American Intelligence
Let’s start with politics. If only because I think the timing of “the gap” (which we’ll get into) falls right along the next presidential election.
I think the next election isn't going to be about the usual stuff. I mean, sure, there will be arguments about immigration and taxes and whatever culture war is trending that week. But underneath all of it, there's going to be one question that drowns everything else out:
What do we do when AI takes the jobs?
Not the factory jobs. Not the truck driving jobs. Those have been the hypothetical punching bags for a decade. I'm talking about the good jobs. The ones your parents told you were safe. The ones you went $200K into debt for a degree to get. Lawyers. Accountants. Radiologists. Financial analysts. Software engineers. Marketing directors. The jobs that were supposed to be the whole point of "doing everything right."
And I think we're going to see candidates split on this in a way that makes the old Right vs. Left thing feel outdated. One candidate is going to run on protecting humans. Slowing it down. Regulating AI. Keeping people in jobs even if it's less efficient. The other is going to argue that acceleration is the only path forward (that fighting AI is like fighting electricity, and the faster we get through the painful part, the faster we get to the good part on the other side).
Both of them will have a point. And that's what makes this spicy..
The concept of a gap is what I keep coming back to. There's going to be this window (maybe it's already starting) where unemployment surges in a way that looks nothing like any recession we've seen before. Not because the economy is "bad" in the traditional sense. Not because of a housing bubble or a pandemic or a financial crisis. But because a piece of software got good enough to do your job for $20/month.
And here's the thing that makes this particular moment uniquely brutal: it's going to hit during the largest concentration of people in their prime earning years that we've ever had.
Millennials (74 million of them) are now firmly in their 30s and 40s. Peak career. Peak spending. Peak obligations. Gen X (65 million) is right behind them in their 45-59 range, which is historically when earnings top out. Morgan Stanley's research shows that this overlap of prime working-age adults is projected to accelerate into the 2030s. We've literally never had this many people simultaneously in the phase of life where they have mortgages, kids in school, car payments, and the highest financial obligations they'll ever carry.
Now imagine telling those people (tens of millions of them) that the thing they spent 15 years building expertise in can now be done by a chatbot.
These aren't people who can just "pivot." A 42-year-old marketing VP with two kids in private school and a $4,500 mortgage can't exactly go back to community college and "learn to code" (which, by the way, is also being automated, so that advice aged well). They're locked in. They have obligations that were built on the assumption that their income would keep going up, or at least stay flat.
And the cruelest part? Previous economic disruptions always had a retraining runway. Manufacturing declined over decades. You could see it coming and adjust. AI displacement might happen in years (months?), not decades. And the target keeps moving (because AI keeps getting better). So what do you retrain into when the thing you'd retrain into is also six months from being automated?
So how can we politically prepare for this?
Let's say the political response to mass AI unemployment is some version of "we'll create programs to early-retire displaced workers" or "we'll expand the safety net." Sounds reasonable, right? There's just one problem.
We can't afford the safety net we already have.
Social Security's combined trust funds are projected to be depleted by 2035. That's not some far-off hypothetical (that's nine years from now). When that happens, every single beneficiary faces an automatic 21% cut in benefits. Not means-tested. Not phased in. Everyone gets cut. That number grows to 31% by the end of the 75-year projection window (CRFB 2025 Trustees Report)
Right now, today, Social Security is running a cash flow deficit of $250 billion per year. Over the next decade, the program will run $3.6 trillion in total deficits. Medicare's hospital insurance trust fund can pay full benefits only through 2036.
To put it in perspective: fixing Social Security's shortfall alone would require either a 29% increase in payroll taxes, or a 22% benefit cut for all current beneficiaries. And that's just to keep the existing system solvent for people retiring on schedule.
Now imagine adding millions of 45-year-olds who just got replaced by GPT-7 and need some kind of early retirement program or universal basic income on top of that.
The math doesn't math.
Every candidate in the next election is going to be scrambling for ways to support displaced workers. They'll talk about retraining programs and transition funds and dignity-of-work initiatives. But the money isn't there. Which means one of two things has to happen: massive new taxes (probably on AI companies and the businesses that use AI to replace workers), or we just... print more money and add to the pile.
Speaking of the pile.
The national debt just hit $38.5 trillion (JEC). It's growing at about $8 billion per day (that's $93,000 per second) (JEC). The debt-to-GDP ratio is sitting at about 121% (Financer.com), and the CBO projects it'll hit 156% by 2055 (AAF). Interest payments alone now exceed $1 trillion per year (Fortune), making debt service the fastest-growing line item in the federal budget. We're paying more in interest than we spend on defense.
One of the core promises of the AI acceleration thesis goes something like this: AI will make everything more efficient. Energy gets cheaper. Production gets cheaper. Raw materials get cheaper (because mining, logistics, and manufacturing are all optimized by AI). Eventually, the cost of most goods and services approaches something closer to zero. And when that happens, living on a government stipend of $2,500/month isn't just survivable (it's comfortable). Maybe even good.
I get this. Maybe not the timeline (more on that in a second), but the direction. If AI can make a kilowatt-hour of energy cost a fraction of a penny, if it can optimize food production to the point where groceries cost almost nothing, if it can design and build housing for a tenth of current costs (then yeah, $2,500/month starts to look like a pretty decent life).
But (and this is the "but" I keep thinking about): if the cost of everything goes down, doesn't GDP go down too? And if GDP goes down, and our debt stays at $38 trillion and climbing, how the hell do we ever pay it off?
This isn't a rhetorical question. It's an actual economic paradox that I haven't seen anyone in the AI optimist camp address seriously.
GDP is measured in nominal dollars. If AI-driven deflation causes the price of goods and services to drop significantly, nominal GDP shrinks (even if the real output of the economy, the actual stuff being produced and consumed, is exploding). But our debt is also denominated in nominal dollars. $38 trillion is $38 trillion regardless of whether a gallon of milk costs $5 or $0.50.
Historically, the way the U.S. has managed its debt is by inflating it away. You borrow a trillion dollars today, and over 20 years, inflation erodes the real value of that debt. Your GDP grows in nominal terms, the ratio improves, and you can service the payments. It's not pretty, but it works.
Deflation does the exact opposite. It makes the debt heavier in real terms. Every dollar you owe becomes more valuable, not less. The IMF's research on historical deflation episodes confirms this (deflation consistently worsens debt-to-GDP ratios and makes government revenues harder to collect).
So we're potentially facing a world where the economy is producing more abundance than ever before, but the financial system (which is built entirely on the assumption of perpetual moderate inflation) breaks down. That's not a minor technical problem. That's a fundamental contradiction at the heart of the AI utopia thesis.
But here's where I need to be fair, because there's a counter-argument that's more than just hand-waving. And honestly, the more I dig into it, the more compelling it gets.
What if GDP doesn't shrink? What if it explodes?
Here's the thing about economic history that the deflation doomers (including me, three paragraphs ago) tend to forget: every single time technology has made existing goods cheaper, humans have invented entirely new categories of things to spend money on. And those new categories have always outpaced the deflation in the old ones.
Think about clothing. The cost of basic apparel has plummeted over the last century (a t-shirt costs essentially nothing compared to what your great-grandparents paid). But did the "clothing economy" shrink? No. It spawned a $1.7 trillion global fashion industry (Statista). The commodity got cheap. The experience got expensive. Food tells the same story. Groceries got cheaper (adjusted for inflation, Americans spend less on food at home than almost any point in history). But restaurants became a $1 trillion industry (NRA). We didn't stop spending on food; we moved up the ladder from "sustenance" to "experience." Music is another good example. It went from $20 CDs to essentially free streaming. The music recording industry collapsed. But the live music industry exploded to $40 billion (CMI). Podcasts created a $25 billion market that didn't exist 15 years ago (PBS). The format died; the category thrived.
Economists call this the "quality ladder" effect, and it's happened with literally every technological disruption in history. Basic goods deflate, but humans keep climbing the ladder toward newer, weirder, more premium things to want. We are extraordinarily good at inventing new desires.
So apply this to AI. Yes, the cost of a basic legal contract goes to zero. But what about an AI-designed, continuously-updated, self-enforcing smart contract that manages your entire financial life? That didn't exist before. It's a new product, a new market, a new slice of GDP. The cost of a standard eye exam goes to zero. But what about the neural-linked bionic eye with augmented reality overlay and infrared vision? That's not cheaper healthcare (that's an entirely new product category that could support a multi-billion dollar industry). A regular X-ray costs nothing. But the AI-powered full-body molecular scan that catches diseases a decade before symptoms appear? That's a premium product for a market that doesn't currently exist.
And then there are the frontiers that are just... new. The space economy alone is projected to hit $1.8 trillion by 2035 (according to McKinsey and the World Economic Forum), potentially reaching over $2 trillion (McKinsey/WEF). That's asteroid mining, orbital manufacturing, space tourism, lunar resource extraction, space-based solar power. Almost none of this exists as a real economy today. AI is the enabling technology for nearly all of it.
The MIT economist David Autor found something remarkable: 60% of employment growth over the last 80 years came from jobs that literally didn't exist at the start of that period (MIT News). Not jobs that evolved. Jobs that were invented. "Social media manager" wasn't a job in 2005. "Cloud architect" wasn't a job in 2010. "Prompt engineer" wasn't a job in 2022. If that pattern holds (and it has held through every technological revolution in modern history) then AI won't just automate existing work. It'll create entirely new categories of economic activity that we genuinely cannot imagine yet. Just like nobody in 1995 could have predicted that "influencer" would be a career path, nobody in 2025 can predict what the economy looks like when AI is a utility.
BlackRock is calling this "deflationary growth" (GDP rising while costs compress simultaneously) (BlackRock). They argue it might be the first time in history where rising output doesn't require rising input costs. If that's true, it could actually help the debt situation: the denominator in the debt-to-GDP ratio grows because new economic value is being created, even while the cost of existing goods falls. Goldman Sachs projects a 7% global GDP boost and 15% U.S. labor productivity increase boost from full AI adoption (GS Global GDP, GS Productivity)
So which is it? Does AI-driven deflation crush the economy under the weight of $38 trillion in nominal debt? Or does AI-driven innovation create so much new economic value that GDP grows faster than the deflation in existing goods?
The honest answer is: nobody knows. And that's exactly the kind of uncertainty that makes for an interesting transition period. Both forces are real. Both are powerful. And we're about to find out which one wins out.
Right in the middle of all this chaos (the AI displacement, the retirement crisis, the debt spiral) America is about to experience the largest transfer of wealth in human history.
It's called the Great Wealth Transfer, and the numbers are staggering. Baby boomers and the Silent Generation collectively hold about 64% of the nation's $190 trillion in wealth (Fortune). Over the next two decades, they're expected to pass down somewhere between $84 trillion and $124 trillion to their kids and grandkids (estimates keep getting revised upward as asset prices climb) (Cerulli). Gen X is set to inherit about $39 trillion, and Millennials are on track to receive roughly $46 trillion (Merrill Lynch)
The timing is almost poetic. Just as working for a living might become optional (or impossible, depending on your perspective), a whole generation is about to receive the biggest financial windfall in history. Many of the people getting displaced by AI might simultaneously be getting a check from Grandma's estate that covers their mortgage for the next decade.
But I'd be lying if I presented this as some kind of clean solution. Because the wealth transfer is wildly, almost comically, unequal. The wealthiest 10% of households will give and receive the vast majority of that money. The top 1% alone accounts for about 42% of the expected transfers (Wikipedia). A third of Millennials expect an inheritance, but only about a fifth of Boomers say they actually plan to leave one. And healthcare costs in those final years have a habit of eating through savings faster than anyone expected. And didn’t we mention something about Social Security being gone soon?
So for the software engineer making $180K who just got laid off? Maybe she's got a trust fund coming. But for the accounts payable clerk making $55K who also just got replaced? Probably not. The wealth transfer could actually widen inequality during the transition, not narrow it. The people who need the cushion most are the least likely to have one.
Still (at a macro level) the fact that trillions of dollars are shifting to younger generations right as the economy restructures around AI is... something. It's not a solution.
We already ran a small-scale experiment on what happens when people's basic needs are met without working. It was called COVID.
During the pandemic, the federal government added $600 per week (later $300) on top of regular unemployment benefits. The result? The median UI recipient was making 145% of their previous wages (NBER). Over 75% of unemployed workers were earning more on unemployment than they had at their jobs (Congress.gov/NBER). And what happened next became one of the most politically charged economic debates in a generation.
47 million Americans quit their jobs in what got dubbed the "Great Resignation" in 2021(Wikipedia). The labor force participation rate dropped from 63.4% to 60.2% at its trough and settled around 61.5% for much of 2021 (Richmond Fed). Workers 55 and older essentially never came back. 26 states cut benefits early to try to force people back to work. The states that cut early did see some return (unemployment-to-employment flows jumped 14-15 percentage points) (NBER Working Paper 29575). But the overall effect was smaller than almost everyone predicted.
The popular narrative was simple: "lazy people got paid to stay home." But the actual research tells a more interesting (and more relevant) story. The NBER confirmed that benefits slowed the return to work, but researchers consistently found the disincentive effects were "smaller than expected." It wasn't just about the money. Workers in service and hospitality described 80-hour weeks, constant risk of harassment, during a pandemic, for wages that had stagnated for 30 years. When given a moment to breathe, millions of people did something the economic models didn't predict: they fundamentally reassessed whether the deal they'd been offered (trade your life for a barely-livable wage) was actually worth taking.
If COVID was the appetizer. AI is the main course.
The pandemic gave a relatively small subset of workers, mostly in lower-wage service jobs, a temporary taste of meeting basic needs without working. The effects rippled through the economy for years. Now imagine that scenario, but permanent, and hitting the professional class, and at a scale 10 times larger.
But there's another side to the COVID experiment that's just as important, and it reveals a gap within the gap.
Some people used the pandemic pause to start businesses, learn new skills, spend time with family, rethink their priorities. The rate of new business applications hit an all-time high in 2021 (Census Bureau). Others... didn't. Some people genuinely thrived without traditional employment. Others felt lost, purposeless, and depressed. The pandemic was a Rorschach test for how humans handle a sudden absence of mandatory work, and the results were deeply mixed.
That divide (between people who flourish without structured work and people who flounder) is going to be one of the most important and least discussed dimensions of the AI transition. Because it's not just about whether we can provide for people financially. It's about what happens to people psychologically when the thing that structured their entire adult life disappears.
There is one potential bright spot in the displacement story, at least in the short term: if AI is coming for white-collar knowledge work first, there's going to be a surge in demand for people willing to work with their hands.
Think about it. Somebody has to install, maintain, and repair the physical infrastructure that AI runs on. Data centers need electricians and HVAC technicians. The robotics revolution needs people who can service, calibrate, and troubleshoot actual machines. Solar panels and wind turbines need to be physically mounted. AI-designed buildings still need to be built (for now). The trades (plumbing, electrical, welding, heavy equipment operation) are among the last jobs that require a physical human presence, and they're already facing massive labor shortages.
For a displaced marketing VP or financial analyst willing to swallow their pride and pick up a wrench, there's a genuine window of opportunity here. The pay in skilled trades is often better than people assume, the demand is enormous, and AI can't (yet) send a robot to fix your furnace at 2am.
But I'd be dishonest if I didn't flag two problems with this narrative.
Problem one: history tells us most people won't make that leap. The track record on retraining is, frankly, dismal. The U.S. Government Accountability Office identified 47 different federal employment and training programs and concluded that "little is known about the effectiveness" of most of them (GAO-11-92). Multiple evaluations of federally funded displaced worker training programs found they did not increase employment, income, or the likelihood of receiving benefits. The Trade Adjustment Assistance program, specifically designed for workers displaced by foreign competition, actually left many participants earning less after completing it (Heritage Foundation). Brookings noted that the U.S. spends just 0.11% of GDP on workforce development (a fraction of what other developed nations invest) and that "the U.S. doesn't do a good job, under any circumstances, of helping people make those transitions." (Hechinger Report) When given the chance during COVID, many displaced workers chose not to retrain into new fields at all (they waited, they collected benefits, they hoped their old jobs would come back). The psychological barrier of going from a corner office to a construction site is enormous, even when the economics make sense.
Problem two: the blue-collar bridge has an expiration date. Humanoid robotics is advancing faster than most people realize. Goldman Sachs projects global shipments of 50,000-100,000 humanoid robot units in 2026, with costs dropping toward $15,000-$30,000 per unit (Goldman Sachs). Tesla's Optimus, Figure AI, Agility Robotics, and a wave of Chinese competitors are all racing to commercialize general-purpose humanoid robots for manufacturing, logistics, and eventually household tasks. Citigroup predicts 1.3 billion robots in operation by 2035 (Citi GPS). The blue-collar jobs that seem safe from AI today might have a 5-10 year runway before humanoid robots start competing for those too.
So the opportunity is real, but it's a bridge, not a destination. And if the retraining data tells us anything, it's that most displaced workers won't cross it voluntarily (especially not the credentialed professionals who've spent their entire lives being told that manual labor is what you do if you can't get a desk job).
I want to zoom out for a second and talk about something that goes beyond economics. Because I think the financial stuff, as scary as it is, might actually be the easier part of this transition.
The harder part is meaning.
American identity is built on work. It's the first thing we ask strangers at a party. "So, what do you do?" Your job isn't just your income (it's your identity, your social status, your daily structure, your sense of purpose). We literally define ourselves by our occupations. Saying "I'm a lawyer" or "I'm an engineer" isn't just a description of how you earn money. It's a statement about who you are.
Now imagine telling 30 million people (accomplished, credentialed, hard-working people who did everything they were supposed to do) that they're not that thing anymore. Not because they failed. Not because they weren't good enough. But because a large language model got good enough to do it faster, cheaper, and (let's be honest) probably better.
That's not just unemployment. That's an identity crisis at a generational scale.
We've already seen what happens when communities lose their economic purpose (the opioid epidemic, deaths of despair, the hollowing out of manufacturing towns across the Midwest). Those were concentrated in specific regions and demographics. AI displacement will be everywhere, and it'll hit the educated professional class that has largely been insulated from economic disruption for their entire lives.
The mental health implications alone are something nobody seems to want to talk about. What happens when millions of type-A achievers (the people who defined themselves by their performance reviews and their titles and their corner offices) suddenly have nothing to perform at?
(This is the part where someone usually says "people will find meaning in hobbies and relationships and community!" And sure, eventually, some will. But the transition from "I'm a senior VP" to "I'm really into pottery now" is not a smooth one. It's going to be ugly and painful and filled with a lot of people feeling lost in ways they never prepared for.)
But here's something worth sitting with: work-as-identity is actually a pretty recent invention.
The ancient Greeks would have found our obsession with work bizarre (they considered labor something to be avoided so you could pursue philosophy, athletics, and civic life). Renaissance Europeans defined themselves by family lineage, faith, and intellectual curiosity, not their trade. The idea that your job is your identity is really only about 150 years old, a product of the Industrial Revolution and the rise of corporate culture. "What do you do?" as the first question at a party is an industrial-era invention. We just don't realize it because it's all we've ever known.
Which means the current relationship between work and identity isn't some eternal human truth. It's a cultural artifact. And cultural artifacts can change. They have before. They will again. The question is whether this generation can make that shift, or whether it takes a generation that never knew the old model in the first place.
Here's what I think makes the next election truly unprecedented: the traditional political coalitions are going to shatter.
Think about it. The "protect human workers" candidate would normally be a progressive, union-backed Democrat, right? But AI displacement is going to hit lawyers, accountants, radiologists, financial analysts, marketing executives (these are historically Republican-leaning professionals). Suburban dads who voted for lower taxes their whole lives are suddenly going to find themselves on the same side as union autoworkers, both asking the same question: "Why is the government letting a machine take my job?"
Meanwhile, the "accelerate AI" candidate could come from either side. Silicon Valley has traditionally leaned left, but the accelerationist techno-optimist crowd (the "e/acc" people) skews libertarian. You could see a coalition of tech billionaires, free-market Republicans, and national security hawks (more on that in a second) all arguing that slowing down AI is the real threat.
And speaking of national security (this is the argument that makes the "pro-human" position really, really hard). Even if the U.S. wanted to slow down AI adoption to protect workers, China won't. The CCP isn't having a national conversation about displaced middle managers. They're going full speed ahead. So any candidate running on "slow down AI" has to answer a very uncomfortable question: are you willing to let China win the AI race to save American jobs in the short term?
That's not a hypothetical gotcha. That's a genuine, legitimate tension with no clean answer.
And if it feels like a hypothetical, it shouldn't. Because we already know exactly what happens when the U.S. falls behind a cheaper competitor. We watched it play out with manufacturing, and the scars are still fresh.
Economists call it the "China Shock." When China joined the World Trade Organization in 2001, American manufacturing got hit by a wave of cheaper imports that fundamentally restructured the economy. The numbers are still debated, but the landmark research by MIT's David Autor and colleagues found that Chinese import competition accounted for roughly 59% of all U.S. manufacturing job losses between 2001 and 2019 (Autor/Dorn/Hanson via USITC). The Economic Policy Institute puts the total displacement at 3.7 million jobs (EPI). Nearly five million manufacturing jobs disappeared between 1998 and 2019 (CNBC). The computer and electronics sector alone lost 1.2 million jobs (EPI).
But here's the part that should scare anyone thinking about AI: the workers didn't bounce back.
The research is devastating on this point. Manufacturing job losses caused by the China Shock converted "nearly one for one into long-term unemployment." (NBER working paper) Workers didn't shift to new sectors. They didn't move to booming cities. They stayed in their hollowed-out communities, and those communities fell apart. The effects persisted for nearly two decades after the initial shock. Incomes dropped. Government transfer payments went up but couldn't offset the losses. Family formation declined. And then came the deaths of despair. Research published in the American Economic Review found a direct link: areas more exposed to Chinese trade competition saw significantly more fatal drug overdoses among working-age adults. One study estimated that up to 92,000 male and 44,000 female overdose deaths between 1999 and 2017 were predicted by the decline of manufacturing alone (ScienceDirect). The opioid epidemic didn't just happen to coincide with manufacturing collapse (it grew directly out of the economic hopelessness left behind). Over a million Americans have died from suicide, drug overdose, or alcohol-related causes since 2006, disproportionately concentrated in exactly the communities that lost their economic purpose (Brookings)
Economists had assumed the labor market was dynamic enough to absorb the disruption. They were wrong. The human costs were staggering and the policy response was completely inadequate.
Now apply that pattern to AI. But instead of furniture factories and textile mills, imagine its law firms and accounting departments and radiology practices and software companies. Instead of blue-collar workers in the Rust Belt, it's white-collar professionals in every zip code in America.
China has explicitly declared its goal of becoming the global AI leader by 2030. Their 2017 AI Development Plan laid it out in black and white (DigiChina), and they've been backing it with massive state investment, talent pipelines, and a military-civil fusion doctrine that turns every commercial AI breakthrough into a potential defense capability. DeepSeek's release of its R1 model in January 2025 was a wake-up call (a model matching OpenAI's performance at a fraction of the cost, built despite U.S. chip export controls). A RAND report found that Chinese AI models now run at roughly one-sixth to one-fourth the cost of comparable American systems.
Here's the scenario that nobody in the "slow down AI" camp wants to confront: if the U.S. pumps the brakes on AI development to protect American knowledge workers, and China doesn't, we don't just lose a technology race. We get a replay of the China Shock, except this time it hits the professional class.
Think about it. If Chinese AI becomes the global standard (cheaper, widely available, open-source) then companies around the world start using Chinese AI tools instead of American ones. The legal work that was going to be automated by American AI gets automated by Chinese AI instead. The software development, the financial analysis, the medical diagnostics, all of it flows through Chinese platforms. Except now, instead of that economic activity showing up in U.S. GDP (funding tax revenue, supporting the programs we desperately need to help displaced workers) it shows up in China's GDP.
That's the brutal irony at the center of the "pro-human" political position: the jobs are probably going away regardless. The question is whether they get replaced by American AI (which at least keeps the economic value onshore, funds the tax base, and gives us the resources to manage the transition) or by Chinese AI (which doesn't). Slowing down doesn't save the jobs. It just ensures that when those jobs disappear, the economic benefits flow to Beijing instead of Washington.
The China Shock destroyed millions of manufacturing jobs and we spent two decades failing to help those communities recover. An AI China Shock could do the same thing to knowledge work, except faster, broader, and with even less warning. And this time, the displaced workers won't just be in specific factory towns. They'll be in every city, every suburb, every professional office park in America.
That's why the national security argument isn't just about military AI or autonomous weapons (though that matters too). It's about whether the United States maintains the economic engine it needs to fund its own transition. You can't build a bridge across the gap if someone else owns the construction equipment.
Okay. I've been pretty doom-and-gloom, so let me talk about what happens if we actually get through the gap.
Because the thing is (and I do believe this) what's on the other side could be genuinely extraordinary.
If AI delivers on even half of what the optimists promise, the world looks something like this: Energy is so cheap it's practically free (advanced solar, fusion, AI-optimized grids). Healthcare is radically better and cheaper (AI diagnostics catch cancer years early, robotic surgery is precise and affordable, drug discovery happens in weeks instead of decades). Food production is optimized to the point where hunger is a logistics problem, not a scarcity problem. Housing costs plummet because AI-designed, robotically-built homes can go up in days. Education is personalized, lifelong, and essentially free.
In that world, $2,500/month from the government isn't scraping by. It's a comfortable life. Maybe even a good one. You're not working because you have to (you're free to work on what you want to). Start a small business (AI handles the boring operational stuff for you). Make art. Coach your kid's baseball team. Write the novel you've been thinking about for 15 years. Travel. Learn a new language just because you're curious. Actually be present for your family instead of grinding out 60-hour weeks to afford a life you're too busy to enjoy.
The barrier to starting something drops to nearly zero. Want to open a restaurant? AI handles your supply chain, your books, your marketing, your health code compliance. You just cook. Want to build an app? Describe it in plain English and AI builds it for you. The creative and human parts of any endeavor become the only parts that matter.
That's not a fantasy. That's a plausible extrapolation of where the technology is heading. The question isn't whether we get there. The question is what happens in the 5 to 15 years between here and there.
And that's really what this whole thing is about. The gap.
The gap between "AI takes your job" and "AI makes your life amazing." The gap between "the economy as we know it breaks" and "a new economy that's genuinely better emerges." The gap between "you can't afford your mortgage" and "you don't need to worry about money."
I have two boys. They're 3 and 5. And I think about the gap constantly through the lens of their lives.
By the time my oldest leaves the house, it'll be 2038. By the time my youngest does, it'll be 2040. Right in the window where, if the optimists are even half right, the other side of the gap starts to become real. The deflation in goods is well underway. The new industries are forming. The space economy is hitting its stride. The meaning of "work" has already been redefined by a generation that never had the old model.
Here's what I keep coming back to: the chaos of the gap, the part we're about to live through, won't feel chaotic to them. It'll just feel... normal. The way the internet felt disorienting to our parents but was just the water we swam in. My kids won't mourn the loss of the 9-to-5 because they'll never have expected it in the first place. They won't have an identity crisis about not being a "senior VP" because that was never going to be how they defined themselves.
Maybe they define themselves by what they create (not for a company, but for the joy of it). Maybe "what do you do?" gets replaced by "what are you working on?" or "what are you curious about?" or just... doesn't get asked at all. Maybe hitting the gym on a Wednesday at 11am, playing nine holes in the afternoon, volunteering at a food bank, learning Mandarin because it sounds cool, coaching your kid's soccer team (maybe that's not a consolation prize for losing your career). Maybe that's just... a life. A genuinely good one. One that the Greeks would have recognized instantly even if we can't quite see it yet from where we're standing.
I think my kids are going to be fine. I think they're going to live in a world that's wealthier, healthier, and freer than anything we've experienced. But I also think the next 10-15 years (the years where we are navigating the gap) are going to be profoundly difficult in ways nobody is adequately preparing for.
That gap is going to be the defining political, economic, and social challenge of the next decade. And here's what scares me most about it: the people in charge of navigating us through it (the politicians, the policymakers, the business leaders) are themselves products of the old system. They got where they are by mastering rules that are about to stop applying.
We're going to need leaders who can hold two contradictory ideas in their heads at the same time: that AI displacement is a genuine crisis that could cause real suffering, and that trying to stop AI is both impossible and undesirable. Leaders who can build a bridge across the gap without pretending the gap doesn't exist.
I don't know if those leaders exist yet. But the election is coming, and we're about to find out.
The only thing I'm certain of is this: "left vs. right" isn't going to cut it anymore. The real divide might be "human vs. machine" (and the most unsettling part is that the right answer might be both).
If you're reading this and feeling anxious about your own career, your own mortgage, your own kids' future (good). That means you're paying attention. The worst thing any of us can do right now is assume this doesn't apply to us. It applies to almost everyone. The best thing we can do is start talking about the gap honestly, without pretending we have answers we don't have.
Because the gap is coming. And the only question is whether we build a bridge or fall in
