Loading summary
Transcript1 lines
- [00:01]
A
Today on the AI Daily Brief we're discussing the AI Doom cycle and how we can move out of doom desperation into a place of enlightened anxiety. 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 Blitzy assembly and Section. To get an ad free version of the show go to patreon.com aidaily brief or you can subscribe on Apple Podcasts. If you want to learn more about sponsoring the show, send us a Note@ SponsorsIDailyBrief AI last two things. First of all, I have an open job for a growth engineer. This person doesn't have to be pre AI technical, but you do have to be a master of Claude Code or Codex and really interested in value added ways to grow this audience by doing cool stuff for them. Again, you can find that at Jobs AI DailyBrief AI and lastly we are registering for Cohort 3 of Enterprise Claw. You can find that at EnterpriseClaw AI Final note, today's episode is one extended episode instead of a split between headlines in main. Turned out all the headlines fit in the context of the main, but I'm sure we'll be back with our normal format tomorrow. Welcome back to the AI Daily Brief. As I was preparing the show, I noticed that a lot of the stories that people have been discussing over the weekend and the ones that I wanted to cover had in a strange way a relationship to each other that was worth exploring. Not because they were directly related, but because they were all part of something that I've been thinking about for a while, which as of this episode I am calling the AI Doom Cycle. Today I want to explore what the doom cycle is, why different types of people in different contexts fall in different parts of it, and how I think we can get to the far side of it, which I believe is the healthiest place from which to actually engage with the big questions around AI and whether it's policy or something else. Now you guys probably recognize the inspiration for this, which is of course Gardner's famous technology Hype Cycle chart. The idea of the Hype Cycle chart is that new technologies tend to, in their argument, follow a pattern as they diffuse throughout society. It kicks off with an innovation, trigger the sparks of that new thing that becomes available, surges up to what they call the peak of inflated expectations. This is the very top of the hype cycle. When everyone's excited about a thing, it's it often comes with big capital injections in the form of venture capital and investment. It's the part of the curve where a big chunk of people are convinced that this new thing is the new thing that's gonna change everything. The peak of inflated expectations doesn't last all that long. And from there, you crash back down into the trough of disillusionment. This is the period during which expectations have bottomed in many cases, not because the technology is bad, but because it hasn't lived up to its inflated expectations. A lot of times, this is, in and of itself, a vector of time, where it's not even a reassessment of what the technology can do, but a recognition that the timeline for it doing those things is a lot longer than previously thought. Now, not every technology climbs out of the trough of disillusionment. Some stay mired there for a very long time. Metaverse, I'm looking at you. But for those that do, the next stage is what Gardner called the slope of enlightenment. This is where, freed from those inflated expectations, people actually start to figure out the right ways to use that technology. The things that it's actually good for. The intersection with people's lives or workflows that make it valuable, if not as earth shattering as maybe people once thought, that leads eventually to the plateau of productivity, where the technology is actually diffused into the world and is a valuable part of whatever environment it's supposed to be around, even if it never reaches the glorious heights of expectations that it once had. Now, a lot of ink has been spilled on this chart, including much research that suggests that most technologies don't actually follow it exactly. But I think the reason it has resonance is that it is intuitively reflective of our individual relationships with technology. In many cases, and I think one of the most revealing things that is true in almost all cases, is that even when we realize early on that a thing is going to be extremely impactful in the world, we just tend to wildly underestimate how long it's going to take to make that impact. The AI doom cycle, in that. In this case, I'm describing the emotional and cognitive states of people and their relationship with AI more than the technology itself. The five stages in this journey that I've identified are, first, skepticism and disbelief, followed by AI psychosis, which is, of course, the AI can do everything stage, followed by doomed desperation, real world recalibration, and eventually enlightened anxiety, with anxiety being the portmanteau of anxiety and excitement that GPT4 all the way back in the day argued strenuously to me was the right way to describe this combination of thrill and caution that many people in the AI space felt. Now, of these for our purposes today, the one that I'm going to spend the least time on is the skepticism and disbelief. This is a category, frankly, that I think is a bit on the wane in many cases since the launch of ChatGPT in late 2022. I think that much of the skepticism and disbelief came from simply not actually interacting with the technology or interacting with AI only on the neutered freeze settings, especially early on, and never having come back to try it again. You might remember we had that explosive deep seek moment back at the beginning of 2025 when the deep Seq app rocketed to the top of the app charts, actually displacing ChatGPT, also wiping out hundreds of billions of dollars in market value on Wall Street. Part of the reason, in fact, the biggest part of the reason that Deep SEQ had such a moment was that they were giving away a reasoning model for free. And that was for most people, the first time they had gotten their hands on a reasoning model. Yes, OpenAI had introduced 01 back a few months earlier in September of 2024, but it was not widely accessible, and even when it did become a little bit more available in December of that year, it was strictly behind a paywall. So when people got their hands on the Deep SEQ app, which was Powered by their R1 reasoning model, it felt categorically better to the ChatGPT style AI they had used, because it was to the extent that there is still skepticism and disbelief today, it often comes down to either a non updated priors about what AI can do B because it was bad at the thing that you were interested in it doing for you before, and so you assume it still is bad at that thing, or frankly, it's attributable to people having AI skepticism be their business model, although even that's being called out more and more recently. But where we go next in the doom cycle is up to the AI psychosis category. I'm using the not so generous term here to acknowledge the fact that for a lot of folks, once they start really appreciating how powerful AI is, they start to see its implications everywhere, perhaps even overestimating what it can do, at least in the short term. Now, AI psychosis as a term is generally used pejoratively, and I want to be clear that that's not my intention here. What this is shorthand for in the context of our AI doom Cycle chart is simply the state of being peak convinced that AI is going to change everything. Which is where we get to our first story from the last couple of days, which is the reversal in Citadel CEO Ken Griffin back in January. Griffin was, relative to many of his finance industry peers, deeply on the skeptical side, and given that he was at the time the 39th richest person in the world with a net worth just shy of $50 billion, his skepticism was enough to make headlines. At a panel discussion at the World Economic Forum, he basically claimed that AI was all hype, with the specific intention of that hype being drumming up support for the massive amount of investment which was required to keep the infrastructure build out going. Griffin said, you're not going to generate this kind of spend unless you're going to make a promise that you're going to profoundly change the world. Referencing his own work in stock analysis, he described, AI generated reports that look impressive at first blush, but once you dig deeper, it's in his words, all garbage. Griffin's mind has apparently changed. During a discussion at Stanford Business School earlier this month, he noted that AI had become profoundly more powerful, his words, than it was nine months ago, which allowed Citadel to use it in a wider range of cases. He commented, for the first time AI is real. Now we'll hold aside his dismissal of all the value that people got out of it for the three years previous to the ascent of these coding models and focus on where he's landed. He said, to be blunt, work that we would usually do with people with Masters and PhDs in Finance over the course of weeks or months is being done by AI agents over the course of hours or days. He said the firm is seeing a 15 to 25% productivity boost, adding when you're seeing really high level research being done by AI engines, it's quite eye opening now importantly, this filled Griffin not with excitement but with trepidation continuing. He said, these are not mid tier white collar jobs. These are like extraordinarily high skilled jobs being I'm going to pick a word, automated by agentic AI and I gotta tell you, I went home on Friday actually fairly depressed by this because you could just see how this was going to have such a dramatic impact on society. And so what you're seeing with Ken Griffin is the fact that very often an increase in one's belief about the power AI can also start to nudge them towards the next category on the doom cycle, which is doom desperation. And this is of course where you extrapolate out from the way you see AI changing the workforce right now, the way that is changing how companies think about hiring and the implication of that is fewer jobs. One of the biggest jobs doomers right now is former presidential candidate and UBI advocate Andrew Yang, who took the chance to quote tweet a video of the Griffin conversation, adding, finance will quickly follow. Tech and Automation via AI Now Andrew and Ken are far from alone in their concerns. Anyone who's been listening to this show certainly knows that it seemed like every couple of days one AI executive or another has taken it upon themselves to go out and do a media tour about just how many jobs are going to be leveled by the technology that they are racing with everything they have to build. And although we didn't get a set of new comments over the last couple of days giving a sense of how resonant and widespread the doom desperation is right now, a whole bunch of speeches from earlier in the year resurfaced and got another round of life thanks to reprinting in magazines like Fortune. A couple of days ago, Fortune, for example, resurfaced comments from Microsoft AI CEO Mustafa Sulaiman from a couple of months ago with the headline Microsoft AI Chief gives it 18 months for all White Collar Work to be Automated by AI. The Wall Street Journal also resurfaced videos of Anthropic CEO Dario Amadei putting big numbers on his prediction, like a 10% unemployment rate overall and a 50% unemployment rate for entry level white collar jobs. That same interviews from around that time also predicted that the cost of software was going to go to zero, which in and of itself would have implications for jobs in that industry. Which is not to say that in the tech industry there isn't some skepticism of this or Anthropic's motivation for it. Soda on X summed up what many people feel when they wrote it's this narrative that allows dario to raise $30 billion at 900 billion. Genai and LLMs are implicitly framed as a zero sum narrative. Still Menlo Ventures Dee Dee Das painted a picture of a Silicon Valley tech industry vibe that is, in a word, pretty dreary. And to use our frame clearly in this doomed desperation mode, On Friday, Didi posted something that went wildly viral with about 11 million people seeing it. The vibes in SF feel pretty frenetic right now, he wrote. The divide in outcomes is the worst I've ever seen. Over the last five years, a group of around 10,000 people employees at Anthropic, OpenAI, XAI, Nvidia, Meta, TBD and founders have hit retirement wealth of well above $20 million. Everyone outside that group feels like they can work their well paying but less than 500k job for their whole life and never get there. Worse yet, layoffs are in full swing. Many software engineers feel like their life skill is no longer useful. The day to day role of most jobs has changed overnight with AI as a result. 1. The corporate ladder looks like the wrong building to climb. Everyone's trying to align with a new set of career paths. Should I be a founder? Is it too late to join Anthropic or OpenAI? Should I get into AI? What company stock will 10x next? People are demanding higher salaries and switching jobs more and more. 2. There's a deep malaise about work and its future. Why even work at all? For quote unquote peanuts? Will my job even exist in a few years? Many feel helpless. You hear the permanent underclass conversation a lot, especially from young people. It's hard to focus on doing good work when you think, man, if I joined Anthropic two years ago I could retire. 3. The mid to late middle managers feel paralyzed. Many have families and don't feel like they have the energy or network to just start a company. They don't particularly have any AI skills. They see the writing on the wall. Middle management is being hollowed out in many companies. And 4. The rich aren't particularly happy either. No one is shedding tears for them and rightfully so. But those who have quote unquote made it experience a profound lack of purpose too. Some have gone from less than 150k to over 50 million in a few years with no ramp. It flips your life plans upside down. For some, comparison is the thief of joy. For some, they escape to New York City to live life. For others still, they start companies just cuz often to win status points. They never imagined that by age 30 they'd be set. I once asked a post economic founder friend why they didn't just sell the company and they said and do what? Right now everyone wants to talk to me. If I sell, I will only have money. I understand. Didi concludes that many reading this scoff at the champagne problems of the valley. Society is warped in this tech bubble. What is often well off anywhere else in the world is bang average. Here, unlike many other places, tenure, intelligence and hard work can be loosely correlated with outcomes in the Bay. Living through a societally transformative gold rush in that environment can be paralyzing. Am I in the right place? Should I move? Is there time still left? Am I going to make it. It psychologically torments many who have moved here in search of success. Ironically, a frequent side effect of this torment is to spin up the very products making everyone rich in hopes that you too can vibe code your path to economic enlightenment. Now, I won't get into all of the critiques of this, which of course are a big part of the reason that 11 million people have engaged. I will use Buco Capital as a sum up for all of them. When he wrote I can't stop thinking about this post. If you do one thing today, I encourage you to give it a thoughtful, thorough read and then commit to never living your life this way. Life has wasted success on the people described in this post. It really is completely pathetic. They say that comparison is the thief of joy. Look no further than this post for validation. It is indeed true. On their deathbed they will realize they have lived their life completely wrong. Don't let it be you. The point, for our purposes, is that part of the reason that doom desperation is so widespread is that a lot of it is coming from the people who are building this technology. Which, by the way, is also why a lot of the animosity is so white hot with people outside of that cultural bubble asking the obvious question of if this thing is going to be bad for us, why are you building it? And that anger is getting louder. One of the media's favorite themes right now is commencement speakers getting booed off the stage for mentioning AI. On Friday, the graduating class at the University of Arizona loudly booed Google co founder Eric Schmidt during his commencement speech. Now, it is certainly possible that Schmidt's status as a tech mogul colored the crowd's reaction. His entire speech was essentially about how the tech industry had failed the younger generation throughout the social media era. He was already getting booed as he claimed the damage was unintentional, stating, in the years after I graduated, no one sat down and resolved to build technology that would polarize democracies and unsettle a generation of young people. That was not the plan, but it happened. As an aside, at some point I'm going to do an entire show about how what I believe is increasingly broad consensus that social media has just been net bad for the world impacts the way that people look at AI. But I digress. But even if they were primed already, the boos reached a crescendo. As Schmidt said, the question is not whether AI will shape the world, it will. The question is whether you will have shaped artificial intelligence. The question is whether you will have shaped artificial intelligence. We do not know. We do not know the precise contours of what this if you'd let me make this point, please. Yikes. Still making it seem like this is a pattern and not a one off Days earlier, another speaker who was completely removed from the tech industry garnered a similar response. Gloria Caufield, the VP of Strategic Alliances at Tavistock Development Company, was loudly booed at the University of Central Florida graduation. She stated, the rise of artificial intelligence is the next industrial revolution, and the boos were loud enough to derail the speech, with Caulfield asking the crowd to let her continue. Journalist Alex Cantrowicz wrote, this is incredible. Artificial intelligence getting booed out of the stadium. In any commencement speech it's mentioned maybe telling college students AI was taking their job wasn't the best strategy. Journalist Eric Thompson retweeted that and said, AI being booed in commencement addresses seems pretty natural to me. It's really, really unusual for the people building and selling a new technology to promise that it will destroy people's livelihoods. Whether you consider this to be horrendous marketing or actually a craven justification for VC investment, or really just honest communication or rationally overemphasizing the probability of a long tail catastrophic outcome or whatever else, the point is, it's very unusual. I think today's 22 year olds should probably familiarize themselves with Claude and ChatGPT, but I don't entirely blame them for booing a technology whose architects have said this will destroy your jobs. And I think that's an important point here. It would be wildly over simplistic to just say graduates hate AI. But as Trevor Garcia put it on X getting a job out of college was already brutal. The promise of AI squeezing the market even further while billionaires promise it will create millions of jobs they can't name doesn't exactly inspire confidence. These kids, he says, are not anti technology. They're anti being told to be excited about something that feels like it was built for everyone except them. All right folks, quick pause. Here's the uncomfortable if your enterprise AI strategy is we bought some tools, you don't actually have a strategy, KPMG took the harder route and became their own client Zero. They embedded AI and agents across the enterprise. How work gets done, how teams collaborate, how decisions move not as a tech initiative, but as a total operating model shift. And here's the real unlock that shift raised the ceiling on what people could do. Humans stayed firmly at the center while AI reduced friction, surfaced insight, and accelerated momentum. The outcome was a more capable, more empowered workforce. If you want to understand what that actually looks like in the real world, go to www.kpmg.us AI. That's www.kpmg.usa AI with the emergence of AI code generation in 2022, Nvidia master inventor and Harvard engineer Sid Pareshi took a contrarian stance. Inference, time, compute and agent orchestration, not pre training, would be the key to unlocking high quality AI driven software development in the enterprise. He believed the real breakthrough wasn't in how fast AI could generate code, but in how deeply it could reason to build enterprise grade applications. While the rest of the world focused on co pilots, he architected something fundamentally different. Blitzi the first autonomous software development platform leveraging thousands of agents that is purpose built for enterprise scale code bases. Fortune 500 leaders are unlocking 5x engineering velocity and delivering months of engineering work in a matter of days with Blitzi. Transform the way you develop software. Discover how@blizzi.com that's B L I-T Z-Y.com youm've heard me talk about Assembly AI and their insanely accurate voice AI models, but they just ship something big. Universal 3 Pro is a first of its kind class of speech language model that lets you prompt speech recognition with your own domain context and vocabulary instead of fixing transcripts and post processing. It's more flexible than traditional ASR and more deterministic than LLMs, so you get accurate output at the source and can capture the emotion behind human speech that transcripts often miss. All without custom models or post processing hacks. And to celebrate the launch, they're making it free to try for all of February. If you're building anything with voice, this one's worth a look. Head to AssemblyAI.com freeoffer to check it out. 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 sec tion AI.com and here and you can probably feel it is the low part of our cycle. There's a reason I called it doom desperation. But luckily for you, if you are still hanging on, this is not where our story ends. From here we move to real world recalibration. By this I mean a step away from the loudest voices and the most media friendly extremes to an actual look at what's happening in the real world, whether it validates those narratives or not. Now to be clear, real world recalibration is not just about counteracting those doom narratives. In some cases, real world recalibration is going to be confirming of the concern. Certainly the best example of this right now are what feels like the never ending string of layoff announcements. A recent example comes from Meta, with that company telling around 10% of their staff that their last day will be on Wednesday impacting about 8,000 people. Now sources said that the official explanation is to run the company more efficiently and offset other investments, meaning presumably that payroll spending is being transferred into infrastructure spending. Wired reports that the looming layoffs and the broader AI restructuring have driven morale to new lows. One anonymous employee said, everyone is unhappy. The only people who are not unhappy are literally executives. Sixteen employees said that recently installed screen tracking software said to be used for AI training has contributed massively to the gloomy mood. One policy staffer commented, I don't know anyone having a good time. The vibe is a bit over it. Lack of connection to the mission, upcoming layoffs, American employees being used to train the AI models that will replace them. This is some of the easiest real world evidence to see, which is why it tends to dominate the discourse. But there are in fact other things happening at the same time. Let's take the conversation around token maxing. This was maybe our biggest theme from last week. The ideas surrounding this are some of the biggest themes that we've been exploring on the show recently. For a bit of background, for the last couple of months you've seen companies experimenting with some version of leaderboards or incentives for people to consume the most tokens. Now of course, Goodhart's law says that as soon as you make something a metric, it ceases to be a good metric because people just game the system. And that's exactly what's happening at places like Amazon and before that, Meta. But holding aside the gamesmanship, the positive idea of Token maxing itself of mass wide scale experimentation and using AI as much as you can to discover how it's going to actually be useful is starting to crash headlong into the real constraints in the physical world that are forcing a business model shift from the labs themselves. Back in April, before most people were paying attention to this, anthropic had started to shift their enterprise customers to usage based pricing for CLAUDE code and cowork. The $200 flat rate subscription with generous usage limits, I.e. subsidies, is no longer available for enterprise seats. Instead, enterprise customers have been switched over to a $20 per seat rate with all usage billed. This, however, got louder in the last couple of weeks as GitHub also moved to token based usage and most controversially of all, claude updating their policies to basically make it so that any usage of CLAUDE outside of a specific anthropic owned harness like the Claude code applied would be billed on a token basis rather than lumped together with generous subsidy in a flat rate plan. As we've discussed extensively, this is a structural necessity and an inevitability of a world in which token demand starts to exceed token supply. We are at the very beginning of a structural compute shortage that involves basically all of the components required to get more compute online, from electricity to memory to chips and beyond. This is not something that is going to resolve fast, meaning that the only mechanism for solving it in the short term is to use market forces, that is to raise the cost of tokens sufficiently so that all of the tokens that are available flow to the people who are most willing to pay for them. I think you're going to start to see a lot more reports like this one from Max Weinbach who wrote I don't know why it's happening, but around 5 of my friends mentioned that they had AI spend budgets added to their given AI platforms at their respective jobs this week. It was effectively unlimited before and to give a sense of just how much these companies were subsidizing things previously, the GitHub Copilot subreddit is filled with people sharing screenshots of Copilot's estimator of what their usage would cost under usage based billing. This marzipan 9239 in a post titled aptly I'm Cooked Dog, Ain't no Way show that while their current billing was $451, their usage based billing would be $11,432.22. Yuzuko's current billing was 39, but their usage based billing would be 5,851. 77 Ginger Tapers is currently spending $54.43 which while more modest than the others, was still a 22x subsidy with their usage based billing equivalent being just under $1200. When you start to see the numbers, I think you get why this is not a short term trend, but is in fact structural. And what that's going to mean for good and ill is a lot less raw consumption. Just to try stuff out, companies are increasingly going to be back into an ROI mindset where they want to see a direct relationship between token spend and value created. A couple of weeks ago Axios published a piece titled simply AI can cost more than human workers now. And boy, if that cost reality isn't going to shift the thinking on just how fast humans get displaced by their automated equivalents. And this is what I mean when I say real world recalibration. The constraints of the real world, in this case the physical real world, in which AI, it turns out, is an expensive, capital intensive technology, is changing just how quickly companies could even theoretically automate everyone away. And so real world recalibration is what starts to flow into what I've called the end state of enlightened anxiety. And hopefully it's clear from the word excitement that I am not saying that the end of the AI doom cycle is a 100% Pollyannish Kumbaya. Everything's going to be fine. It's enlightened because instead of being generally anxious and assuming that everything's going to change and it's going to change tomorrow, we can start to get more specific and clear about what's actually happening, what we think is going to happen next, and what we should do about it. Enlightened excitement also allows in a lot more space for nuanced discourse. What I mean by that is things like when we talk about AI taking away all the software jobs, being able to zoom out and see that even as AI has come around, we've also been recalibrating for post Covid zero interest rate hiring or being able to see that if in fact we are in this structural shortage of the amount of compute that we need along with every other component in the AI supply chain, that the quote unquote jobs that AI will create aren't just interesting, unimaginable ones in the future, but some very clear ones right now. The boom, for example, in data center construction jobs, looks very different in a world where it's a short term one time shift versus one where we're talking a generational buildout. And that recognition makes it a little bit easier to see, perhaps, why blue collar unions are increasingly trying to provide a bridge between big tech on the one hand and local communities who have all sorts of reasonable skepticisms of data centers in their communities on the other, and find a way to make these things be net positive for everyone. Another type of discourse that real world recalibration and the new state of enlightened anxiety allows for is for economics discussions that don't fit inside tweets and headlines, but actually explore how savings and surplus from one industry and type of work that will be changed flow into others. I think we'll look back at economist Alex Imas's essay what Will Be Scarce as one of the more influential in 2026, in which he explored how the relational sector, where the provenance of human creation or human service is actually part of the economic value of a thing, is likely to rise proportional to the savings, and where it doesn't matter if a human or a computer creates something. I've noted importantly that Alex's essay isn't just playing in tech circles, but is also finding its way into larger political discourse vis a vis people like Ezra Klein, who discussed it in his recent New York Times essay why the AI Job Apocalypse Probably Won't Happen. Real world recalibration and enlightened anxiety look like OpenAI in anthropic, launching massive consulting efforts because it turns out no matter how powerful AI is in the lab, it smashes against the wall of institutional and human inertia that is the corporate sector. It requires intensive amounts of work to actually be integrated and close the capability gap between what it can theoretically do and what it's actually doing. Paul Baum nailed it when he jokingly tweeted anthropic knows they are weeks away from AGI, which is why they are working with companies like Accenture, Deloitte, PwC and others to build joint centers of excellence and training and certify 30,000 PwC professionals on Claude. And to put a fine point on this, there was genuinely for many years inside the labs an extreme debate between the enterprise business focused folks and the research side of the organization about just how much they should focus on business model at all. If AGI is the big game and the only path to AGI is from the researchers, there is at least a compelling theoretical argument that all of your resources should go into that bigger play, not distracting yourself by helping heavy industry in the Midwest figure out how to use CO work. And yet here we are. The constraints of the real world have won and demanded that these efforts go into place and consume significant resources in the process. There's also just different ways to look at even the frustration that we're seeing on these college campuses. Journalist Joanna Stern retweeted the Eric Schmidt booing and says this wasn't the case a year ago when I gave my commencement speech almost entirely about AI. I find this pushback unsurprising and even encouraging. I say encouraging because it means there's debate and discussion. And by the way, one commencement speech that didn't get nearly as much attention was Jensen Huang's at Carnegie Mellon, which one news outlet called the most anxiety relieving speech in the AI era. This is because Jensen clearly genuinely believes, and has been saying so forever, that AI doom is not destiny, that we, including those students in the audience, have power to shape what the next phase looks like. He talked, as he always does, about the generational opportunity for the United States to re industrialize and restore its capability to build about all the specific new jobs that are going to be needed in the very short term for that re industrialization. And whether one agrees or not, it's very clear that from a sheer efficacy standpoint, this type of message is a better place to build from than the presumption that what's next is already written and we have no control. You can see the rest of the AI industry finally getting the memo. Basically starting at the beginning of May OpenAI's Sam Altman finally decided to run away from the doomsday story and recommit to a different type of narrative. On May 1, he said, we want to build tools to augment and elevate people, not entities to replace them. A week later, he said, way cooler to help software developers Pokemon evolve into superheroes than to try to replace them. And even more recently, about a week ago, he actually went further and painted a picture of how our changing relationship with work could actually make our individual lives better as well. He wrote, kicking off a bunch of Codex tasks, running around with my kid in the sunshine and then coming back at nap time to find them all completed makes me very optimistic for the future. And lastly, the point that I want to make about this enlightened anxiety state is that it's not just that we feel better, but when we are all operating in this place, we can have more interesting, sometimes harder conversations about policy and what we should do if AI is going to inexorably take all of our jobs. There's not a lot of policy discussion there. It's basically just UBI or die. If on the other hand, it is going to be disruptive we but in more specific and discrete ways and over a longer period of time than the doom prophesiers thought. That gives us a lot more opportunity for more discreet and better policy. I'm not co signing on this idea, but I think it's going to lead to conversations like this one from Matthew Iglesias who writes Allow data centers, but only if 20% of the GPUs are set aside as affordable compute for low income agents. I don't actually even know how tongue in cheek he was being, but why isn't that something we could discuss? My show over the weekend was about the potential for AI inequality in the circumstance of compute shortages, but right now local communities and policymakers have a lot of power to set the rules by which data centers are going to play. I don't believe that a lot of them would balk at something like this, at least not from the conversation about it. Or another one from Mark Cuban. On Friday he tweeted we should federally tax tokens at the provider level. Not a lot less than $0.50 per million tokens. It'll accomplish four things at least. 1 It will push big AI players to optimize tokenization, caching, routing and localization, which will 2 reduce energy usage, saving them in energy costs more than what they paid in tax, and reducing strain created by the growth in energy consumption, which will 3 generate maybe $10 billion a year to start, but over the next 10 years could grow 30x to 100x which will 4 create a source of funding to pay down the federal debt or deploy in response to the things AI brings that we don't expect or don't like. At some point the models will pass it on to customers. Of course that's okay. Customers will have the ability to choose between providers or to do everything using open source models locally. Thoughts? And boy did people have a lot of thoughts. But man, if that isn't a more useful conversation to have than any of this BS about a permanent underclass. Right now the vast majority of the world still lives in this first hump of the AI doom cycle, somewhere between skepticism and disbelief, AI psychosis and doom desperation, or some part of all of it at once. I think the faster more of us can move into the real world, recalibration and enlightened anxiety, the better off we're all going to be. For now, however, that is going to do it for the AI Daily Brief Appreciate you listening or watching as always and until next time, peace.