, and as someone who spent a long time in that field, frankly, outside of a very small handful of tokens, the there there in terms of usage has never been a thing. It's all been speculation. AI is wildly different. You can't have a single platform like ChatGBTBT, grow from zero to 700 million users in two and a half years and explain it away by just saying that they're pricing it too low. Maybe they are pricing it too low, and maybe that makes you squint at

Starting point is 00:24:59 OpenAI's business model. The argument that that's inflating consumer usage just doesn't hold water. Let's take the more recent example of these web coding tools. The reason that these companies might be losing money is because the customers want literally as much of the product as they can possibly get their hands on. That's wildly different than only using it because it's cheap. One of the reasons this challenge is coming up right now is that in addition to more and more vibe coders coming online who had never interacted with code before, more sophisticated developers are starting to employ strategies and workflows

Starting point is 00:25:31 that involve spinning up multiple background agents at the same time, which definitely do work in the background consuming a huge number of tokens. By one example, the number of tokens processed by Google jumped from $480 trillion in May to $980 trillion in July, a 104% growth in just a couple of months. And for those who have argued that costs weren't going to come down, like very notably Goldman Sachs back in that report from last year, that's also been wrong.

Starting point is 00:25:59 In fact, the cost of inference has come down dramatically, frankly, much faster than anyone would have anticipated. It's just that demand is going up faster. And so companies adjust. Repplet, for example, changed from charging for each completed request to instead charging for the amount of time and compute used, and their margins went right back up. The reality is the people consuming tokens for coding haven't really been put in a position yet where they really have to make the choice at scale

Starting point is 00:26:25 between the most day-of-the-art model and the more cost-efficient models. I'll remind only that when GPT-5 came out, it was priced about 10 times lower than anyone thought it was going to, clearly making efficiency and cost a vector of competition. The point of this is not, that markets aren't overpricing certain stocks, or that there are inflated at expectations that AI will at various points not be able to meet, it is only to caution against blithe bubble comparisons that too lazily pull comparisons from the past when they simply don't apply to the new reality. Some think it's a generational thing. If there's one thing Wall Street really hates, it's losing to a kid. Young Macro writes, capital management will soon become

Starting point is 00:27:08 an activity exclusively done by chronically online zoomers. He continued, just like everyone would have called you insane in the late 20th century if you suggested that a majority of new companies, companies would be launched by the college-age cohort within a decade or two, they've been calling me crazy for insisting on this inevitability for three straight years now. Now, I have no idea if Leopold and situational awareness's ability to keep these high numbers up, will continue. I think that the more likely scenario is that Wall Street humbles itself and just gets on the train. The point is, though, for now at least, AI markets are exuberant and full steam ahead. Now, this is one that is sure to cause some debate. You can't say the bubble word without

Starting point is 00:27:45 surfacing a huge array of opinions, all of which of course are valid. And there's lots and lots more evidence to look at in different places. So I encourage the conversation, get to the comments, have fun with it. For now, though, that is going to do it for today's AI Daily Brief. Appreciate you listening or watching as always. Until next time, peace.

The AI Daily Brief: Artificial Intelligence News and Analysis - Who Thinks There's an AI Bubble?

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.