Peter Teal lets his children use screens for one and a half hours per week. When he said this at the 2024 Aspen Ideas Festival, the audience gasped. Steve Jobs, the man who launched the iPad, told the New York Times his kids had never used one. Bill Gates banned smartphones until age 14 and banned phones from the dinner table entirely. Evan Spiegel, the CEO of Snapchat, the man who runs Snapchat, limits his child to the same 1.5 hours per week as Steel. Steve Chen, co-founder of YouTube, said he wouldn't want his kids consuming short form content because shorter form content equates to shorter attention spans. And Elon Musk said it might have been a mistake to not set any rules on social media for his children. These people built the attention economy. They designed the platforms, funded the algorithms, scaled the engagement loops, and they won't let their own families near any of it. It is the ultimate expression of corporate confidence. A business model so profoundly robust that the people who created it treat their own product with the exact same caution a nuclear technician reserves for unshielded uranium. It's kind of comforting to know that the modern tech aristocracy is governed by the ancient wisdom of the 1980s drug dealer. Never get high on your own supply. That's the headline. And it is outrageous. Understandably, viscerally outrageous. But I think the real story is something different. And I think the framing that most people are using to respond to it is going to guarantee that nothing gets fixed. And we so desperately need it to get fixed, don't we? Today, I'd like to talk through the $30 billion experiment that produced the first generation less cognitively capable than their parents. The poll showing 69% of Americans want to seize half of every AI company's stock and why being mad at AI is kind of like being mad at a hammer and just as useful.

In 2002, Maine became the first US state to implement a statewide laptop program for school children. Then Governor Angus King envisioned a generation with the internet at their fingertips, immersed in information, empowered by access. By that autumn, 17,000 Apple laptops had been distributed to seventh graders across 243 middle schools. By 2016, those numbers had grown to 66,000 laptops on tablets across the state. King's initiative was mirrored nationwide. By 2024, the United States had spent more than $30 billion putting laptops on tablets in classrooms. It was one of the largest educational technology deployments in history. Well, the results are in and they're not good. Earlier this year, neuroscientist Jared Cooney Horvath submitted written testimony to the US Senate Committee on Commerce, Science, and Transportation. His conclusion was basically that Gen Zed is the first generation in modern history to score lower on standardized tests than the previous one. Using data from the OCD's program for international student assessment, which tests 15 year olds across dozens of countries, Horvath documented a stark correlation. The more time students spent on computers in school, the worse they performed. From a purely financial perspective, spending $30 billion to achieve a measurable decline in human intelligence is a remarkably bold strategy. No. Usually, if you want to lose that much money while simultaneously making everyone involved slightly less capable of basic reasoning, you're going to have to buy a social media platform or maybe launch a cryptocurrency exchange. Doing it to school children is a fascinating logistical pivot. A 2021 survey of 846 teachers by the Edweek Research Center found that 55% reported students spending 1 to four hours per day with educational technology. Another quarter reported 5 hours per day, but a 2014 study observing 3,000 university students found they were engaged in offtask activities on their computers nearly twothirds of the time. They were not learning. They were just kind of scrolling around. And the academic research on what happens next is accumulating rapidly. A 2025 systematic review and metaanalysis published in Psychological Bulletin, which is one of the highest impact journals in the field, studied nearly a 100,000 participants and found that short form video consumption was consistently associated with poor cognition and declining mental health across both younger and older users. A 2026 systematic review in the Journal of Adolescence synthesized 23 studies and found that short form video addiction impaired attention, executive function, and self-regulation in adolescent and young adults. A 2025 study published in memory found that fast context switching in general, the defining feature by the way of Tik Tok style consumption, directly contributed to decline in perspective memory, which is just the ability to remember to do things in the future. And the platforms are essentially engineered for exactly this. A November 2025 study led by Baylor University uncovered why Tik Tok in particular is so difficult to put down. It requires the least cognitive effort to consume of any short form platform, less than Instagram reels, less than YouTube shorts, because its algorithm balances relevant content with surprising and unexpected videos at a ratio that keeps the brain in a loweffort, highreward loop. As Gene Twang, psychology professor at San Diego State University, told Fortune, "Many apps, including social media and gaming apps, are designed to be addictive. Their business model is based on users spending the most time possible on the apps and checking back as frequently as possible. The average US child aged 8 to 18 now spends 7.5 hours per day watching or using screens according to the American Academy of Child and Adolescent Psychiatry. 7 and a half hours. That's a lot. That is a full working day of screen consumption every single day by children who were supposed to be developing the cognitive architecture they will need for the rest of their lives. Horvit's view is also very stark. He called Jenzie victims of a failed pedagogical experiment and said whenever I work with teenagers I tell them this is not your fault. None of you asked to be sat in front of a computer for your entire K12 schooling. That means we screwed up. And I genuinely hope Jenzie quickly figures that out and gets mad. Now, I want to be careful here because I am a scientist and nuance matters. Attention and cognition are not static traits. They're more like a muscle. If you spend two years scrolling Instagram, your attention patterns will adapt. But if you do spend two years off it, they just kind of adapt back. The brain is plastic. And the research on long-term permanence is still developing. But what is not in dispute is that the deployment was reckless. $30 billion spent on a nationwide experiment with no efficacy standards, no controlled pilot studies, no evidence base and no mechanism for measuring whether the intervention was helping or hurting. They gave children unfettered access to technology and just kind of hope for the best. Well, it turns out that if you swap a child's textbook for a portal to the entire sum of human chaos, they don't use it to study calculus, but only less informed people could have guessed that. And these decision makers are all incredibly intelligent. Fortune reported in 2017 that Maine's public school test scores had not improved in the 15 years since the program launched. The state's own governor at the time, Paul La Page, called it a massive failure. And Horvath's Senate testimony framed the generational impact directly. He said, "This is not a debate about rejecting technology. It is a question of aligning educational tools with how human learning actually works." Does that sound familiar? a powerful technology deployed at scale without any adequate evidence producing unintended consequences that the people in charge are now just kind of scrambling to address. We've seen this pattern before on this channel and the generation that absorbed this experiment is now the one being asked to embrace the next wave of technology being forced on them.

According to Pew Research Center June 2026 survey of 5,119 US adults, Gen Z, that's ages 18 to 29, is the most hostile to AI of any age group. 48% believe it will have a negative impact on society. And yet, they are also the generation that uses AI the most at 66%. The most engaged and the most hostile simultaneously. Horvath in his testimony basically said, "I genuinely hope Jenz quickly figures that out and gets mad." And this is where I have to push back not on the anger itself. The anger is legitimate, earned, and understandable, just as all feelings are valid. But on the direction it's pointing. I see it in the comments on this channel every single day. LLMs are stupid. AI doesn't work. I'm anti- AI. And I get it. I've said it before. People have had their lives disrupted significantly because of this tech. But being mad at AI is like being mad at a hammer. The hammer didn't decide to hit you. Somebody swung it. AI is a sub field of computer science. That's all it is. It is not venture capitalism. It is not tax structures. It is not wealth and asset inequality. AI is not the labor market. It's also not an economic system. AI is a set of computational methods for pattern recognition, prediction, and generation. Being anti-AI is like being anti- medicine or anti-mathematics or anti-ele electricity. It is a framing error so fundamental that it guarantees nothing gets fixed. If you're mad about job displacement, that's a labor market and corporate governance problem. If you're mad about children's cognitive decline, that is a deployment and education policy problem. If you're mad about billionaires shielding their kids while you're scroll TikTok, that's a resource inequality problem. If you're mad about your electricity bill, that's an infrastructure planning and public consultation problem. None of these are solved by being mad at a neural network. None of them are solved by being anti-AI. And the longer the anger stays pointed at the technologies instead of the systems that deploy it, the longer every single one of these problems persists. As a computer scientist, it pains me deeply genuinely that AI is being conflated with all of this because AI is extraordinary. It is all inspiring. The research breakthroughs happening right now are some of the most beautiful intellectual achievements of our time. And every time someone says I'm anti- AAI because what they actually mean is I'm angry about my job or I'm angry about my kids attention span or I'm angry about my electricity bill. The actual problems drift further from being solved because you've named the wrong enemy. And naming the wrong enemy is the most reliable way to lose. Getting mad feels good, but it doesn't do anything. Once you figured out what specifically you're angry about, and that would be a genuinely valuable exercise, you need a plan. You need a structural reform, you need policy, you need strategy. Anger without strategy is just noise.

And this is where we arrive at the most dramatic expression of public anger towards the AI industry so far. A national survey of 1,690 US adults conducted by research firm Veracite in June 2026 found that 69% of respondents support forcing AI companies to transfer 50% of their stock into a public sovereign wealth fund. In June, Senator Bernie Sanders proposed the American AI Sovereign Wealth Fund Act, which would mandate exactly this, a one-time seizure of 50% of stock from the largest AI companies in the United States. At current valuations, Standards estimated the fund would be worth approximately 7 trillion. Okay, so that's a number that sounds extraordinary and the public support 69% is really remarkable for a policy this aggressive. Meanwhile, 89% support mandatory safety disclosures. The tech sector accounted for nearly a third of all US layoffs in the first half of 2026. Over 166,000 positions eliminated with projections suggesting 312,000 by year end. That's a lot of people. Goldman Sachs has estimated that more than 15 million US jobs could be displaced across a decadel long AI transition. The anger is backed by job losses, rising costs, and a growing sense that the benefits of AI are flowing up while the costs flow down. Now, the political response to this has been nothing short of spectacular. Senator Bernie Sanders proposal to seize 50% of AI company stock is the kind of aggressive portfolio management usually only seen in highstakes corporate rates or maybe pirate movies. It's an incredibly satisfying headline, provided you don't look at the balance sheets of the companies you're actually rating. I understand the sentiment completely, and I need to be honest about this policy. What wealth are we talking about exactly? OpenAI lost $20.9 billion operationally in 2025. Anthropic is not profitable. The majority of AI company valuations are built on venture capital funding and angel investment, not revenue. These are companies that are spending $1.60 for every $1 they earn. Seizing 50% of their stock means the public gets 50% of shares in companies that are right now today losing money at historic rates. The fund wouldn't hold $7 trillion in wealth. It would hold $7 trillion in valuation, which is a very different thing. Valuations can collapse, and they frequently do. Seizing half the equity of a company that loses $20 billion a year is a masterclass in financial massochism. You're not redistributing wealth. You're nationalizing a cash burn. Is the sovereign wealth equivalent of breaking into your neighbor's house to steal half of their outstanding credit card debt because you heard they have a very high credit limit. Technically, you own a massive asset pool, but it consists entirely of invoices. And let's just pause for a second here. If the goal is wealth redistribution, if the genuine legitimate concern is that the profits of technological progress should be shared more broadly, then why start with AI companies that don't have profits? Huh? The S&P 500 is full of corporations generating enormous, real, tangible profits. Apple made $97 billion in net income last year. Why aren't we seizing 50% of that? The answer is that the political energy is directed at AI specifically because people are angry at AI specifically which means the pull is measuring anger not policy preference. We actually have a mechanism for ensuring that profitable companies contribute to society. It's called tax. Corporate tax, capital gains tax, dividend tax. These are the structural tools designed to redistribute the gains of commerce. If the tax system isn't working, and there are strong arguments that it isn't, then reform the tax system. A 50% stock seizure is not taxation. It's partial nationalization of companies that are mostly unprofitable, funded by venture capital, and whose valuations could evaporate in a downturn. The government should not be making investment decisions with that kind of risk appetite because it doesn't represent everyone's risk appetite. That is a gamble with public money. And if the gamble goes wrong, if these companies fail to reach profitability, if the AI bubble deflates, if the valuation's correct, the public isn't holding a sovereign wealth fund, the public is holding 50% of the losses.

The public anger again is completely legitimate. The desire for accountability is definitely correct. This specific mechanism is, in my opinion, the wrong tool. And it's the wrong tool because the anger is aimed at the wrong target. Consider what would actually happen if the US government seized 50% of OpenAI stock tomorrow. It would own half of a company that loses $21 billion per year. If these companies eventually do become profitable, which most independent analysts don't expect before 2029 at the earliest, the fund might actually generate returns. If they don't, the public holds paper in failed companies. and the precedent. A government seizing half of private company's equity would almost certainly trigger capital flight. If you tell every AI startup that the government might take half of their stock, investment moves offshore immediately. The companies you're trying to hold accountable, relocate to jurisdictions that won't touch them. The AI development doesn't stop. It just happens somewhere with fewer protections, less oversight, and no public fund at all. Meanwhile, the 89% who support mandatory safety disclosures, that is a policy that could work tomorrow without seizing a single share. Transparency requirements, auditing standards, independent evaluation of AI systems before deployment. These are the tools that address the actual problems without the risks of partial nationalization. But mandatory safety disclosures doesn't make for as satisfying a headline as 7 trillion sovereign wealth fund. And that's the problem with anger-driven policy. The most emotionally satisfying proposals are rarely the most effective ones.

Let me come back to the billionaires and their children for just a moment because I think the more generous interpretation is important and it actually strengthens the argument rather than weakening it. I don't think Tal and Gates and Jobs necessarily had perfect insight into the long-term cognitive effects of screen time. A CEO who builds a technology doesn't automatically possess the equivalent of 20 years of clinical trial data on its developmental impact. Just like a pharmacist can formulate a drug without having a full understanding of every long-term side effect. Right? Knowledge comes from research and the research takes time and the research on screens on child development is still maturing. What these billionaires did have and what most parents do not is the resources to make a different choice. A parent working a 10-hour shift hands their child an iPad because they need 30 minutes of peace. It's not bad. Parenting is just exhaustion. That's a person trying to survive their day. Teal can just instruct a nanny to handle it. Gates can fund a school. Jobs could design an entirely different child school for his kids, one without the products that made him the wealthiest person in tech. Blaming an exhausted parent for giving their child an iPad after a 10-hour shift is a bit like blaming a stranded traveler for drinking rain water because it hasn't been properly filtered through a Britta filter. They're not trying to optimize the child's cognitive architecture for the year 2045. They're just trying to negotiate a highly volatile 20-minut ceasefire so they can just cook dinner without the kitchen burning down. The 1,600 plaintiffs, 350 families, and 250 school districts who filed lawsuits alleging that Meta, Snap, Tik Tok, and YouTube created addictive platforms leading to depression and self harm in children didn't have the resources to shield their children. They were fighting the same platforms that the founders children were quietly protected from. Australia and Malaysia have now banned children under 16 from social media. France, Denmark, and the UK are considering similar legislation. These are governments trying to do at the policy level. What billionaire parents did at the high school level years ago, limit exposure to technology that they suspect is causing harm. The difference is that the governments are doing it for everyone, not just for their own children. The hypocrisy is that they can afford to opt out of the consequences while the rest of the population absorbs them. The same technology that rich parents can afford to avoid is the technology that exhausted parents depend on. And the anger that produces, which is completely valid, gets directed at AI and tech and screens when it should really be directed at the systems that created the resource gap in the first place. At the labor conditions that leave parents too exhausted to supervise their children's screen time. At the education system that spent $30 billion on laptops without evidence they would help. at the corporate governance structures that allow companies to fire thousands while posting record profits. At the tax system that allow the wealthiest people on earth to accumulate resources so vast they can insulate their families from the consequences of the very products that made them wealthy in the first place.

As a computer scientist, this conflation wounds me. AI is a beautiful field. The mathematics underneath it are elegant. The research breakthroughs are genuinely all inspiring. The potential applications in medicine, in accessibility, and scientific discovery fill me with a kind of joy that I suspect only nerds fully understand. And every time AI gets used as a shorthand for venture capitalism, labor exploitation, asset inequality, and corporate recklessness, the actual technology, the actual science gets buried under grievances it did not cause and cannot solve. AI didn't create wealth inequality. AI didn't design the attention economy. AI didn't decide to spend $30 billion on laptops without checking whether they worked. AI didn't choose to fire 166,000 people in 6 months. People did. Systems did. Incentive structures did. And until we start naming the actual causes and building actual strategies to reform them, nothing changes. No matter how angry we get, what people are asking for, whether they articulate it this way or not, is societal change. And AI is the catalyst for that conversation. But catalysts are not causes. This kind of reckoning about labor, about inequality, about the responsibilities of powerful institutions would have arrived with or without artificial intelligence. The labor market was already under strain. Wealth inequality was already obscene. Education policy was already failing children. AI accelerated all of it, made it visible, gave it a name and a face, but the face isn't the cause. The structural conditions are the cause. And the structural conditions need structural responses. Tax reform, not stock seizure. Education policy grounded in evidence, not in sales pitches from technology companies. Labor protections that don't evaporate every time a co discovers automation. corporate governance that holds executives accountable for the consequences of their deployment decisions, not just their quarterly earnings. Public consultation requirements for infrastructure projects that affect communities. These are not radical proposals. I'm not suggesting something crazy here. They're the boring, difficult, incremental work of building systems that distribute the benefits of progress more fairly. They don't make for good protest slogan. They don't make for good protest signs, but they do make for good policy. And the anger, and once again, I have to be clear, the anger is earned. I would say this, figure out what you're specifically angry about because that clarity is genuinely valuable in this case. But once you figured it out, you need a plan, not a slogan. Getting mad doesn't change systems. Strategy changes systems. And the most important strategic insight available right now is this. AI is a hammer. It can build hospitals or it can break windows. The hammer didn't choose. Somebody is swinging it. Go find them. Stop being mad at the hammer. Start asking who handed it to the person swinging it, who is profiting from the swing, and what it would take to put it in bitter hands. If you want to understand how the hammer is being swung at specific companies, and why 55% of executives already target replacing their workers with AI, I cover that in this other video that I'm linking on your screen here. Thanks so much for watching. I'll see you in the next one.