Now these all seem quaint and arbitrary. The space on them was almost laughable — 128 kilobytes, even if the box said “1 Mega.” And any “real gamer” had two of them.

How will the world change? Answering this question with any accuracy provides for a real, measurable gain — there are entire shops that spend billions of dollars every year to answer “what will happen next?” This is true for customer actions, for the stock market, for technology and for governments. And still, as far as I can tell, no one has perfectly mapped out the human brain for a single individual, let alone for our entire species. We’re not even sure where we go when we fall asleep.

Magic

It’s a funny thing to live in the modern world. There’s a promise that all of our problems will be swept aside with impending technology. They’ll be whiz-bang good. They’ll both automate our lives and make our lives more enjoyable.

This isn’t how humans work1. Humans have continually rising expectations. They get a computer to answer their questions and build software systems in real time and they reply, “so what?”

Humans, we, also find ourselves in fear of what it can do. Wonder quickly gives way to terror. Let me share a story.

I was, for a time, supposed to learn Korean. My wife is fluent and much of her family lives there. I tried, diligently, to learn the language, but it wasn’t going as well as I had hoped. So, I turned to AI. I used ElevenLabs to first clone my voice and then Gemini to translate my speech. I put my Korean words — I had absolutely no idea what they meant — into the AI and had it say, in my voice, an announcement of our baby.

When we first had it running with my wife’s aunt, she reacted first with a “whoa,” and then “this is terrible. They can steal identities.” Her husband was a bit more congratulatory, but all the while the aunt shook her head and admonished the new technology. She fretted a voice would one day trick her mom into wiring money, or mimic her children’s and convince her they were at risk.

Somewhere in all of this, we find ourselves having to do the right thing, while defining the right thing.

We don’t want to lose the magic in the technology, but ignoring concerns about it and claiming the market will sort it out is, to say the least, problematic. How? Well, let’s turn to Meta.

Think of the children

Social media — when it started, it carried this promise that anyone could connect with anyone. That the walls of fame and normalcy would dissolve, that your connections mattered. That community would win the day.

What it became, though, was obvious in retrospect and the darkest scenario looking forward. It’s now clear that Meta (and other companies) manipulated the emotions of their users, especially children, to sell more ads. They used various ways to algorithmically cause depression, then showed ads that aided in pushing people to believe that they could buy something that would cure it.

This was great for sales. And, over time, Meta achieved a kind of monopoly in social media. Modern social media has become a place for “doomscrolling” — a kind of perpetual “not-good-enough” distortion field that feeds into self-loathing and purchases2.

Meta has to pay up to $17 billion, but the damages will last far longer than that. There will be changes to their algorithm or else they'll face more severe repercussions. All of this was not done by the government taking action (Meta became a powerful force in getting people elected), nor did it have anything to do with consumers suddenly dropping out of a crappy and addictive product. No, it was a court case that forced a massive settlement. This is what finally won. States banding together. And it only took ~20 years. Two decades to sort out a negative influence in the world3.

AI

AI cannot go down the path of social media. If we let the market decide alone, it will become the worst version of itself. A kinda crappy monopoly that uses the worst tactics to make the most money.

This is the way that capitalism works. Crappy monopolies do shady stuff without proper regulation. Don’t believe me? Consider — prior to the FDA, most “drugs” could kill the patient, far more dangerous than the disease. The food would often contain dangerous ingredients in copious amounts — things like glass in sugar and sawdust in bread. First, competition should handle this; then, when monopolies exist, human nature is such that it pushes the lowest quality at the highest price.

What would this mean for a technology as powerful as AI? Well, imagine people strapped into a screen, with Soylent keeping them nourished, while they live fantastical lives in the digital world, and their actual bodies languish. Or, worse, the kinds of horrors that arrived with World War II — the eradication of whole populations deemed “useless” or “unimportant.”

Instead, competition today is doing wonders to keep AI advancing, to keep its pricing down, and to keep the companies innovating at a break-neck pace. It’s also forcing competitors to test each other’s models, and report out when something is causing major issues, such as security concerns.

This is great for the world! Open source models from China could cause this to continue unabated for a time. It is, after all, unlikely that there will be a winner-take-all, but competition consolidates over time. People choose a vendor. They “lock in” to a solution. They opt out of others and decide to go down a certain path.

With open source, it’s most likely that AI will increasingly become a kind of electricity — a utility, of sorts, that provides “intelligence” at the spigot. On demand, always ready.

But that doesn’t mean we’re “in the clear.” In fact, if anything, it would lead to several places where intelligence is monopolized. When the raw intelligence is cheap, the money moves up the stack — to whoever owns the data, the distribution or the government contract. Imagine a kind of monopoly in governmental or military intelligence. Or what about one that does taxes. Or another for law. This is the belief of many a venture capital firm. And they’re scrambling to seize as much opportunity as they can.

After all, there were several places for bread when people started to get sawdust in it.

Regulation

My favorite joke about regulation is stolen from Annie Hall. Two women are in a restaurant, dining. One says, “My, the food here is terrible.” The other replies, “I know, and in such small portions.”

Somehow regulation manages to be both awful and there’s not nearly enough of it. There was, seriously, a bid — the AI Data Center Moratorium Act — to basically stop all data center construction. The main driver of growth in the United States — the key to AI building and the United States remaining a dominant player in a new technology — the solution when confronted with potential badness was to “shut it down,” so that it could be, ostensibly, studied. Ironically, the people who study it are also likely to be using AI to do so.

But not doing anything is also bad. Technology changes the world. It changes the way we interact with one another. It changes how we view ourselves. Look to Looksmaxxing — young men optimizing their faces and jawlines, sometimes surgically — a phenomenon that I briefly touched on.

Sure, there are fads, but how we interact now with our neighbors is far different from how we did even 20 years ago — and social media has much to do with that.

AI can be the ultimate force for good. It can wipe away a lot of the ill that was caused. If it’s properly regulated, distributed broadly, and given widespread access, it can do more for the benefit of humanity and lifting people up than any technology before it.

Or it can be controlled by a few wealthy corporations that dominate over time. This is what we have to pick, right now. Picking correctly looks like the FDA, not the Meta settlement — rules written before the damage is done, not a payout two decades after.

Only later did I learn about looksmaxxing — a kind of attempt to make one’s self as attractive as possible. This includes severe plastic surgery and steroids for 18 year olds. It includes, too, severe dieting and attempts to achieve the bodybuilder aesthetic.

Plastic surgery is expensive. It’s not everyone who can afford to be nipped and tucked. It’s not everyone who wants to be. But an ever larger portion of the population is turning to the idea of something they can control — such as the way they look — instead of something they can’t, like where they were born, or to whom, or what the economy is going to ask of them next.

I sat in my nine-year-old son’s play, an abbreviated version of Descendants, the Disney movie. In it, the children of the historical bad guys (Cruella de Vil, Maleficent, etc.) are banished, until a new king of the good guys invites them to come to the same school.

It is, in a sentence that seems far too grandiose, about familial trauma — about the trauma that we hand down to our children. It’s strange to watch your son play in a production about this, let alone to watch it while you find yourself obsessed with a book about how we as a culture must change to one of us.

Individualism

The idea that a person pulls themselves up by their bootstraps — that we are, all of us, individuals and responsible for ourselves. This changes slightly as we age. And if we’re fortunate enough to have families, then some of us are responsible for those families too. Others never get out of their own sense of individualism.

The book is from Terrence Real, it’s called Us, and he encourages everyone to start thinking, as he describes it, relationally. It’s kind of radical. One of the first things he says is to stop looking for an objective reality. Instead, accept that it doesn’t exist — that the reality we live in is painted both by our histories, our experiences and each other.

Humans co-regulate. We work together to manage each other and our own emotions.

In fact, being social and cooperative is kind of our super power. Take it from Yuval Noah Harari, the historian, who points out that humans took down woolly mammoths or herds of buffalo by cooperating, by working together, rather than competitively. In fact, ancient civilizations doubled as forms of cults, basically. They were large shared beliefs — where the people thought their kings Gods, literally divine. Questioning this could lead to death.

Life is, in a sense, a shared hallucination. It’s a kind of agreement about the stories that we have — that the score of a football match matters, or that money will give you a bottle of water, that a leather purse is especially stylish when made by Hermès.

These stories are our way of cooperating. They extend themselves, even, to capitalism. At its core, capitalism, while considered a system of competition, is a system of cooperation. Before commerce provided a method for people to get what they need from each other, many would simply take it. Through violence or theft, they would take and harm others.

It’s worth saying what that system organized us around, which was work. Your job is how you’re useful to strangers, and so it’s how you eat, and so, eventually, it’s who you are. But it’s still an agreement. And agreements are the kind of thing that can be renegotiated.

The end times

Packed gyms reflect a perverse truth about America — that we’re spending more time alone together. We’re all filling up the gyms to looksmax, but we’re not stopping to improve our relationships with the people around us.

Social media has captured our community and online life has tried to claim our souls. Mark Zuckerberg is the unfeeling model of a billionaire cool guy surfing in a suit. And his platforms laud the superficial — attractiveness, wealth and irony. Nothing has meaning, and it’s devoid of the hard fought lessons of humility and growth. The cycle of harmony, disharmony and repair.

Instead, we’re taught that if we make ourselves the best version we can be, that the world will be better for it. The truth is, that we really are responsible for those around us. We co-regulate them. And we’re responsible for telling them what we think (kindly), being brave and sharing with them the truth that we live with, while honoring their truth.

This sounds fluffy and nonsensical, and impossible. But is it? Cultural tropes change all the time. In fact, for a long time, the idea of community was commonplace. People lived in a shared experience. They went to war together, they gathered at block parties, they celebrated each others’ birthdays. Neighbors would pop by. Home was tended to by strangers, who then became friends.

There was a guy who I had the fortune to know. I met him through a friend who knew him for 55 years. They lived in the same apartment building, but one floor separated. This led to a comedy of errors — they got each others mail and occasionally their packages. The result was that they became lifelong friends. They were the godparents to each others’ kids. This is community.

It’s a unique situation, but one wonders if this happened today what would occur. Maybe the same thing. Maybe they’d find a reason or a way not to interact.

Robert Putnam gave us the phrase, and the data — we’re now bowling alone a lot more than we used to. But we should begin to bowl more on a team.

And this has to do with AI…. How?

The core thesis of this blog is that we have a chance to now change the way we live, work and interact with each other. In short, the change that our technology will force upon us will introduce the ability to rethink our systems. Some things shouldn’t change. Some things should.

The bootstrap story was never only a philosophy. It came with a job attached. It worked because the work was there. It worked because the effort you put in showed up, more or less, in what you got back. And it worked because your job answered the question of what you were for. What do you do is the first thing we ask a stranger in this country. We think it’s small talk. It isn’t.

AI is going to loosen all of that. Not everywhere, and not at the same time. But the line between what you put in and what you get back is going to get harder to see, and it’s going to happen to people inside a single career, which is new. When the answer to what do you do gets shaky for enough of us at once, the story we’ve built on top of it gets shaky too.

That’s frightening. It’s also the opening.

We don’t rewrite our agreements while they’re working. We rewrite them when they stop. The last time work changed this fast we came out of it with the weekend, the pension, the public school. None of that was natural or inevitable. People argued for it during a period when the old story had come loose.

We should admit to ourselves that we are a species that needs to cooperate. This comes with tribalism. It comes with nationalism. It comes with groups who move, herd-like, against things they view as threats.

But it also creates community, it builds friendship, it reduces loneliness and despair. How we treat each other matters.

Now that we’re looking at a future rewritten by our technology, it’s time to consider how we can rewrite how we treat each other too.

To which he replied with a very apt question, “Well, what would you do differently?”

The question demands empathy, and when framed that way, it’s kind of captivating. The editor-in-chief stammered out her reply, “Well, I think you supporting Ukraine was the right thing to do.”

It’s hard not to believe that Musk is in part or very self serving. And so, when he told the Economist that money would basically be unimportant with AI it was impossible not to think, “easy to say for a trillionaire.”

And he has said something similar before in the past, offering it up as a kind of panacea for the potential ills that AI will bring. This time, though, he brought a bit of economic theory to the party.

He broke it down like this — inflation is, essentially, the amount of money printed divided by the amount of goods and services produced. AI and robotics in his estimation will greatly increase the amount of goods and services produced. This means that these technologies are deflationary, and, “the government can simply add some zeros to their money database.”

This is reminiscent of another argument he made before, which is that the technology itself is basically mandatory to save the government and its benefits1.

And, it’s easy to start to take him seriously2. But let’s look at what inflation really is and what this all could really mean.

What exactly is inflation

The core formula for inflation is pretty simple to calculate — in a sense it’s the consumer price index — this basket of goods that the government tracks — did the price of those go up or down over time.

This ignores the underlying mechanics of what makes the price go up or down, and Musk’s main point about technology having a deflationary impact. This is because if it becomes far cheaper to produce those goods (or in the case of AI, you remove the main cost driver of the price of those goods — labor), then you’ll have far cheaper goods. This, as the technology progresses and the initial costs of labor indicates that Musk’s main point about “an infinite economy,” i.e. all of anything can be made at any time for next to no expense, interesting, but hand-wavy, at best.

Musk puts in the numerator the amount of money that the government produces, but as well see later, that’s also a specious claim. This rudimentary short-hand for inflation — amount of money the government prints has become a kind of standard, but ignores a lot of the complexity of what the government puts its money to and how, and subsequently where that money ends up.

So let’s unpack first the price of goods and services, then the value of where and how the government spends its dollars.

Would a rose by any other name smell as sweet?

There are highly, highly automated areas in life where the cost of goods or services continues to go up. And the value of many goods that could be considered automated or free in fact now cost far more than they used to.

Take education3. The price of education has skyrocketed, far beyond inflation, for the last several decades.

This is even during a period when becoming educated on a topic has become far cheaper and easier than ever before. The Interest was invented in the early 1990s, and access to it spread in 1995 and shortly thereafter.

The ability to access the world’s knowledge should have made people brighter and less needy to go to academic institutions to learn. Yet, the opposite happened, the world learned about a few elite universities and how to get into them. They learned more about how to game the system. The world created entire feeder industries to getting into a top college, rather than colleges becoming less attractive.

Because the value of education isn’t in the education, it’s in all of the ancillary aspects of it. If one goes to Harvard, as a friend did, they get access to an alumni network and a brand that is forever with them. Harvard opens doors in a way that no previous pedigree would.

Harvard changes fortunes. And that is because it also acts as a kind of filter for elite industries seeking out top candidates. Immediately upon reading “Harvard” or otherwise, a recruiter and hiring manager trusts that Harvard did the work of vetting the character, work ethic and fortitude of a candidate.

This is broadly true. The value of a thing is more than one output. Or rather, the utility of that output has intrinsic and intangible components to it.

It’s unlikely, therefore that if AI or robots suddenly infilitrated education, the price would drop dramatically. The value of the education is not in any one component, but in a brand, legacy and history — it’s in things that can’t be automated.

This is true of so many aspects of life. Take health care where the value is not in the life-saving surgery, but in getting the best doctor or the closest doctor. The value could be dropped there, especially for routine surgeries or for robotic surgeons that are produced en masse and spread world wide, but if there’s a dire emergency, people will pay. And in the case of something like saving a life, they’re go into debt or more to do it.

Or take the price of land. While the various aspects of what can be built on the land could become less expensive with more fixed cost, the land itself is unchanged. Even the nature of what’s considered valuable4 on that land is unconcerned with the price of land itself.

Monopolies

A thing that holds sway in a market is monopolistic power. Highly automated fields today that mass produce goods still are not free.

In the case of things like bandages and gauzes there is a price that is both above zero, i.e. profitable, and less than any new entrant could compete with. In a fully automated world, this can be pennies, or even fractions of a penny, but it’s unlikely that you’d find a person who would choose to not make money on something they build.

This, generally, is considered “pricing power” in business. And, as part of that are core intangibles — things like brand power and monopolistic control over the market.

In the past, markets have been busted up because they represent an unfair scheme for a customer, i.e. a company with monopolistic power can push prices up. This was the case of the train barons — they controlled the railroads, and therefore could charge whatever they wanted on the lines. But today’s monopolies often flip it — they hold prices at a level that prevents competition.

And unless the government takes over all production — which is essentially the definition of communism5 — the owners of the businesses will continue to both want to make a profit and be intelligent enough to find profit.

Inside of this pricing power, in many places, owners actually price higher than would seem influenced by the market to price out the lower tier of consumers. For example, an efficient market would assume that every building owner would want to maximize the rent they receive from their tenants, in order to both recoup their original investment and keep occupancy rates high.

This, though, is not the case. Often, there are owners who don’t have to rent out their properties, who can sit with them vacant, or a high percentage vacant. Often there are owners who would rather have properties sit vacant than lower the rents, due to the various externalities of lowering the rent.

And these owners don’t feel pressure outside of making a return on the initial capital. If they can achieve profit off of higher rents and fewer tenants, then that’s perfect for them.

Monopolies, and there are many of them6, distort markets and have no reason with automation to suddenly stop charging customers.

Automation ≠ value

Aside from the various parts mentioned for monopolies and the fact that many costly parts of the economy would be untouched by automation, it’s also true that as the price of goods drops and has dropped, humans have a tendency to devalue them.

Take a aluminum can. Once viewed as a marvel in creation — the refinement of aluminum was so challenging that it was placed atop the Washington Monument — a symbol of the great affection that the United States had for its first president.

Now, aluminum cans are so plentiful that they’re often thrown away or discarded on the side of the road. It’s unsurprising to see aluminum as, basically, trash.

If you offered someone a 9-inch, 100 ounce pyramid of aluminum, a treasure in the late 19th century when the monument was built, today, they would be unlikely to pay for it7 or they would only insofar as a paper weight.

There are many such goods in the world like that. And it’s not that the cumulative value goes to zero when the price of one thing is devalued, rather it’s that the value shifts to something else.

Paper at one point was so expensive that students would buy a few precious sheets, which they’d reuse and turn over and write in tiny script to cover with their notes. Now it’s so plentiful, we throw it away when it comes in the mail without even looking at it.

This has been true for as long as humans have existed. We value the rare and devalue the abundant.

And it’s true of even experiences. Take David Foster Wallace’s essay on a cruise ship experience, where he marvels that he can watch Jurassic Park whenever he wants, and order a sandwich at any hour of the day and receive it within 15 minutes.

After a few days of this, he found himself annoyed that he couldn’t receive the sandwich in 5 minutes instead of 15. He found himself frustrated that there wasn’t more selection8 , and that he would watch Jurassic Park.

He began to devalue all of the things that were so abundant on the cruise ship and search for the things that were rare — the ability to walk outside of the confines of the ship, the ability to have some time outside without being in a public space, etc.

And then he became miserable because he couldn’t have those things — those things are antithetical to the cruise ship experience.

The mass production of goods and services will actually make a few special experiences far more valuable. Let’s say that there are robot barbers everywhere, then a human barber, with great skill actually becomes far more valuable. But the economics change — while the human barber with great skill could command huge prices for haircuts, and fly around the world to cater to billionaires that would value their services, the vast majority of barbers could not make a living. They could not compete with the robotic barbers.

Government spending

While the majority of this essay has been on the production side of the equation, it’s important to also talk about the supply side, or in this case, the government printing money.

The machinations of government funding are, to vastly oversimplify, complicated. It involves a kind of constant pork-filled process where money goes out to any number of social or political causes that are either in vogue at the time or meant to aid in re-election bids. Or the money gets cut, programs defunded, and what was once considered important becomes passé.

This is how we have bridges to nowhere, and money to fund any number of absurdist programs. This is how we can have people in the Department of Defense sitting on three-legged stools to save money, while the defense contractors make incredible profits.

So, the money that would go to people or projects or things, would almost certainly be tied to some kind of outcome. And if it wasn’t, if, as in the case of the checks mailed to everyone during the pandemic, or the small-scale tests of universal basic income, the money, ultimately ends up in the hands of the people who own the credit cards, who own the properties, who can charge rents.

A natural experiment

We went through the process of printing excessive amounts of money for several years after the Great Recession through “quantitative easing.” This process ended up adding massive amounts to reserve currency and balance sheets, without creating a net positive economic impact. This was because of where the money went. It went to the hands of the wealthy, to help them navigate a massive crisis.

The gains of that for ordinary people were felt in the stock market, the bond market and in asset prices, like homes. So asset prices went up, while the general economy stayed basically the same. This, at the time, was heralded as a triumph.

A fair shake

Let’s take a moment to consider seriously the idea that anyone can have anything at any time made on demand, for free. This is the core idea of Musk’s “infinite production of goods and services.” This would, basically, break economics. The Consumer Price Index as all zeros, or, for a time, severely negative, would make it irrelevant and not able to measure anything. Inflation as a concept would, with it, become irrelevant. An infinite expansion of goods and services would be able to survive an infinite number of zeros on a government’s balance sheet, which could fund any and all pork projects, all bad ideas, stimulus checks and more.

But too it’s silly to think that rents wouldn’t go up with this infinite money — that landlords couldn’t charge trillions of dollars a month to live in a place. That banks that held mortgages and fixed-rate loans wouldn’t go bankrupt in the hyperinflation that would follow.

Indeed, Musk’s idea is that that the government would basically have to add an infinite number of zeros onto a system that reaches zero. This would be the only way that the economy could stand still. And at this point breaks core economics. This is Musk’s double point — that both traditional economics will break and money itself won’t matter.

Money as a story

The problem is that money will matter for the same reasons that it matters now — because it’s a story of value. Money stopped being a thing a long time ago. It’s a story now. We assign value to it, and the more convincing the story, and the more people who are convinced by it, the more valuable it is.

This might seem like lunacy, but consider the value of a dollar. It’s decided by a number of economic indicators — the reserve currencies of a variety of countries, for example, or the exchange rate, or the level of inflation.

But fundamentally what these measurements tell us is how seriously people are taking the story of money. At its core, it’s just a number in a digital space that goes up or down, and that becomes really important when you need to get some groceries.

The problem is that we don’t have a particularly good alternative for these things, and it’s most likely that Musk is, well, wrong. That the shift in value will go to more bespoke, more high-end, more custom and human and tailored items. That the value of things will not be in having anything at any time that is useful, but having the right thing, the perfect thing that represents you or who you want to be.

The second that you can’t get anything at any time, then Musk’s argument breaks. Then the economy bifurcates between the limitless, abundant and inexpensive (but likely not free) goods, and the extremely expensive, high-end luxury goods and services.

And, while so much of the world will become automated, art will still exist. It’s just that we all will become artists. The most beautiful outcome is that we all become craftspeople, artisans, making things that we deeply value.

One caveat

There is a place where Musk’s argument does hold up — and that is in the limitlessness of the digital world. AI already can create digital experiences and spaces that far exceed what humans can, at a breath taking speed.

And humans have never exactly existed in the digital world. Rather, this infinite expanse of software can be truly realized. There too the constraints are physical — they’re in compute and in energy, both of which are being solved by the significant spending of the major tech giants.

But in the next few years, the price of software will be basically zero. The price to build software will be basically zero. Any any every service imaginable digitally will be zero. The economics will break.

Money won’t matter for software. What will matter is how it interacts with humans and enriches our lives.

What will the world look like? I recently flew on a United plane and the Starlink advertisement with Arthur C. Clarke saying, “The only thing we can be sure of about the future is that it will be absolutely fantastic,” must have played half a dozen times1.

Clarke was more confident than I am. The future could be more abundant — in opportunity, in choice, in intellect — than we can imagine. Or it could be a hellscape of inequality and neo-feudalism. Or it could just be kind of crappy: a weird, half-hearted attempt at social welfare, constantly undone by a dying empire2.

Here are all three.

Abundance

The technology came on quickly. AI automated whole swaths of work. And yet, as intelligence became cheap, the world began to value things it hadn’t before. Human skills — creativity, empathy, sociability — commanded huge new prices. The biggest rewards went to people who could find new problems worth solving, the right problems, and imagine simple ways to solve them. Or who could sit with other people and help them understand what it means to be human. Or who could navigate the human dynamics that most find impossible.

So much had to change, and it did. The education system got a complete overhaul. The wealth generated by the new tools was heavily taxed. This wasn’t unprecedented — we’d done a version of it before. For most of history, brute force dominated work; the mind was secondary. Then the steam engine took over the physicality, and the ability to design systems came to mean far more than the ability to use a shovel. The economy became knowledge-based. Then the knowledge stopped mattering, and we made the leap again.

And such abundance. More food than makes sense, at prices near zero. More software than anyone can use. Robots built the infrastructure, and the simultaneous discovery of new architectures and new materials revolutionized what people could build. Entire worlds towered up over the skyline. The space elevator carried masses to the stars.

The social welfare system was tightened and improved. It caught those lost in the acceleration while the benefits of the technology spread broadly. The government embraced the tools itself, cut its own costs substantially, and freed up capital for investment and a greater role in citizens’ lives.

In short: the government got a spine again. It taxed the wealthy. It committed to a more equal and inclusive society. And somehow, it didn’t manage to mess everything up in the process. The lessons of the last 200-year transition — into the knowledge economy — served us well in the transition out of it.

Dystopia

The headsets were the first part of the “uniform” to become free. Then the haptic gloves, then the bodysuits, then the link directly to the spinal cord. Then came the nutritional concoction that provided just enough to stay plugged in, and the caretaking robots that washed people and cleaned them up.

This too became free. The bodies, eventually, found their way into containers — people hovering in perfectly tempered water, stacked like drawers in a dresser, pulled out when deceased and otherwise left alone3.

Not everyone is in the pods, of course. There are the wealthy owners of the businesses giving all of this away. First they built the data centers, then the headsets, then the gloves. Later they built the pods, and collected the UBI checks sent to everyone occupying them.

Everything was free, except nobody was.

Their workers are robots and recursive AI, and they spend the world playing emperor — yachts and helicopters, big fish and huge waves. They are the inheritors of the earth.

Here, the government does largely nothing. It allows capital to swallow the world and intervenes only when demand destruction threatens to outweigh the cost of a tax. The world has become a mass company town, with a few firms taking all the benefits.

But what of the people in the drawers? They’re perpetually online, living hollow shells of their dreams in a place that is undoubtedly real, full of fidelity, with ample opportunities to spend their dollars. The economy inside is infinite. They can live with whatever struggles they wish, and the game masters craft worlds that perpetually challenge them — the AI forever tuning the difficulty.

And the people who opt out of these “free” lives? They live like the homeless of today, wandering and subsisting on what they can find. The funny thing is that the world outside would be absent much of today’s chaos. It would mostly return to what it was before humans came. Robotic caretakers keep the parks and forests in check. The suburban sprawl was simply unnecessary.

Half-hearted

The strangest thing about this world is how familiar it looks. Nobody built a space elevator. Nobody built the pods. The suburbs are still there, a little shabbier.

The technology arrived on schedule. The response didn’t. Congress held hearings. A commission was formed. A UBI pilot launched in three cities and died in an appropriations fight. The benefit that finally passed was means-tested into uselessness — enough to keep people off the street in the cheap states, not enough anywhere people wanted to live. It phased out at exactly the income level where someone might have started to climb.

Unemployment settled somewhere in the twenties and everyone stopped calling it an emergency. The official numbers got creative. If you’d given up looking, you weren’t unemployed anymore; you were “detached.” A lot of people were detached.

The companies kept growing for a while, selling to each other and to the shrinking group with money, until they couldn’t. Some regions got the abundance — the compute corridors, the coastal metros where the owners of the machines still needed lawyers and restaurants and each other. Everyone else got what Flint got, in slow motion, over forty years: the plants close, the tax base collapses, the services degrade, and the country calls it a local problem. Except now there’s no Sun Belt to move to. The escalator goes everywhere.

The government didn’t do nothing, which is the tragedy of it. It did just enough to prevent the crisis from ever getting bad enough, all at once, to force the real answer. Every collapse was partial. Every bailout was temporary. Every reform was pilot-sized. A dying empire doesn’t fall; it defers.

People adapted the way people do. They doubled up in houses. They traded work off the books. They built things in garages and sold them to their neighbors — a small-craft economy nobody planned and nobody taxed, growing in the ruins of the official one. Which is either the seed of what comes next or a coping mechanism. Nobody in this world can tell, because nobody in this world is measuring.

The future wasn’t fantastic. It wasn’t a hellscape. It was Tuesday, indefinitely.

Clarke was right about one thing: we can’t be sure. But these aren’t three doors we’ll open blind. They’re three directions from where we’re already standing, and the difference between them isn’t the technology — it’s everything we decide to do about it. We don’t get to pick a future. We only get to steer.

I’ve thought of myself, at times, as a modern day IBM consultant or a kind of pseudo inventor — a fanciful designation, sure, but one that I wanted to be when I was a kid, so especially nostalgic for me.

But consistently people are not warm to the idea of AI.

The reasons are many, but the vague attack I heard over the weekend was that it was “destroying the planet.”

This could mean that it’s destroying the world environmentally. Or that it’s deepening inequality or that it’s going to cause some kind of irreparable harm.

This particular criticism was from a liberal friend who spent years as a teacher and as a writer. And when she mentioned freelance writing I had to pause and think, “Will this still exist in a few years?” Well, my response is — sort of, but it’ll likely be different. This, to me, felt salient and important. And it’s worth examining closely.

The other many criticisms of AI don’t amount to a lot, frankly. Water consumption has been way, way over hyped.

And we live on a planet that is more than 70 percent water. Desalination plants in places like Aruba and Israel have given those previously desert climates all the fresh water they need and more.

Electricity, often too, is cited as a reason to stop adding data centers, but we have the technology and know how and capability to add abundant, clean energy to the grid. It’s called nuclear fission — it’s been around for ~70 years, and it’s the dominant energy form in many places, such as France.

Don’t like nuclear (too many concerns about things blowing up?) Then I don’t blame you, still consider the below:

Solar power is now the cheapest in the world. It's also by far the most installed form of new energy. It's growing everywhere. The sun provides way more energy than we need, and increasingly humans as a species are getting good at harnessing that energy.

Still not convinced, then consider fusion energy — literally making mini suns (or at least the same way that the sun creates energy) on earth. Most recently, a startup out of Massachusetts applied for their fusion reactor to attach to the grid. It’s a lengthy process, so it will still take a few years to go through all the appropriate regulatory hurdles. Nonetheless, the progress is real.

The energy from this fusion reactor will likely, you guessed it, supply power to data centers.

I don’t mean to be dismissive, just point out that while these are problems (desalination can be an environmental non event or environmentally destructive, depending on how it’s implemented) they can be solved. And more importantly, with the right motivation, capital will and is solving them.

Rather, the pessimism about AI seems centered on one significant issue — unemployment, and with unemployment, inequality. This issue is something that capital is also pointed on “solving.” I’m part of that “solution.”

But the pessimism here isn’t necessarily that AI can take over vast swaths of work. It’s that people do not have a recourse against the impending automation. Government doesn’t appear to intercede on behalf of people, at least not in a way that resolves the problems and provides better outcomes.

Take a look at various inner-city communities that have had decades of “investment” by way of government policy. The change has been slow and painful. Black unemployment has run roughly double the white rate for as long as we’ve measured it separately — today it sits around 6.6 percent, against 3.6 percent for white workers. Better than the past, by far. Still stubborn.

This is after WIC, food stamps, low-income housing, school lunches, affirmative action, forced school integration, our first African American president and so much more. And here’s the uncomfortable part: for all of that spending, we don’t really know how well any of it worked. Even Barack Obama — who came up through the exact institutions these programs were built to open, Columbia and Harvard — has said he has “no way of knowing” whether affirmative action helped him get there. If the most celebrated beneficiary can’t draw a clean line from the policy to the result, what does that say about our ability to measure the rest of it?

Which is the harder question underneath all of this: can government even be a good social custodian? Can it reliably step in for people and produce a better outcome — or does it mostly spend, declare progress, and leave the results in the air? The honest answer is that the record is mixed at best. And that is the real engine of the AI dread. Not that the machine is coming for the work. That the institution we’d count on to soften the landing has a shaky history of softening anything.

Now those programs are being slowly dismantled. The ideal scenario is a world that doesn’t need those programs, where the schools on the South Side of Chicago are as good as the ones on the North Shore.

But those programs are being dismantled and there’s nothing here to take its place. Instead, we have a lot of reason to be pessimistic. Social media came on like a drug that never got regulated. Internal Meta documents, unsealed in litigation, indicate the company found its own products worsened users’ depression, anxiety, and loneliness — and shut the research down rather than publish it. Children became addicted to screens with almost no enforcement in the United States to limit or discourage it, even as some studies suggest test scores in certain cognitive domains have started slipping for the first time in a century.

What are people to think? The world just minted its first trillionaire and there’s AI slop in their feeds.

The promise of AI is that it’ll take over work, but it’s more than that — it’s taking over things they love to do. It’s generating videos that they like to make, it’s taking over their art, it’s making men look like women and providing therapy to the lonely. It’s taking over huge swaths of writing. The world, goofy as it was, is becoming increasingly AI. We stuck a whole generation into a digital world then we said, “We don’t really need you here anymore.”

The technology is representative. It’s a kind of idea. It’s the idea that people will no longer be needed. It’s that a tool will suddenly render the contributions of many, the gains that so many have made, irrelevant or useless. That humans will become useless in this world.

And, yet, to quote William Faulkner, “I decline to accept the end of man.”

We have always been in a race between our humanity and our technology. We have always been on the cusp of doom. The future belongs to the people who seize it and use it as a tool.

I watch my nine-year-old make Minecraft armor with Claude, and he does it because he can. To him, Claude is just skipping a four-year degree. He’s still going to have a week-long summer camp building with the tools to make Minecraft even more customizable.

He does these things because they’re fun. Because he can brag about it to his friends. Because he can.

AI expands the scope of what humanity’s capable of. It, much like the industrial revolution before it, will expand our capacity. And now, we’re invited to contemplate what it means to be human.

And this is a question that only we, no machine, can answer.

This did not prevent his studio from unionizing against him, which as Acquired just laid out in a fantastic episode, led to Disney getting excited about trains, which led to, yes this is true, Disneyland. So, in a way, anger at communist sympathizers led to Disneyland. Consider that the next time you tuck into a turkey leg with a Dopey hat on.

There was a time when many in the United States feared that it would go communist. This was, ironically, not during the Great Depression, when the New Deal was the last resort for so many. Rather, it came after the end of World War II, when the concern was that America would lose the Cold War, when it became a conversation about too much government overreach.

So where are we today? Well, we’re not yet at a place where we’re targeting our concerns for communist sympathizers. And increasingly, it looks like socialism is back — there’s a pent up demand for something new, something different, an opportunity that we haven’t had in a long time.

I’ve often wondered exactly what caused this extreme dislike and distaste of communism. I say this almost in the abstract. It’s a thing foreign to me. It’s like disliking carrots. Like there are probably many reasons why people don’t like carrots, but to be really upset at carrots feels peculiar.

This was before I met a refugee from Vietnam. He was furious at the notion that communism could be considered “OK,” by anyone. He would likely throw the computer across the room and stomp on it and maybe even give me a call to tell me how much of a fool I was.

Communism, for him, destroyed his home, forced his family to leave their native country and brought about decades of suffering in a foreign land. And truly, if you look at the countries that have gotten some form of communism, they do terribly compared to their direct counterparts. The two obvious examples of natural experiments in world history here — North Korea and East Germany — which fared (or are faring) much worse than their capitalist counterparts.

Put this way, few things in history have caused quite as much human suffering as communism. Pair this with the decades-long crusade against communism, the active campaigns to root out communism, and the triumphant sense that people had when the Cold War was won, and you get a unique cocktail for a hatred of, a loathing of, anything that smells like communism.

So, we went from implementing socialist policies that still form the backbone of our social welfare state in the 30s and 40s to hunting down communists in the 50s to a Hippie, free-love communist ideology in the 60s, to a decade lost to the cocaine in the 70s to winning the war against the communists in the 80s and 90s, to every country entering free markets in the ensuing decades — the world carried the anger at an ism without it staying the same. Communism, for what it’s worth, has changed.

Now when young people talk about socialism (which is, importantly, distinct from communism — a nuance that’s often ignored in modern debates), they’re not talking about the kind that was dominated by Pol Pot and Stalin. They’re not talking about the kind that hurt or killed millions of people. They’re talking about the kind that helps people live knowing that they can feed their children.

And listen, if you’re skeptical like me — consider a real number. In America, roughly 48 million people — yes, about one in seven of us — live in food-insecure households. More than a third of single-mother households (about 37 percent) experience food insecurity. The price of groceries went up. It has a real impact on people.

If that’s not enough, then consider that the largest city in America just handed its keys to a democratic socialist who ran, more than anything, on the price of food. Zohran Mamdani won New York promising a network of city-owned grocery stores — a “public option for produce” — to make eggs and bread cheaper in the neighborhoods where, as he puts it, putting food on the table “feels out of reach.” When he announced the first one, he sited it at La Marqueta, the East Harlem market Fiorello LaGuardia built in 1936 to do exactly the same thing. A New Deal mayor’s answer, dusted off for 2026 — the same year we minted our first trillionaire.

In this scenario, people are turning to old promises for solutions to new problems. They want a new new deal.

It’s a shame that the solutions are also, seemingly, old. They are pushing for more unions — the very thing that Disney chafed at. They’re pushing for profit sharing, corporate ownership, robot taxes.

I don’t want to say that they’re all equally bad. Some are worse than others. (The worst idea I’ve perhaps ever heard is a moratorium on data center construction.) But they don’t exactly solve the problem.

The country’s problems are many: we have an explosion in inequality. We have a major divergence in living standards. We have a trillionaire and deep poverty. We have communities where things have not gotten better. We have technology that will accelerate the displacement of work. We have a clear and material impact on the living standards of millions.

This isn’t necessarily a long-term issue. Long term, humans can thrive. If you don’t believe me, watch a 9 year old use an LLM. Mine does. He makes games with it, he builds “dragon armor.” To him, an LLM is a sometimes annoying assistant; it’s a tool. He never once worries about it taking over his “job.” Because he never once thinks he should be doing this work. He’s doing whatever he needs to do to make something that’s fun to play, that he will be proud of.

This is the crux of a future enabled with LLMs. People will still create. They won’t overtake us. They won’t dominate our brains. They won’t make us lazy and undeserving, but they will make it really, really hard for many folks who have learned a certain way of working to change.

And the change is coming quickly. So quickly that there are tens of millions, potentially hundreds of millions, of people who are going to be in really challenging circumstances with work.

They see it coming. They want to have a government that works. They want to have the pressure taken off of them and be part of the great abundance our technology can provide.

And, they’re going to vote that way.

Welcome back, socialism. You’ve been gone for quite a while. I wonder if this time we’ll get a Disneyland (or a much more unfortunately named Muskland) on Mars.

If you assume that the whole market is possible, i.e. that everyone will buy a smartphone eventually, you can understand how to market to them by finding where they are in the adoption curve.

Typically, it takes a while for a market to develop. Smartphones, now almost impossible to escape, were first introduced in their current instantiation in 2007. It took until 2013 for them to become the majority in the US. And even now, nearly 20 years after the iPhone, there are people who fight the adoption of smartphones. Only when there’s no other choice, when the default is a smartphone, when land lines literally will not be installed in a house, do they accept the device.

And the iPhone really unstuck the market. The BlackBerry was beloved by its users, but couldn’t find broad market adoption because of its clunkiness. The first device to carry the name launched in 1999 — though it was a two-way pager, not a phone; the first BlackBerry phone didn’t arrive until 2002. So the timeline there was roughly eight years in the “innovators” section before the form factor changed in 2007, which then brought it six years to the early majority, and in totality nearly 27 years to get nearly everyone in the market with some kind of smartphone.

AI as it exists now (essentially a chatbot) flew past the innovators and early adopters stages. Roughly 1.5 to 2 billion people worldwide now use a chatbot — ChatGPT alone crossed a billion monthly users by mid-2026, with Gemini, Copilot, Claude and others adding hundreds of millions more. But if we accept that the whole world could use one, and roughly half of those are currently too young to (we’re giving some really rough figures here), we’re still in the early majority for chatting.

The other use cases of LLMs, and there are many, are nowhere near as mature. And the companies and valuable use cases of today are unlikely to remain the most valuable uses tomorrow.

And the same is true for companies.

AOL is the poster child for early internet; Yahoo for early search; MySpace for early social media; Napster for early music sharing; none of these exist in their first forms, even though for a time they seemed unassailable.

I would hesitate to say with certainty that a company is going to die. And I’d go further: I think a lot of the pieces that claim to know what’s going to survive or thrive are, basically, snake oil. But there are a few trends that I think we can point to with some certainty.

Open source large language models are now approaching parity with all but the frontier proprietary models, and are therefore essentially “free,” meaning that for a great many use cases you no longer need to pay Anthropic, OpenAI or otherwise for access to the technology.

Local hardware is getting increasingly optimized to run AI — this is clear in the product announcements from Nvidia, and the various hardware that’s coming out. There’s a continued push to put very capable large language models on device.

Companies are seeing a lot of traction at the application layer — Cursor, Perplexity and others booking billions in revenue extremely quickly. And even the model labs are migrating value upward: Anthropic’s explosive growth (a $47 billion revenue run rate by mid-2026, up from $10 billion the year before) is driven as much by Claude Code, an application, as by raw API access.

These things have significant repercussions for the next phase of LLMs. First, it implies that the really valuable use cases have yet to be discovered. Next, it implies that any company whose primary value is reselling raw API access as a revenue generator likely won’t be around in a couple of years. Finally, it implies that the simple things we can build with LLMs are, genuinely, valuable.

Look at Cursor, Lovable, Claude Code, Granola or Wspr. All of these are solving simple problems — making coding far easier and accessible, summarizing meeting notes, doing really good transcription. This is not the kind of pie-in-the-sky use case people dreamt up when they’d talk about automating whole flows of work. Rather, they’re just small and simple technologies that help people work better.

The angst about the future of work is two-pronged:

People often work kind of bullshit jobs that they know can be automated. Their job, like their real job, isn’t “making corporate strategy and aligning stakeholders,” it’s making slides for their boss after he gets out of a meeting with their SVP. And this task, once involving hours-long tedium, can now be done by Claude Cowork faster and more accurately, based on notes taken by Granola.

The government appears incapable or unwilling to help. It does not reflect nor represent the will of the people, and no matter who is elected, they all seem basically like crooks. There’s a tremendous, deep pessimism against American leadership.

Bullshit jobs may be bullshit, but people count on them. And while the right thing for the economy and for a company is to automate this bullshit work, freeing up human potential for so much more, the world is also not set up for it.

This is the problem we’re confronting. AI will not take over the world and kill everyone. It’ll make powerpoint creation trivial. And that will automate a lot of jobs.

And where will those people go?

Let me put that a different way: one man has almost 4 percent of the United States’ GDP.

My mind couldn’t help but wander to king of the hill — the kids’ game, where everyone tries to knock the kid at the top off the mound. I also couldn’t help but feel a tinge of jealousy.

Then I thought about the relative distance of various increments — tens, hundreds, thousands, millions, and on.

Annoyingly, nano banana wouldn’t even put billion on the same scale, let alone trillion. Apparently a thousand billions on a line would just run past the screen and keep running.

To put it in more concrete terms. A pixel is about 1/96 of an inch — roughly a quarter of a millimeter, the size of a grain of table salt. So: 1 pixel is a grain of salt; 10 is the width of a matchstick; 100 is a quarter; a thousand is the long edge of a sheet of paper; a million is a walk down a long city block; a billion is the drive from Chicago to Indianapolis; and a trillion pixels makes a line long enough to wrap around the Earth’s equator about six and a half times — roughly two-thirds of the way to the Moon.

No wonder nano banana refused to make it.

It’s hard to fathom this level of wealth. It’s not really opulence. And, truly, there’s a part of it that is wildly impressive.

Yet, weirdly, we’ve been close to here before. At his peak around 1913, John D. Rockefeller controlled roughly 2.3 percent of U.S. GDP — extraordinary, though still short of where Musk sits now. Rockefeller, the man who modernized business, who created the standard for oil, the one who destroyed competition and integrated horizontally.

Rockefeller was a fascinating guy. Also a deeply religious man who gave away his fortune to charity. He created the standard not only for oil and for business, but for the charitable act too. He founded several of the most elite universities in the country, including institutions for medical research — places that have prolonged the lives of many millions of people.

What a weird paradox the great-man version of history is. Without Rockefeller, what would we have? Would someone else have risen to fill his place? Is it just cognitively easier to think of Rockefeller as the great man, or was it inevitable that we’d produce someone like him?

Or Genghis Khan or Winston Churchill or whomever. This version of history insists that singular, focused individuals (I realize “great men” is a bit much) move history — not the structures and circumstances they’re standing in.

What will Musk bring? What has he already? Electric vehicles, reusable rockets — what will come from his legacy?

He seems decidedly uninterested in giving away his money. And for a Diablo completionist, that kind of makes sense. He wants to rack up as many points as possible. He wants to be at the top of the leaderboard.

In fact, Bill Gates — a far more charitable person — gave away so much of his own money that he never reached for the trillionaire crown, though he could have. Had he not started his own foundation and given away his fortune, he’d be even wealthier now than Musk.

During this earlier era, the world was also slightly crazy. It was the Gilded Age — the stretch between the Civil War and the First World War. It was the time of corruption in the presidencies.

The Gilded Age was lousy with this kind of thing. Union Pacific Railroad insiders created Crédit Mobilier, a dummy construction company they “hired” to build the transcontinental line and then billed the government at wild markups — paying themselves with public money — and when an investigation loomed, they simply handed discounted shares to the congressmen who might have stopped them. It reached Grant’s own vice president, Schuyler Colfax. Grant’s Secretary of War, William Belknap, ran a quieter racket, taking kickbacks for the monopoly licenses to operate trading posts on military forts, the cost passed straight to soldiers who had nowhere else to shop. And there was Tammany Hall, which ran New York on loyalty and kickbacks.

Just recently, a federal judge indefinitely blocked the creation of a $1.776 billion fund — an “Anti-Weaponization Fund” that grew out of Donald J. Trump’s own lawsuit against the IRS over the leak of his tax returns. His Justice Department settled that suit and, in doing so, stood up a $1.776 billion pot that critics call a slush fund for his loyalists. The administration has since told Congress it won’t proceed, but it hasn’t formally killed the fund — and Trump has mused aloud about reviving it — which is why the judge blocked it indefinitely rather than taking the government at its word.

It’s not exactly the same. The Gilded Age was called that because of the gold-covered crap.

It was the end of an era, but the era repeatedly resisted reform, right up until the huge destruction of World War I. That brought reform to the world — and that reform was inadequate. World War II followed barely two decades later.

Musk is now a trillionaire — the first one ever, made official when SpaceX went public in June 2026 in the largest IPO in history, repricing what he already owned right past the trillion-dollar line. He beat Rockefeller. His accomplishments are remarkable. And the kind of corruption we’re watching with Trump isn’t exactly easy to pull off. You have to be president to direct your own Justice Department to settle a lawsuit with yourself.

But part of me really wants to stop and sit with the selfishness of our leadership. The accomplishments are, undoubtedly, remarkable — yet Rockefeller modeled a kind of giving that generations followed. Rockefeller was also deeply Christian.

I’m not sure what Musk believes — the most I can make out is that his bible is The Hitchhiker’s Guide to the Galaxy — but it doesn’t really matter. There’s a “take all you can, as long as you can” mentality here, accompanied by the refrain that “government is incompetent, so you don’t want them holding the money.”

It’s especially funny to hear the latter from the people in government actively making it incompetent.

How long will this Ayn Rand self-centeredism prevail in our culture? And what will jar us out of it?

Hopefully not a major war.

Now there are conservation efforts to protect that same sea turtle — preserved beaches, protected mating grounds, fenced-off sections for the eggs. The intervening five centuries saw humans move from predator to protector. It is one of the rare times we managed the switch.

The sea turtle is lucky compared to the other creatures we’ve encountered. The woolly mammoth, the dodo — we have a habit of causing extinction wherever we go. And, no, this is not a European problem. Extinction followed humans as they spread across the world over tens of thousands of years. Nearly every place that had its own megafauna — a sprawling, biologically diverse ecology of large animals — watched those animals collapse soon after people arrived.

Humans migrated, in a sense, in search of abundance. We traveled thousands of miles chasing it — temperate forest to rain forest, savannah to swamp, always after the next meal.

The hunger was real. Perhaps you, too, had to read A Day No Pigs Would Die, the novel set in the 1920s in which a boy named Rob helps save a neighbor’s cow and is given a piglet in return. (The book came out in 1972, so I’ll assume the statute of limitations on spoilers has run.) The piglet, Pinky, becomes Rob’s best friend; pigs are notoriously intelligent. Rob dreams of breeding her — litter after litter he can sell. But Pinky turns out to be barren. When winter sets in and the family faces starvation, Rob and his father slaughter the pig they love.

A dramatization, sure, but not a distant one. In Germany during the First World War, blockaded citizens ate what they could find, down to the leather. In the Second, the people of besieged Leningrad were driven to cannibalism.

The human body is built for famine. It’s built for feast or famine, really — for long stretches of hunger after gorging on whatever the kill or the harvest provided. If diabetes and the modern diet tell us anything, it’s that we were never built for abundance.

Consider the lobster. It was once trash food, so plentiful it was fed to prisoners. Only later did it become a delicacy — once we’d caught enough of it to make it scarce.

In fact, most modern scarcity is arbitrary. Very little is genuinely scarce anymore. The obvious case is the natural diamond, a product that somehow still commands a market even as its synthetic counterpart has surpassed it in quality and collapsed in price. Or caviar — and having eaten several rounds of it in various corners of the world, I can say with total conviction that it is not better than a really good cheeseburger.

Consider, too, the two Koreas. In the South — one of the most technologically advanced economies on earth — the birth rate is the lowest of anywhere. In the North, where people still face deprivation and outright starvation, families remain larger and the population has held on.

It’s a paradox. When we talk about AI, we almost always talk in the grand language of abundance. What we rarely admit is that humans are terrible at managing abundance.

• What do we do with an abundance of time? We spend it doomscrolling, wired into our dopamine receptors, eating a “feed” that grows more filled with slop by the year.

• What do we do with an abundance of calories? We get fat. We develop heart disease and diabetes. We build a multibillion-dollar diet industry that never quite seems to help anyone.

• What do we do with an abundance of intelligence? That’s the question of the day. And if the past is any indication, we’ll devalue it — because that’s what we do. We find whatever is scarce and crown it.

There’s a hopeful version of what comes next. In it, we finally turn toward what is genuinely scarce: biology, ecology, the living world. The relationships we actually have — the ones in the room with us — as the online connections reveal themselves to be ephemeral and, in the end, a little silly. In it, the scarce thing we rediscover is meaning.

Because that may be what we’re really short on. A sense of belonging. It’s no accident that church attendance has begun to tick back up.

The danger is that meaning is hard, and scarcity makes people impatient. So we get a lot of simple answers to complicated questions, offered by the loudest voices in the room. That’s the cheap substitute — certainty where meaning should be.

I am writing this as a baby slaps my foot for attention. My baby. I can tell you that the small thing pulling at my sock adds more meaning and purpose and love than anything I have ever encountered — and that no spreadsheet of the economic case against children comes close to touching it.

There was a time, not long ago, when people had to have large families or risk collapsing as a household, a clan, a people. The economic value of a child plainly exceeded the cost. Now the math runs the other way: many would be richer, freer, by some measures happier, without kids. And so fewer have them. The rational case is real. It also misses almost everything.

Which brings me back to the turtle.

Five centuries ago the sea was so thick with them that a sailor thought his ship might run aground. We spent most of the time since proving we could empty an ocean. And then — late, imperfectly, but really — we decided the turtle was worth keeping. We fenced the beaches. We guarded the eggs. We became the protector instead of the predator. The reason why the wooly mammoth didn’t have the same fate is they died off before we realized our folly.

We are about to come into an abundance of intelligence. History says we’ll squander it the way we’ve squandered time and calories and every other thing that stopped being scarce. Maybe. But the turtle is proof that we are capable of the other move — of looking at something we were busy destroying and deciding, in time, that it mattered.

So consider the sea turtle. And consider what, in an abundant future, we might finally decide is worth protecting.

A whole culture of the internet rose up in this kind of soup of experimentation — a post–Cold War soup of American capitalist idealism. And boy howdy did it ever work!

Google, Amazon, Meta, Netflix, Nvidia, Salesforce, and so many more — trillions of dollars of value created during the rise of the internet, not to mention the trillions more in additional markets for incumbents like Oracle, SAP, Microsoft, and Apple. The rise of SaaS, of Mobile, of the Cloud; the rise of content and “creators”; the sharing economy. On one hand, it’s clear that when you give entrepreneurs the ability to just build and grow and find value, they do so. And they create massive wealth in doing so. I think we’d be hard-pressed to find someone who wouldn’t say this is a good thing.

But on the other hand, these companies are far from startups. They regularly rake in billions of dollars. They own entire sectors of the digital economy. The course of capital unrestrained is monopoly. And while tech companies are nowhere near alone in having monopolistic power over their domains, they have become extraordinarily powerful. Today, the largest social media platforms are owned by Meta, and when an upstart — TikTok — tried to take share, the government forced its owners to sell to American allies.

Google dominates search, and with it so much of the internet’s traffic. It owns the largest media property in the world in YouTube (a startling and true claim). It also now runs one of the most advanced robotaxi companies, with ventures into various other areas, including extending human life.

Google — or Alphabet, as the parent company is called — is often considered the good monopoly. Like Bell Telephone before it, Alphabet uses its power to create and invent new and interesting things: Bell made solar panels; Alphabet made large language models (among other things).

Bell was broken up eventually. Its phone lines were deemed similar to the railroad lines that preceded them — infrastructure important enough to keep unrestricted, so that no one company holds power over the others.

Futility

Yet the largest telephone companies today can trace their lineage to Bell Telephone. These include AT&T, Verizon, and Lumen (formerly CenturyLink). They still dominate communications in the United States.

The same is true of the famous “Standard Oil” — the original monopoly breakup. The largest oil companies in the world today were once Standard Oil companies, or grew from acquiring them: ExxonMobil, Chevron, and BP. Indeed, part of what made John D. Rockefeller so wealthy was the breakup of his company into several successor companies, each of which grew well.

Breaking up modern technology companies would be unlikely to create a new marketplace full of thriving startups. It would be far more likely to make even wealthier the handful of people who already own them.

Policymakers know this, of course, which is why they often scrutinize companies’ acquisitions rather than try to break up firms that have grown so massive that doing so would be largely ineffective.

Influence

But it’s not like these companies don’t understand this as well. They’re led by incredibly smart and sophisticated people. And they have deep pockets. Such deep pockets that they influence the legislation itself.

In her book Careless People: A Cautionary Tale of Power, Greed, and Lost Idealism, Sarah Wynn-Williams details how Mark Zuckerberg and the leadership at Facebook realized the power of their platform to help pick political winners and losers. So great was this power that Zuckerberg was routinely in contact with presidents — Obama, Trump, and Biden — and had audiences with political leaders around the world, including in places like Myanmar and the Philippines, which had far weaker democracies than the United States and elected populists and strongmen.

The influence these social media companies wield has effectively blocked comprehensive changes to the original Section 230 — once intended to keep a nascent industry from being stopped before it started, now shielding multibillion-dollar companies from their part in many potentially damaging societal ills. Consider the loneliness epidemic, the rise in mental health issues, depression among teens, increased suicide rates, and so on. I’m not saying all of this has to do with social media, but any coherent conversation is also laced with what’s been termed the “tobacco strategy” — using “studies,” “experts,” and “scientific research,” often paid for and promoted by your own company, to dispute and fight, every step of the way, any claim that your company or your product is doing harm.

American Money

This cycle is perhaps as American as apple pie. Companies rise up, in part, from governments creating the market. Then, once they become world-class money-printing machines, they lobby that same government voraciously to prevent any significant changes to their business models — which, left unfettered, tend toward that board game everybody hates.

Right now, we’re watching the descent of the parabola for the internet companies. The angel of death is well known; it’s AI. AI’s capabilities can destroy many a SaaS company, but even the large and well-known companies are not immune.

Take Meta, a company this post has slightly vilified. How will people interact with the hum of low-level entertainment? Right now, the “doomscrolling” of youths especially makes up the majority of its revenue.

Not long from now, people could basically get away from their screens entirely. At some point, LLMs will become new companions — on-device, personalized, capable of responding to low-level messages and performing autonomous coordination.

How much of a day is taken up by “Hey, where’s that document again?” Or “Hey, we don’t have access to X — could you give me permissions?” A forgotten attachment, and so on.

“Screens” — these little sheets of glass in front of us from nearly the moment we wake until sundown — are the anomaly in human existence. For the vast majority of human life, entertainment was tied to community; it was working with one another. AI presents the opportunity to leave the digital world behind.

It also provides a chance to create even more immersive worlds. Generative AI can essentially realize the dream of virtual reality with far less compute. This will be a desire for many. As St. Vincent sang, “People turn the TV on, it looks just like a window.”

And, importantly, we’re currently in the early phases of a new technology. We’re at the start of the cycle. Now is the time for optimism, for unrestrained imagination — what can we build?

Eventually, it’ll follow the path of so many before it. Monopolies will be created.

The cost

We do not know what toll AI will take in the near term. The most obvious is the jobs that will be replaced by automation — the companies that will fall to the new technology. This is not insignificant, and it’s happening quickly enough that it’s hard to believe we’ll be able to handle the changes well.

Perhaps it’ll exacerbate our loneliness, our distance from one another. Maybe it’ll lead to greater dissonance. More crappy technology that’s basically meaningless and asserts that people are consumers first and foremost.

And as mentioned, the companies that emerge today will build monopolies tomorrow. They’ll consolidate their interests, then hold their leads.

I decline to accept this fate. Companies will win the market, but they will not define who we are. Humans are so much more than what they consume. They are creators. They are builders. They are a full spectrum of emotion. They are lovers and friends, and each of us plays a role. Sometimes we’re the villain and sometimes we’re the hero. We’re wise at moments and, at least for me, fools most of the time.

AI, to me, represents a chance to take back the digital world from a few companies that have made it just kind of a crappy place to be.

“You may say I’m a dreamer, but I’m not the only one…”

You might guess coffee intake, but not “alcohol consumption.” On the former you’d be correct; on the latter, wrong.

Escape

When people moved to cities following the extreme productivity increases brought about by the steam engine, there was an onslaught of terrible conditions. To manage life in those, people turned to alcohol. The drink of choice in England was gin.

For a time, London was a place where anyone with a still could manufacture cheap, floral booze — the 18th-century “Gin Craze” that left the city very nearly drowning in it. That thirst followed people into the industrial city. In the United States, alcohol consumption peaked in 1830 at a staggering 7.1 gallons of pure alcohol — 200 proof — per adult.

For context, that’s 1,515 standard drinks. Or 4.15 drinks every single day.

Alcohol consumption was higher then than it is now, and as is often reported, people drank to avoid tainted water. But four drinks of gin a day meant one thing very clearly — people were trying to escape.

It’s no wonder that Her Majesty’s forces could recruit men by the hundreds to go off to foreign lands and fight or explore or… Lord knows exactly what. Being somewhere, anywhere, was better than a crowded hovel of an apartment with cramped, dangerous conditions and an alcoholic, abusive family.

The Rhyme

Accompanying it — a rise in prostitution. An 1843 Manchester police report documented 330 brothels containing 973 beds. Local missionaries and medical professionals tracking the data estimated that at least 1,500 women in Manchester relied on prostitution to “eke out the deficiencies of their ill-remunerated labour.”

In New York, contemporary estimates put the number of prostitutes in 1830s Manhattan at roughly 10,000 — though no one really counted, and the figures were as much moral panic as math. As the city grew into a commercial powerhouse after the Civil War, those estimates only climbed. At one point, it was believed, around 2 percent of the city’s women relied on prostitution.

According to some estimates, among American women aged 18 to 34, the share working as “creators” on adult platforms is likely somewhere between 1.6 and 2.5 percent. Rates of addiction and drug misuse among people 12 and older are climbing too, especially with cannabis. From 2021 to 2024, the share of Americans 12 and older with a cannabis use disorder rose more than a full point — from 6 to 7.1 percent. Some of it grown at home, the way the gin once was.

It appears we’ve already begun to rhyme with the past.

What Is the Government For?

So much of the industrial revolution has not been dealt with from a policy side. Today we, rightly, have permits and policies that prevent the kind of slumlording that dominated city life for the early decades of the industrial revolution. But we don’t have a good answer to who should be responsible for health care, for example. Is this a “right” or a service? Who should pay?

Take a seemingly innocuous item — public transportation. It’s almost universally good for countries to have fast, reliable ways to move people and goods around. Many countries — Japan, China, various parts of Europe — have high-speed rail, which lets millions of riders a year efficiently cover huge distances without the hassle of flying.

Even inside cities themselves, an effective public transportation system aids the local economy. People can use buses, trains, or otherwise to get to work, buy groceries, visit other parts of the city, and more.

And yet, it’s hard to find something more controversial — on a small scale or a large one. Country voters don’t want to subsidize cities, and they don’t want to sell their land for huge-scale projects, like the high-speed rail in California that is the definition of a government boondoggle.

This large question — “What should the government be responsible for?” — plagues many democracies. Quite literally, the conflict remains between the people who went to the city and the ones who stayed in the countryside.

A New Revolution

And now, in the midst of this, we have a new revolution. AI promises to expand humans’ mental productivity by some 50 to 100x. The steam engine expanded our physical productivity; now AI is expanding our mental productivity.

The fear is that people will be forced into an economic reality with few options, and so they’ll find their way into ever-increasing economic choices that are difficult at best and problematic at worst.

When the Tools Outran Us

Eventually, government — especially in cities — began to take the needs of the city seriously. They instituted trash collection and public transportation and zoning. They cracked down on alcohol consumption and prostitution. The rules were far from perfect, but they made life far, far better.

More importantly, society caught up. Opportunities caught up. It became clear that college and additional training were necessary, and access to those institutions mattered. Innovation brought about democratization — not just in places like Ford’s assembly line, but also in the index fund. Each time, access to capitalism opened up a bit (and people made way more money).

But it also set up a moment when the technology outpaced people’s ability to handle it. There were tanks and bombs and machine guns, cameras and chemical weapons. Fighter jets and satellites. Bombers with munitions that could, as Kurt Vonnegut would write decades later, turn Dresden, Germany, into the surface of the moon.

Society couldn’t handle the biggest crash of them all. The Great Depression — the fallout of which fed the rise of fascism and the deaths of tens of millions of people.

It’s a stretch to say that all of this was because of the steam engine. It’s not. But in humanity’s history told through the lens of technological progress, time and again it’s clear that scientific and technological progress outpaces our ability to manage it. The gap between those things has brought disaster.

Updating the Apparatus

So what exactly should we do? The world is still trying to catch up to the internet and social media — a revolution that wasn’t quite the steam engine, but definitely was the printing press.

The best thing would be to accept that first attempts will likely not be right. There will be many scenarios where well-intentioned policy is basically bad policy. The politicians and the public need to get:

• much better at adapting;

• much better at experimenting;

• much more comfortable with being wrong.

There’s a dream state for the world that any techno-optimist jumps to. And it is, to paraphrase another Kurt Vonnegut quote, a place that is beautiful and nothing hurts.

Without a doubt, the world will and can be abundant. It can be peaceful and full of opportunity. It can exceed even our wildest expectations.

But we won’t get there the way we are set up today. If the industrial revolution is any indicator, everything will change. What will education be like? What about city life? Will we want to live in suburbs? What will beauty standards be? Will celebrity be the most attractive thing, or anonymity?

The truest thing, to me, is that we need to update the apparatus for policy change. This is a fundamental need. Outside of this, we’re in for a long, long period of change.

Sebastian Haffner, the author, fled in 1938 — before Germany invaded Poland — with his pregnant Jewish girlfriend, whom he would later marry, when it became clear that his homeland had transformed into something unrecognizable.

Throughout the process, it continually seemed to Haffner like things would return to the way they were. They, of course, didn’t — until nearly six years of war brought the gnashing teeth of Hitler’s war machine to a stop.

There’s a lot one can observe, almost passively, in Haffner’s writing. “Wow, it can happen anywhere,” or “the people never expect it to happen to them,” but the biggest wonder, to me, is that Hitler’s Germany was so hungry it could devour up the whole land of Germany, then its neighbors, and then much of Europe and beyond before the lie was exposed for what it was — that the idea that there was some genetic basis for superiority in a group of people was, patently, false.

This reminds me of an old adage, often attributed to Mark Twain (though it more likely traces back to Jonathan Swift): “A lie can travel halfway around the world while the truth is putting on its shoes.”

The simple lies win first, it seems. People are easily divided, and they can find reasons to dislike the other, or to make each other into the other fairly quickly.

I’m guessing you already know where I’m going with this. There’s a strong streak in the United States right now that’s decidedly anti-immigrant (they would say illegal immigrant, so to honor them, I’m including this parenthetical). They want to “round them up and kick them out,” but before them one could argue, as many have, that the other included Black Americans, Gay Americans, Women, the Irish, Germans, Italians and Catholics.

The other, too, seems to include billionaires and technologists, creators of AI. The other lives on the coasts or in “the flyover states” (one that would personally irk me whenever I would hear it on the coast); everybody claims to be more clever, to have solved the problem. Everybody seems to have turned on one another.

Everyone is, it seems, slightly ageist or classist. They’re all a little bit racist, easily offended, put off by the hot new trend, or, simply, scared.

I remember a friend and I found ourselves having a confusing debate. On a one-to-one basis, people were probably kinder than they ever had been. But on a macro level, people seemed more distrusting and downright terrible to each other than they had ever been. The dissonance there was… shocking.

In his TED talk, Rich Benjamin — who is Black — recounts a resident of an all-white community telling him, in essence, “I don’t have a problem with you. I just don’t want to live near you.”

Benjamin describes the white people he met as genuinely affable.

When did Haffner, I wonder, see average Germans become OK with smashing Jews’ windows?

And while this is truly terrible, and a lingering question, this is the kind of thing that the left points at all the time with some smugness and says, “see, we’re preventing the bear from eating us all with our anti-fascist rock.”

And at some point, the right created their own rock, which is, “we prevent communism by making sure that wealthy people aren’t taxed. Ever.”

And yet, if you found yourself in a conversation with the average American, you’d probably have the same experience that Benjamin did. It would be mostly pleasant. You’d see each other, I hope, as people, as fathers, mothers, brothers, sisters, sons and daughters. You would, I hope, not find a reason to view them first as a Democrat or Republican. Or as a Black person. Or as a German. Or as a Jew. Rather, I hope that you’d find dignity in the person next to you.

Then why is it so hard for people to find common ground when we suddenly do apply labels to them? These distinctions are exactly what weak leaders use to provide simple answers to complex problems. Because it’s far easier to say that it’s the immigrants’ fault than an interdependent weave of well-intentioned choices hollowed out the middle class and then most of the midwest due to demand for profits on the one hand and overreach from the unions on the other. And those same decisions led to an insurgent Asia and now an absolute economic powerhouse in China, a country that literally manufactures everything the world needs in such abundance the only rational policy seems to be tariffs.

Kick out the immigrants is just tidier. It’s simple. A thing that tired and worried minds can rest upon.

But this is how dystopias happen. We take the simple choice. We allow for ourselves to continually be fooled. The thing that shocks me about Haffner’s writing is just how long the Germans were at it.

Hitler came to power in 1933, but the Nazi Party’s roots reach back to 1919. And while the term “eugenics” was coined in Britain, it was the American eugenics movement — and especially California’s mass sterilization program — that directly inspired the Nazi Party. The path to “racial purity” was, in some ways, paved in the U.S.

So, then, why didn’t it take root here? Well, for one thing the United States has always been a bit of a rascal. It’s constantly building itself up and tearing itself down. It’s forever under construction. Anytime you have one person saying “God Bless America,” someone else is saying, “not so fast.” And both have the right to do so.

It’s because you can sit down next to someone who you should, on paper, disagree with about everything, and still find a way to make each other laugh, or hear about their pain, and become slightly whole yourself again.

It’s because, at the end of the day, as Bob Dylan sang, “Even the president of the United States sometimes must have to stand naked.”

And we all have the right, each one of us, to say “man, what a fool.” Together, or separate, from many, one.

America, the land of milk and honey. America, the land of blood and tears. America, a home for you and me.

Don’t let them divide us. We’re all the same in the end.

The machinery to make this work might make many queasy, and others question the civil rights implications. Suspend disbelief for a moment, and say that if people want expedited permitting, faster licensing, digital passports and more, they also need a bioidentifier on file with the government. Verification then becomes very simple — a scan of the eye or fingerprint.

There will be plenty of discomforts about a government using advanced technology to become far more efficient. Overreach is easy to imagine. Predicting violent crime with 99.99 percent accuracy — would you implement it? What about for crimes against children? This was the plot of Minority Report, and I remember distinctly telling my philosophy teacher father that losing 1 in 10,000 people, while preventing the crimes the other 9,999 would commit, was, in short, worth it. He disagreed. As would the ACLU.

Still, let’s say that AI makes a version of this real. Crime rates fall. Planning and permitting become trivial. The bureaucracy that requires many hands to shepherd anything through gets dramatically thinner. And let’s say, somehow, the government doesn’t abuse the technology.

I know — the thought of a competent and well-functioning American government strikes many as not just laughable but downright frightening. The concern often gets whispered in a single string of numbers: 1984. And many of our institutions, not least the separation of powers, were specifically designed to thwart governmental efficiency.

Still, with a highly functioning bureaucracy, the broader question is what the government can devote its time to. In effect, what is uniquely human? I think two things:

• The process of proposing and debating legislation. This can and should be fueled by AI research, experimentation, and scenario planning, but the process itself must be dictated by humans — because of our humanity, our ability to empathize and reason.

• Community organization and connection. What can the government do to foster deeper connections among neighbors and citizens? It does very little of this today (the usual refrain is that it shouldn’t), but a series of policy choices could enrich communities far more than we do now.

Start with the first.

Politicians are elected to represent the people. Mechanistically, the system often feels broken, but the truth is it works — potentially too well. When the populace is angry and demanding change, they find candidates who are outsiders and aggressive. They throw them at office like, as Michael Moore put it, “a molotov cocktail.”

The change candidate is then booted out when the change doesn’t come, or comes too slowly, or any of the other ways this typically goes. The party usually goes with the candidate. This cycle can take a while, but it’s constantly weeding out bad ideas until the ones that actually make a difference survive.

This isn’t far from what the philosopher Karl Popper argued about open societies — that they must go through a perpetual process of piecemeal experimentation, where ideas are debated, tested, and discarded.

We could say the current administration’s anti-immigration stance is, essentially, being tested. And the voters not only asked for the test but, increasingly, are repulsed by how it’s being carried out. The same goes for the tariff policies, or lower taxes, or any number of the administration’s signature moves.

It’s frustrating that democracy doesn’t feel more technocratic — plenty of policy wonks warned against both tariffs and the immigration crackdown. But this is, in a sense, how democracy functions. It’s not about running scenario planning; it’s about convincing the population that your ideas are best. That’s a human thing. AI may make it more efficient, but the messiness is important. The electoral process and the sloppy testing of policy should be preserved.

The danger of handing policy to AI is efficiency — that AI will choose the most efficient solutions rather than the most humane ones. It will optimize for GDP growth instead of helping a mother take care of her children. Perhaps it could manage both. And our democracy is hardly without major flaws — but it’s at least self-aware enough to keep trying to correct them. The mother can vote, after all.

The concern with AI taking too big a role in policymaking, without human politicians in between, is that it would silence voices, optimize for the wrong thing, and, in effect, remove self-determination.

But AI could fundamentally speed up the implementation of laws. This is the kind of thing it should do. It should make gun violence less likely. It should make IRS tax collection simpler. It should make permitting straightforward and seconds long.

This is complicated. Laws are often written to be vague, and optimizing around vagaries almost guarantees some bad outcomes. Implementation — even writing a law for an AI rather than for a human — is not simple. Judges and the Supreme Court effectively exist because laws need interpretation.

Still, let’s assume AI has created significant savings in government — potentially billions of dollars annually. What, then, is the core function of government beyond setting policy? If not enforcement, what could it shift its focus to?

My answer: it should help people form deeper communities at the neighborhood level.

This was for the longest time deemed unimportant. The serious work of government was first ensuring that any use of violence was essentially only done by the government, and then making sure basic services were met by a shared pool of capital. The first goes back as far as government itself. The second — deciding what counts as a service in the public interest — continues to be debated today, but largely started during the shift to industrialization.

What I’m proposing is a third role: government as a steward of community. The structural pieces could look like this:

• Funds distributed at the neighborhood level, dispersed by neighborhood captains or leaders and earmarked for hobbies, clubs, and groups — bowling leagues, garden clubs, book groups, whatever the neighborhood actually wants.

• Funds for small, local chapters of community-based organizations — the Rotary, the Lions Club, or new organizations that don’t yet exist.

• Training in the basics: how to recruit members, found a club, run the books, keep things going past the founder’s enthusiasm.

• Physical infrastructure — meeting halls, low-cost rentable spaces, the kinds of places where a club can actually meet without renting a hotel ballroom.

The goal is to make it dramatically easier for people to spend real time together.

The need for this isn’t theoretical. Robert Putnam documented two decades ago that Americans were “bowling alone” — the post-war scaffolding of fraternal orders, civic clubs, and church groups was disappearing, and nothing was replacing it. Since then it’s only gotten worse. The Surgeon General called loneliness a public health epidemic in 2023.

If AI eats white-collar work the way some of us expect it to — and that’s the world I write about elsewhere — community is not a nice-to-have. It’s the thing that keeps people anchored when the old anchors of work and routine come loose. Flint, where I’m from, learned this when the auto plants closed. The economic damage was real. The community damage was worse, and it lasted longer.

A government with the bureaucratic load lifted off its shoulders has a chance to do something genuinely new — not just deliver services, but help people deliver community to each other. That’s the work AI can’t do for us. That’s the work worth doing.

The ring has powers wholly familiar to a Lord of the Rings fan — turn its setting and Gyges becomes invisible to those around him.

Gyges quickly realizes the power of his newfound treasure. He travels to the royal palace, seduces the queen, murders the king, and usurps the throne for himself.

The story predates Tolkien by a couple of millennia. It’s from Plato’s Republic, written in 375 B.C., and serves as the philosophical entryway to a discussion of justice and the social contract. Plato’s broader point was that it didn’t matter who you gave the ring to. The outcome would be the same. They would find themselves on the path to corruption, because the ring removes the only two things keeping them in line — reputation and law.

We have a strange mythification in the United States of the wealthy. They earned their money, the thinking goes, and with that they get pretty much whatever they want. The Wolf of Wall Street is a hedonistic glorification of reckless debauchery (it’s also absolutely fantastic cinema).

There are often questions about whether someone deserves their money, or whether the right personalities emerge from the hellfire. I’d argue it doesn’t much matter who you are when you go into the process. The ring of power does the same thing to everyone.

Which isn’t to say the wealthy operate without checks. An entrepreneur would argue, rightly, that they are kept in line by the same two things that kept Gyges in line before he found the ring: reputation and law. Lose either and the floor falls out.

This is exactly what dark money does to American politics. It removes both.

Take the NRA, which spent $180 million in the 2024 election cycle and only $10 million on politics directly. That figure might appear high, but it’s actually a significant decline from previous cycles (the group spent $54 million in 2016). The real money I can’t track because it’s so-called “dark money,” bundled into political action committees and never broken out. In 2024, that amounted to $1.3 billion.

What are those mailers? What are those issues? What exactly are the Super PACs and the people contributing to them paying for? It’s really, really murky on purpose. Groups like the NRA originally formed around efficient ways of combining money from businesses and individuals who shared a common interest — in their case, gun rights. The cause was named. The money had a face.

But now the common unifier is often a candidate, or a much vaguer cause (don’t raise taxes, say). And the vagueness is the point. It gives a flexibility that is invisible. We’ve taken away the reputational risk and the legal framework for policing our politicians, and what’s left is a never-ending barrage of mailers and money spent on a few pillar issues that never quite get addressed anyway.

Meanwhile, the politicians enjoy the scapegoat. They don’t have to act on controversy. They don’t have to touch the difficult and thorny topics. They get the ability to say “I told you so,” even when they didn’t — because they didn’t say anything at all.

This is the banality of it. Not cigar-smoke conspiracy, not villains in a back room. Just a slow, structural disappearing act. Money without a name. Causes without a face. Officials without a position. Each piece, on its own, a fairly boring procedural detail. Together, a ring slipped onto the finger of an entire political class.

Plato’s point, remember, wasn’t that bad people do bad things when given the ring. It was that anyone does. The shepherd was, by all accounts, a decent man before he found the chasm. Decadence, in Plato’s telling, isn’t a character flaw. It’s what happens to character when no one is looking.

The scary part isn’t that the people in power are uniquely corrupt. The scary part is that they don’t have to be.

She is “working class” in America 2026. And while one could claim her situation is unique, there are many people with similar circumstances. Some live with family for many, many years. Others will work menial or many jobs just to eke by. Some aren’t fortunate enough to have parents who can take them in and help raise their children.

She’s elated if she earns $30,000 (pretax) in a year. A $100 monthly expense — like a karate class for one of her kids — is a big stretch for her. And, again, she’s not alone.

In fact, 28 percent of the U.S. population earns less than $30,000 in a year. The average individual income in the United States is $62,000, with most households having two earners. This brings the average household income to $121,000. The median household income, a better reflection on where we are as a country (the really low and really high numbers don’t skew it like they do the average), is $83,000.

There are people who earn $83,000 a month. These folks are wealthy — they’re in the top 1 percent of earners. But they’re nowhere near the top 0.1 percent who make $1 million a month. But who are they? How do they get there? Could my niece be one of them?

Well, maybe…

The truth is that the highest earners own the businesses. And I’m not referring to a Silicon Valley entrepreneur. Those folks, if you’re wondering, tend to make $120,000 or so a year — they do that to show that they’re disciplined with the cash that VCs give them. They make that money in an effort to grow the value of their stock.

Rather, it’s the people who own small businesses, who have grown those businesses over many years. They’re the unsung millionaires in the economy. And, they tend to be in a place now where they’re selling their businesses. They’ve grown them over 30, 40 or more years. And so they’re turning to places that have cash and want the consistent cashflow from well-run businesses.

They’re turning to folks who have a lot of money. Like literally trillions of dollars. They’re turning to people who make it their business to find ways to make vast amounts of money. These folks buy businesses like the restaurant that my niece works at, or the veterinarian office where I take my dog, or the medical office where I take my two sons.

If you’re not familiar with it, there’s an entire apparatus designed to aggregate the vast assets and wealth they’re entrusted with, and transform it into ongoing cash flow. Dear reader, it is time I told you about the world of Private Equity.

Today, many of the best small businesses — the ones that have found the best market, produced the most reliable products, or crafted compelling moats — are purchased by private equity firms. They come in, buy the business, seek to take the cost out (this happens in a variety of ways and is slightly euphemistic), and essentially turn them into cash cows.

Do your kids need braces? They could be getting them from a doctor paid by a private equity firm. See, an orthodontist could, over their lifetime, start up a family practice. They could pick a good neighborhood. They could even live in said neighborhood. The ortho could build a reputation and brand and decent clientele putting braces on the kids in the neighborhood. In some cases, they could be the friends. And over 40 years of consistent returns and growth, adding more chairs, staying on top of improvements (Invisalign, etc.), the ortho could have a practice they’re proud of.

But, insurance is a real pain in the neck, and man do they want to retire. Our ortho just got approached by these nice guys in vests who offered to buy them out for a few million dollars — roughly four times the revenue in a year. What a nice retirement plan. And, they’re going to keep the business largely the same.

So the ortho takes the deal. They go off to the Bahamas and sail around and chase after colorful fish and fly back up for the holidays sometimes. The private equity firm, in turn, looks for opportunities to reduce staff and cut costs. In some cases, too, they could load the practice with debt and pay themselves a dividend from the proceeds. They often use technology to reduce costs through centralization — let’s say they have a dozen orthodontist offices, what if they could put them all under one megapractice, which then handles most billing and invoicing?

This is the kind of thing that happens regularly for really well performing firms. And I don’t begrudge anyone in the process. It’s one of the various financial apparatuses that exist out there. And to me it’s honestly pretty incredible. People have a way to sell their businesses that they’ve done a brilliant job in building.

And after the sale, if private equity doesn’t do a good job with the business or doesn’t want to continue it, then the business dies, and the world turns, and some other enterprising person rushes in to take that space. This is a free market and an important part of creative destruction.

The point is that this world, which has billionaires on private jets buying up small businesses, is completely divorced from my niece who is barely eking out a living.

And that’s not to say that my niece couldn’t find her way into starting a company that private equity one day would buy. Her sister, coincidentally, did start a business as a barber. This one day could be many barbershops, and robotic barbers could man the stores.

This is one of the scenarios that people lean on when they’re considering a future with AI that doesn’t result in some kind of dystopia. They believe that mass entrepreneurialism will save us.

But it’s not clear to many how they would make the leap. Our system is not set up for this kind of thing. We don’t have efficient public grants (they exist, but in a byzantine system that I’ve never figured out), nor do we have an education system designed around a future where people do not perform the kind of drudgery that employs so many today.

Private equity firms are at the forefront of using AI. There are non-fantastical and immediate benefits in places like sales and marketing, or an AI receptionist instead of a real one. And one day perhaps there will be a fleet of robotic teeth care workers, not human dental hygienists.

Today, many technologists earn triple my niece’s salary to aid private equity in the transition to a more automated future. This will lead to greater proliferation of the technology and, ostensibly, a potentially automated wait staff at your next visit to a chain restaurant or accounting firm you didn’t know was owned by private equity.

This too might be a bubble. For a time, legions of consultants who helped companies prepare for Y2K. It’s hard to say exactly what will stick and what won’t. In the 90s, there was massive economic expansion as the internet made personal computers “must have” items. This resulted in the “roaring 90s,” which also benefited from a rise of optimism following the fall of the USSR. Capitalism, it had seemed, won a long, cold war.

There was also a growing population, permissive policies, no clear international challenger, obvious sectors to grow into (international), a very cooperative global order, and lingering blue collar positions. Globalization had not yet hollowed out factories in places like Greenville, Michigan, and investments felt plentiful.

Today, there are many economies. The economy for my niece makes $100 a month feel like a massive burden. For many at the top, $100 feels like a tip on a nice dinner — like buying a pack of bubble gum.

This is capitalism, and without mitigation in its ability to monopolize, capital will grow and grow and grow. There are, without a doubt, opportunities. Entrepreneurs and technologists can create value and gain wealth for themselves and the investors who work with them.

Some might argue that my niece could have done the same. And maybe she could have. Maybe she should. But also, let’s not ignore the vast barriers that exist in establishing a network and expertise in a nascent industry. Today, many break in by going to elite universities, but even getting into those places requires a certain level of privilege.

Needless to say, I’m not convinced it’s entirely an individual’s fault. Besides, does it really matter? The fact is that many are scraping by for grocery money. This is not the responsibility nor the problem of Private Equity. Nor the technologist. This is a societal problem. What does it mean that a whole year for a mother of two working three jobs is a tenth of a bonus payment (or less) for someone who works in private equity or AI?

It would be tragic for our society, which soon could be enriched beyond our wildest dreams by a new technology, to not find meaningful ways to ensure that all of its members of society are also enriched.

It’s not easy, though. Bad ideas and bad policy are easy. We have a lot of it.

And with these new problems and technologies, we should seek to avoid perpetuating a broken welfare system. It does mean, though, that we need to find new policies that aid the many workers who will soon be or already are at a crossroads.

Consider the Industrial Revolution. Never before had more wealth been created. Never before had technology so thoroughly transformed the world. Productivity exploded. One person could suddenly do the work of 50 men. I sometimes joke that I came from a family of ditch diggers. They lost their jobs to a machine, but when steam engines made digging ditches easier, we suddenly got a lot more ditches…

My family might have lost work (I cannot verify they actually were ditch diggers, btw), but they found a way, as did every generation, even when the conditions seemed hellish or the odds low.

When the industrial revolution started, the conditions in cities were terrible. It took generations for them to improve through good policy.

We have the benefit of understanding the past and learning from it. We can say with certainty that new policies must be created and that the past is likely a bad predictor of what policy will work for the future. But we can also say that a new type of politician is necessary, one unafraid to shepherd us into the future and respond to the complexity presented by our own magnificence. And leave us no doubt, the future can be magnificent. A world of true automation, where the things of our dreams can come to be.

We’re on the cusp of that. And we should find a way to enable everyone to benefit. I believe my niece, like all of us, is on this planet to do great things — how do we enable her to thrive?

It’s an easy, albeit reductive, thing to say. The social compact of capitalism has been: business owners take on risk, start companies, hire employees as they grow, and those employees in turn spend money on the goods and services of other companies — a “rising tide lifts all boats” scenario.

In a strange twist of industrialization, history books attribute a key insight to Henry Ford on the production of cars — that to pay his workers well would mean they’d become customers for his cars, especially as he used his invention of the assembly line to reduce the cost of producing one.

One could claim that software should have, in some ways, broken capitalism. After all, take a good that is infinitely replicable, with no cost, and put it out into the world. And for a time, this core economic attribute of basically having a fixed cost to create (in hiring of software engineers) but zero cost in distribution wasn’t practically enforced. Windows updates cost, for a time, hundreds of dollars each year.

It wasn’t until the Internet spread broadly and companies like Google and Salesforce (yeah, those guys, believe it or not, they created the SaaS business model) pioneered new ways of capturing value that there were entire categories of businesses created that didn’t exist before.

It was once hard to conceive of a company that could make $402.8 billion in the last twelve months primarily by tracking how people click on websites, and in so doing predicting what they’d spend money on. And yet, here we are. Google (aka Alphabet) is a $4 trillion company.

There are a few things that I’d be willing to say, concretely, about generative AI in the future.

It will create a lot, as in a whole lot, more code. It already has. It will also make a whole lot more writing, imagery, videos and anything else that can now be generated in a few clicks.

In traditional economics, the vast proliferation of things tends to reduce their perceived value. This is considered the “baseball card economy,” as my history teacher used to say. The fewer of something the more valuable they are, or at least considered — this is not exactly true in modern business and economics, but the point stands. Commodification is the destruction of value.

AI will shift white-collar labor dynamics in a way that, so far, they have been untouched.

Industrialization

For a time, blue-collar work in America was a dynamic economic engine. There were car mechanics, line workers, assemblymen. There were contractors, plumbers, maintenance people. There were tire manufacturers and line cooks.

People were paid handsomely to do routine tasks on an assembly line. These tasks were repeatable, and legions of consultants and engineers spent careers finding ways to make them simpler and simpler. Ironically, that made them easier to automate. Then, when the work was picked up and placed in Japan, then Korea, then (later) China, automation became the expectation.

These countries became experts at automation. Today, there are whole dark factories that build massive quantities of high-tech industrial goods without a single person. The automation that began in the 70s has neared completion today.

Parallels

Today, we’re facing the start of a new era. For the longest time, a kind of white-collar work seemed beyond automation. That drudgery is already being absorbed by large language models.

Take research tasks, like reading through earnings transcripts, or the news. These are done more simply and effectively by an LLM.

But take even low-level code. Building a website is now trivial. Making it SEO- or GEO-optimized is equally trivial. Maintaining it too. As recently as the start of this year, I was quoted between $50,000 and $75,000 over a quarter by multiple design firms to build a website — not even SEO-optimized. Today, Claude does both for the price of a subscription: $20 a month, or bump it to $100 if you run out of credits.

I think about the factories that used to herd people in. Lines of grey-faced workers streaming through fences, punching cards. Those people were automated “away.”

The country tried to shift. It suggested college for everyone. Education systems that gave pieces of paper for, essentially, technical vocational work. There were degrees to basically learn crappy user interfaces for standardized products. Entire courses were built to learn Microsoft Office — these were eventually taught to kids in high schools as a way to modernize workforces.

And liberal arts degrees became a fast track to latte-making for those who hadn’t finished high school or gotten a GED.

This educational system took a long time to get to — some could argue that it started at the end of World War II, 80 years ago. And now, it’s expensive, bloated, and inefficient, and, increasingly, it’s turning out workers for a white-collar world that appears to have questionable prospects of a future.

The drudgery will be automated. LLMs can do it better. They will perform a level of white-collar work today done by people or services firms. Tomorrow, just like blue-collar automation, entire chunks of work will disappear — specifically the repeatable, easily articulated kind, the work with process-flow documentation and the like.

Beyond the scary stuff

This kind of work is the kind of work that people, rightfully, panic about. Automating this kind of white-collar work is hard to quantify. If we look at a company like Block, it’s roughly 40 percent of their workforce. Perhaps for a manufacturing company, it’s less.

Yet consider: when the world first began to be automated, a new profession was born — “consultant.” IBM launched a family of computers in the 1960s meant to help companies digitize.

For a time, IBM consultants in their black ties and suits were iconic. They were the future. They were the Google of their day. In 1960, they employed 104,000 people — in 1970, nearly 270,000.

This is not to mention the thousands of people employed by Apple, Microsoft, or the broader IT industry that grew up as a way of making sense of these new systems inside of a company.

The broader point is generally turned to when people panic about the impending job loss: that there will be new kinds of work. Maybe, indeed, there will. It’s impossible to say exactly what, though.

And in the case of blue-collar work, there really never was, unless you count being a stocker at Wal-Mart or a fulfillment person in an Amazon warehouse. And here, too, the companies are working hard to automate those roles out of existence.

So the honest answer is that a lot of work is going away, and we don’t yet know what, if anything, replaces it.

Abundance and its asymmetries

Inequality, as a general rule, doesn’t follow a “normal” distribution (like height). It follows a power-law distribution — meaning there are outliers that sit way, way outside the clustered norm.

The emphasis traditional economics places on income will be especially challenging when income itself approaches zero.

Take seriously, for a moment, a world where the price of every good and service is essentially zero. An AI-enabled world makes this possible. Distributing software costs nothing, and core needs like energy become far cheaper with abundant renewable generation (solar, wind, nuclear, and so on). And even when coal still needs pulling, robots can do it — fixed-price, multiplying human output, requiring only an initial investment.

This system, though, would require initial capital. There would be owners — people who start the companies that build the robots and the AI, or who invest in them. They would want to recoup (and then some) their investments. The result would be massive, unparalleled inequality without worker recourse. Historically, from the Black Death through the sit-down strikes, worker scarcity led to greater societal inclusion and equality. Now, robots threaten to erase that dynamic entirely. If there’s no need for workers, there’s no need for inclusive systems. Scarcity gets solved by more robotics or software.

This is a world of abundance. Potentially, abundance in everything — intelligence, resources, goods and services. A world where, hypothetically, anything anyone can imagine would be theirs, either digitally or (within the confines of physics) in reality.

New value, new economics

So what, then, becomes of value? Well, if anything is possible, it’s the ability to create. It’s the ability to create things of universal demand and value — it’s the ability to find problems that should be solved and solve them in an elegant way.

This is of value for large companies and for entrepreneurs. It’s of value for the world. There are marketplaces to be built, and taste will be valuable for knowing what to leave out instead of what to include.

There was a time when clay was free. If it existed in the soil around you, you could have it. Then it got expensive. Then it got cheaper. Now, it could become free again.

Consider the arc. Clay was there for anyone willing to dig. When people moved into cities, getting it required knowledge and someone who owned, ostensibly, a clay mine where miners worked by hand. Eventually it became highly processed, and pulling vast amounts out of the earth was trivial. Now we’re on the cusp of huge-scale, essentially free clay.

So what was valuable when clay was free? The pottery and the art made with it. What will be valuable when clay is free again? The pottery and the art made with it — the expression drawn out of it.

But how on earth do you value something that moves the soul? How do you capture that craftsmanship? How do you build a GDP from artisans?

I don’t know, but I suspect we’ll find out.

These roles are heavy on research, heavy on orchestration and architecture, and heavy on interpersonal skills. As a Financial Times article recently put it, humans are becoming a kind of “glue” — theyGiver hold projects, organizations and other people together.

Companies will, without a doubt, have fewer people in them. There's a large category of white-collar work that is routine, repetitive and easily described: making dashboards, invoice processing, standard status updates, first drafts for contracts, bank reconciliation, meeting coordination, travel booking. All of it will be automated.

But that doesn't mean that a travel agent will be automated, or an FP&A person. The tasks go away. The job changes.

Let's take an office — how many people are really the creative, imaginative ones? The ones digging deep into their work and surfacing things that actually matter? Let me be more specific: Take a financial analyst. Traditionally, they'd have some skills in SQL or data analysis. This has been a prerequisite for the role, and a lot of the training they receive — from their four-year college degree to their early positions — is oriented around it. Over time, they learn what to look for, and their experience guides them.

Now, Claude Code, Codex and many other LLMs can write this SQL and perform this analysis significantly faster and better. This is a skill that's inherent in the profession, but it's not the thing that separates the best financial analysts from the standard data analysts. What separates them is craft.

Claude Code and the like are incredible tools. They can be used in a variety of ways, but so too can a hammer. The same tool can be used in surgery or in building a house — in removing nails or cracking rocks. The craftsman knows how to use it, and much of that knowing is inherent. Just as the financial analyst knows the right questions to ask, the carpenter knows the right place to hit a nail.

This feels a little funny, but it's important — there are craftsmen around us, everywhere. People who take pride in their work, who build things that bring happiness, that make life easier and better.

There's a lot of not that, too. A favorite word right now is “enshittification” — the idea that digital products have begun to get worse over time, or at least stop getting better. There are a lot of really crappy digital experiences out there, and a lot of people want to blame capitalism. But, in part, it's due to a lack of competition — something that a sudden democratization of software development should help solve.

So while the jobs of the future will likely become more and more craftsmanlike, it's also true that whole job categories will emerge that don't exist yet — just as they always have.

Think about it: digital marketing managers, digital ads managers, community managers, digital designers — literally any profession done on a computer. This entire class of work, which now employs most white-collar workers, didn't exist 60 years ago. The advice then, as echoed in the 1960s classic The Graduate, was to buy into plastics — the cutting-edge technology of the time.

Sixty years is a long time, but many of these professions — community managers, social media managers — which are considered crucial to companies today, also didn't exist when I graduated from college. I'm only 38 years old.

When Facebook launched in 2004, it would have been hard to predict that an entire class of employment would emerge around optimizing ads for it, or that whole industries and agencies would specialize in managing ad spend. And yet here we are.

In fact, the careers that people predicted would rise with AI have already burnt out. Prompt engineering, for a time, appeared to be the new frontier — working with the models, coaxing the right output. But it turns out that everyone needs to know how to work with the models, and the models themselves are the most capable prompt engineers there are.

So what will be the new social media manager? The new digital campaign person? How will a company derive $100 billion in revenue each year from the way people interact with a website?

Right now, it's just too early to say. People seem to treat AI as though it's, in some way, the final technology — the last thing we'll ever build. But it's not. It's a tool. A tool for craftsmen to build with. It will enable a kind of knowledge work that is unencumbered by complexity and time and training — but that still requires craft, taste and knowledge.

The trouble is that it's so new we're having a hard time being imaginative about what comes next. When the ditch digger lost their job to the backhoe, I'm certain a lot of ditch diggers were annoyed. But then we got so many more ditches.

That's a silly example, but consider how much more we can build today, and how much more quickly. The only limitation is policy. It's not a lack of capability — it's a congestion of rules around what people are allowed to build.

We would do a significant disservice to the capabilities of AI — and to our own capabilities — if we put policy barriers up now, before we even know what it will enable.

The only thing to fear is fear itself.

Yet still, there are so many areas that have appeared immune to automation. Head to the DMV and get a driver’s license. I once stood in a six-by-six square in the DMV to get my Illinois driver’s license, in one of three places in the greater Chicagoland area that could do so, masked up during Covid. It took forever. The picture was terrible. I have complaints.

In another instance, which I’m sure is both unsurprising and familiar to anyone who has gone through something similar, I wanted to buy a new car. I had paid off my last car. I was going to trade it in. Yet, I needed the title of the car, which I hadn’t gotten or misplaced when I paid it off. So, off to the DMV I went. When I got there, I learned that I needed to have different paperwork than what I had, which I had to request from the state — this was a title change that had somehow not gone through when it was paid off. So, I didn’t buy the car. Instead, I submitted a request to the state to change the title owner from Toyota to me — a process that would take six weeks. Only then could I go back to the DMV, get the title, and finally trade in the car.

Throughout this and so many other processes, I couldn’t help but think how byzantine it all was, and how Amazon could somehow deliver me a coffee maker from Japan the next day, but a simple paper change for my own car took multiple weeks from a city that’s six hours away.

In my marketing brain, civic tech is basically worthless — I mean that in terms of true value of the software. To start, there aren’t many buyers. The folks who derive value from the government tend to mostly be lobbyists or interested in influencing the laws, and they often like to work in the quiet rooms and not in the public.

There are some categories that have traditionally been “money makers” — journalism, for example, served the public a steady stream of information about legislators, but social media has essentially reduced the value of that “news” down to nearly zero. Now citizen journalists are the ones who profit off whatever they put out, often it’s disconnected from the inner workings of the government.

If, instead of making money off the body of government, i.e. the things that they put out, you could, hypothetically, make money from the government itself. Here, again, it’s the kind of pain that would make an entrepreneur shiver. Nobody is harder to have pay than the government. The politicians are the ones who benefit from various forms of obfuscation and are unlikely to want this to change. And, while labor unions can be beneficial (there are studies that show they help protect the middle class), the unions that are the strongest are the public unions. For technology to be useful, it has to be used — and introducing a technology that suddenly would make a lot of the work of an administrator unnecessary would be problematic.

Also, government spending is intensely scrutinized. Any spend on technology, especially startup technology that could fail, would be viewed very harshly by taxpayers who want to know where their money goes.

This is all to say that it’s hard to make money from governments, from almost any angle. So, it’s unsurprising that technology adoption, of the kind that makes an Amazon package zip across the world while a piece of paper languishes in a file somewhere, would be sorely lacking. For the longest time, to build a worthwhile piece of software, you needed a software engineer. Getting one was expensive, especially a top-tier talent that would be necessary to ensure the ongoing quality of the code.

This presents a unique opportunity for LLMs. If LLMs essentially provide a level of quality for code that is at or above the best software engineers in the world, and they do so for the extremely low price of $200, say, a month, then what will likely happen to code? Well, a lot more of it will be written, especially in places that were deemed uneconomical prior.

It’s essentially free to have a world-class software engineer build and maintain a piece of software that can automate things like identity verification and biometric scanning, and provide significant bureaucratic speed improvements to a whole host of tasks.

Imagine all of the tasks that could be done that previously didn’t make sense economically to do:

• Hyper-Local Permit Automation: Small municipalities often lack the budget to build custom software for niche permits (sidewalk repairs, block parties, or signage). LLMs can generate and maintain the logic for these “long-tail” forms without a dedicated dev team.

• The “Plain Language” Layer: Translating byzantine legal codes and “legalese” notices into clear, actionable instructions for citizens. This previously required thousands of hours of manual rewriting; now, an LLM can act as a real-time interpreter for every government document.

• Legacy System “Wrapping”: Many state agencies run on 40-year-old COBOL backends. Instead of a $100M “rip and replace” project, LLMs can be used to write and maintain sophisticated middleware that allows modern front-ends to talk to ancient databases.

• Mass-Scale Transcript Synthesis: Automatically summarizing thousands of hours of public school board meetings, city council sessions, and zoning hearings into searchable, categorized datasets for journalists and citizens.

• Personalized Benefits Navigation: A “concierge” that cross-references a citizen’s specific situation against city, state, and federal programs (SNAP, housing vouchers, tax credits) to find eligibility overlaps that human caseworkers are often too buried to spot.

• Proactive Infrastructure Triage: Analyzing thousands of citizen-submitted photos of potholes, cracked sidewalks, or downed lines to automatically categorize severity and route them to the correct department, reducing the need for manual intake.

• FOIA Fulfillment Automation: Redacting sensitive information (PII) from massive document dumps and organizing files for Freedom of Information Act requests — a process that currently consumes millions of man-hours.

And this is just the start!

Government is an apparatus built on language and administrative tasks — two things that are likely to be vastly sped up because of LLMs and generative AI. There is no shortage of efficiency opportunities, and they’ll be economical because the code itself is so inexpensive to write.

Imagine a more interactive government, brought to you by AI.

Or imagine all of the people who could benefit from sound legal, medical or financial advice but can’t pay for it today; or a companion that would make people feel less alone. All of this and so much more becomes possible as the creation of software drops close to zero.

We’re on the cusp of truly unprecedented changes in the world.

But one would be remiss on talking about these huge changes without talking about the number of people it could affect. For many places, local government is the largest employer. These jobs aren’t exactly high skill, and many of the employees get them after a bit of college or a high-skill diploma, but they put food on the table for families.

This is the crux of the concern — we’re transitioning to a profoundly different world, and we’re doing it so quickly that many are at risk of being left behind.

Dear past me,

The crash came swiftly, or so it seemed. In retrospect, the foundations had been hairline-fractured for years. The 2028 election was a fever dream of grand promises, the usual witch hunts, and a startling amount of common sense. It shocked everyone. When President Gutierrez emerged as a dark-horse candidate and rode a wave of populist fury to the White House, the markets finally spooked. The collapse wasn’t just sudden; it was total.

It felt as though years of slow rot were laid bare in a single, breathless moment. Unemployment had been ticking upward for a decade, but once the stock market cratered—driven by a mass exodus of assets anticipating the new tax codes—the floor dropped out. Then the housing market followed. When those second and third “investment” homes flooded the listings, prices didn’t just dip; they plummeted.

It was terrifying for a time. The knock-on effects were a landslide: 401ks vanished overnight, and the national debt moved from “massive” to “catastrophic.”

Yet, Gutierrez was savvy. She harnessed the crisis to fuel her narrative. Her solution was unapologetically simple: tax the gains where the gains were made. The argument wrote itself. While the rest of the country spiraled, the architects of the new age—AI, fusion, robotics—were living lives of staggering opulence.

“The return to the middle class,” she famously promised, “requires reclaiming the future from the billionaire elite.”

For all the initial panic, the outcome felt strangely inevitable. Democracy occasionally manages to find the right leader for the right crisis. The resulting policies were a tidal wave, a New Deal for a world that had forgotten how to build. People are still hurting, of course, but unemployment has finally stabilized at a “reasonable” 25 percent. The talk in D.C. now isn’t about finding jobs, but about what happens when work itself becomes obsolete.

We even have our first trillionaires. Surprisingly, it wasn’t Musk. It was a researcher in the Philippines who actually cracked the code on cancer—for real this time. He had a brilliant patent attorney, and now he’s quite literally orbiting the Earth, eyeing a plot on the moon.

That’s where the ultra-wealthy congregate these days. They claim Earth is becoming “uninhabitable” due to the new regulatory climate, so they’re building a paradise among the stars. Some are even tinkering with engines that approach light speed. It’s the kind of high-tech fantasy we used to watch in theaters. I wonder if my children will be resentful that we didn’t buy a “pre-construction” lot on Mars. Jury’s still out; I still prefer a lake view with only one sun.

Still, we aren’t out of the woods. How does a society cope when life expectancy jumps from 80 to 800? The clergy are in a state of perpetual existential crisis. It would be funny if it weren’t so heavy.

But the automation... you wouldn’t recognize the world. The DMV is now a bastion of efficiency—you can get a license in five minutes, provided you still want to drive. Most don’t. It’s mostly just hobbyists on tracks now. Everything else is handled by the machines.

And the cost of living? Dirt cheap. Food is nearly free, provided it isn’t prepared by a Michelin chef—they’ve inexplicably decided to charge more for the “human touch.” Everything is clean, too. Little bots sweep the sidewalks, and the air is silent because everything is electrified. Remember the copper shortage scares? Graphene wire solved that. And fusion? We have miniature suns powering Chicago, New York, and Boston.

Traditional business is, in many ways, a relic. The Gutierrez administration gutted the regulations that protected incumbents, and when the new tools made building nearly free, the old fortunes evaporated. The new tax brackets finished the job. They’re aiming for a “Great Society” 2.0: new education, new ways to reward human contribution, a new... well, I won’t spoil the surprises.

The shift is revolutionary, like moving from the plow to the steam engine, but at the speed of a fiber-optic signal.

A bit of advice: don’t bother with real estate in the city unless you just love the noise. When location no longer dictates your livelihood, the skyscrapers start to look like very expensive fossils. As for stocks? Buy in—though knowing what I know, the SEC might have some questions about the timing.

The future is already here, and it’s unrecognizable. But you have a seat at the table—use it. Stay curious, stay a little loose with your plans. It’s a roller coaster, past me—don’t forget to enjoy the drop.

“Hey! I wanted to float something past you. If you know of any copywriting positions open or opening up shortly, would you mind passing them along? The agency [redacted name] was with just laid off more than half of the staff. Seems like agency work is not as good as it used to be.”

I sighed. My response was one of resignation. I told this friend, “Yeah, of course, but to be honest, I was thinking how agencies will soon be automated entirely anyway... The LLMs can do copywriting and editing really, really well.”

“I hate that but also love that for you lol,” came her response.

These particular friends—whose names I won’t be using—have two kids and live in the greater Midwest. They do the right things. They work hard. They sacrifice. They’re active in their church. And exactly none of that matters to the agency that laid one of them off this week, nor the customers who decided they don’t need that agency anymore.

All over America, tales are emerging about AI replacing jobs. It’s starting slowly, but the value is clear. Customers can do entire categories of work better, faster, and easier themselves than they could before. If they once paid someone for this work—take website maintenance or building, for example—that person or group is no longer necessary.

Want the numbers? In Chicagoland alone, admin and information-heavy support roles saw a contraction of roughly 52,384 workers from 2020 to 2025. Entry-level roles are especially impacted; 66 percent of enterprises reported reducing entry-level hiring due to AI integration. This has led to a 13 percent relative decline in youth hires.

Imagine you’re a young person graduating from college with debt. How would you cope with that debt and no job prospects? And yes, this too affects minorities more. Black and Hispanic borrowers have $53,430 in student loan debt, on average, compared with $46,140 for white adults. Black graduates often find themselves in administrative and public-sector roles—the very ones currently being automated.

Or take the story of a newspaper editor in a small town of 20,000 people. He was new to the paper and wasn’t tenured. The company wasn’t doing well, so in a “last-in, first-out” move, they automated his $55,000-a-year copyediting job. He’s currently driving buses until he can find something better, “or more likely—move to a major city.”

These are obvious tasks for a large language model. In fact, it’s a bit surprising they weren’t automated even sooner. Still, they’re jobs that people are doing and were trained for. They required a college degree—usually, as in the case of my friend, from a liberal arts school that itself is failing.

Of course, there’s more. There’s the guy who disagreed with his boss after the boss used an AI agent to come up with a database solution. The disagreement escalated, and two weeks later, he was out of a job.

“The DBSCAN thing was a strange situation in which basically my boss wanted to believe AI more than he wanted to believe me,” he said. “But in that case, what am I being hired for, if not for my knowledge and experience?”

The truth is that the pain looks largely unavoidable. Three massive systems are converging: AI/automation impacting the job market; the Iran War crushing energy markets; and housing beginning to depreciate as an asset while many are trying—and failing—to sell. The larger, systemic trend is inequality, something that has been going on for years.

What to do? In the near term, batten down the hatches. In the long term, get ready for profound changes to life, society, work, and politics.