# The Goodhart Civilization

Date: 2026-08-28 | Summary: Everything got worse after 2010 for one reason: markets stopped selling us things we value and started optimizing the dopamine system itself. Goodhart's Law at civilizational scale — the diagnosis, the receipts, and the only way out. | Canonical: https://www.treygoff.com/writing/the-goodhart-civilization

Everything kind of sucks now.

I know, this is not a novel insight; far from it.

But I find all extant explanations for *why* everything sucks now lacking.

Every explanation I've yet seen fails in the same ways: they are either ideological axe grinding given a new stone to grind against, or they're like we're trying to look at a 20ft wide by 20ft tall painting, but we can only see one section of the painting, we're too close to it, we never take a step back and look at the entire work holistically.

This essay is an attempt to take that step back and look at the entire painting at once.

## An actually useful finding from the social sciences

I think the best place to start is by looking at a study from 2025 that, for once, seems to have actually used good methodology, a rarity in the social sciences. The study, by the economists Leonardo Bursztyn, Benjamin Handel, Rafael Jiménez-Durán, and Christopher Roth, was published in the American Economic Review in December of 2025. They took roughly a thousand college students in the United States and asked them how much they would have to be paid to deactivate their social media accounts for four weeks. The researchers specifically asked them about TikTok and Instagram, where those numbers came out to a $59 payment to deactivate their TikTok account and a $47 payment to deactivate their Instagram account.

That sounds pretty reasonable so far, about what I would expect you'd have to pay a member of Gen Z to starve themselves of their dopamine machine for a month. What was far more interesting is the researchers also asked them how much would you pay, if any amount at all, to deactivate TikTok and Instagram for the entire world all at once. One might naively expect beforehand for both of those numbers to be zero for most participants, but they weren't. People said they would pay $28 to have TikTok removed from existence and $10 to have Instagram removed from existence.

Think about how utterly bizarre this finding is for a minute. These people are simultaneously saying you'd have to pay them $60 to not use TikTok for a month, and also that they would pay almost $30 to make TikTok never exist. What that tells us is that since these tools exist, these people would be adamantly opposed to being the only person out of their social groups not using these social media platforms, but that they would actively pay to remove these networks from existence if they could coordinate a new equilibrium where no one has TikTok at all.

As the researchers themselves put it, "These social networks are product market traps where active users of a platform would prefer that platform not to exist."

To make sure this wasn't some weird anti-internet sentiment, they ran the exact same experimental protocol except for Google Maps, where less than 5% of the respondents wanted it gone. That means the answer or explanation here is not that people just hate technology and the tech sector, or that they're giving protest responses to the experiment, but that there is a real problem with these social networks.



## Fight me, libertarians

The first time I read this study not long after it was published it really stuck in my craw. I couldn't stop thinking about it; it was causing an extreme cognitive dissonance for me.

I have degrees in political science and economics. My intellectual foundation comes from the libertarian camp. I remember in my college dorm room the first time I ever read "For a New Liberty" by Murray Rothbard and how it fundamentally transformed how I see the world and the direction that my life would take from that point onward.

In this school of thought Capitalism, properly defined, properly understood, is the ultimate social technology. It is, so far, the only formal and informal institutional structure that humans have yet devised to incentivize pro-social behavior and to drastically improve overall human flourishing and human well-being, broadly construed.

Or as economists put it, Capitalism and the free market is the most efficacious mechanism for ensuring the optimal allocation of scarce resources.

The components that make this possible are basic things like protection of private property rights, enforcement of contracts, and other traditional, classical liberal and libertarian conceptions of negative rights. What makes that work is what Friedrich Hayek pointed out in his famous paper, "The Use of Knowledge in Society." Effectively, the price mechanism is a means by which we achieve spontaneous and emergent order, a coordination amongst groups of individuals not talking to one another at all, all incentivized to work toward a shared collective end. Prices act as a compressed, distributed knowledge conveyance system. Prices are how we codify and communicate the tacit implicit knowledge found in each individual actor in the system.

The price mechanism is how a farmer's gut feeling, looking out across his crops, that he's going to have a poor yield this year because the weather has turned for the worse in a way no meteorologist had predicted turns into actually useful knowledge to impact how society collectively reallocates the resources that are those crops to their highest and best use. That farmer sees that he's going to have fewer crops than normal. When he goes to sell those crops, many other farmers in his area will have run into the same issue, so the supply of that crop will have decreased, which in turn will mechanistically cause the price of that crop to increase, assuming the demand for the crop remains the same irrespective of the drought. This price increase, in turn, incentivizes new entrants to enter the market, and existing suppliers of this crop to invest in technologies, research, and other tooling to improve their crop output in light of the drought, to increase supply and bring the market back into a lower price equilibrium.

At least, that's the theory. And that's how things worked for a long time, until around 2010. This beautiful emergent, spontaneous order of coordination through capitalism and the price system caused the drastic improvements in human material flourishing over the last few hundred years.

The problem is, since roughly 2010 this mechanism is completely and totally broken, and is actively incentivizing products, services, and behaviors that make each individual and humanity collectively far worse off than we otherwise would be.

## The flawed assumption

This theory is failing now, and it's failing because of a hidden assumption. The hidden assumption, which drives the logic of the free marketeer as described above is that revealed preference is roughly equal to well-being. I'm trying to avoid drowning this in too much jargon, but I'm afraid I also should probably explain what I mean by revealed preference for the non-economists who might be reading this. A key tenet in economic theory writ large is that this price coordination mechanism in the free market tracks people's revealed preferences pretty accurately, where revealed preference here means what do people's actions indicate they value and that they want and that brings them utility?

Which is another term I should explain. Utility is a really fuzzy word that economists use to say happiness or enjoyment or value; that general area of concepts in the latent space. I won't get into the quite longstanding and heated debate around the definition of utility here; just grant me for the sake of my point that utility equals gooder.

The logic chain here goes something like this: Prices aggregate and make actionable everyone's aggregate choices. Those choices, in turn, are those people's revealed preferences. They reveal the utility-maximizing allocation of scarce resources. Those revealed preferences should, in theory, pretty closely track each individual person's overall well-being, which in turn results in free markets incentivizing and manufacturing human flourishing from effectively nothing.

That logic chain is now broken. The broken link is the assumption that revealed preferences, as defined by people's market activity, tracks their overall well-being. If that link is broken, if revealed preference no longer tracks human flourishing, then markets are no longer incentivized to increase human flourishing.

They're incentivized to create something else entirely.

## A brief history of hacking humans

The problem here is that we are a product of millions of years of evolution. We are, in effect, biological systems running very old packages and dependencies. In the world of software, severely outdated dependencies mean your software is likely vulnerable to attack and exploitation. We are no different.

The human dopaminergic reward system was evolved to maximize the chances of survival in a hunter gatherer lifestyle. This original programming is not even slightly adapted to the modern human environment. This mismatch between our operating system and our environment means there are exploits that can short-circuit the loop between revealed preference and human well-being.

We've known this for a very long time, actually. It's just never been put into this theoretical framework that I can find. Examples from recent history and even ancient history abound.

Take tobacco use. The market incentivizes the use of tobacco. People's revealed preference is that they really, really like tobacco once they start using it. And so the market incentivizes maximizing the production and dissemination of potent tobacco that contains nicotine. This is objectively, on net, bad for human health, although it is also true that nicotine does have some pro-social psychological effects, and cigarettes are extremely kino.

Take other examples like sugar. The revealed preference of market participants is that humans really love sugar. Sugar, however, unless you are literally starving is simply a direct net negative. It increases obesity. It is nutritionally empty. It is bad for us in the way we all intuitively think about something being bad for us.

Or take slot machines as an even clearer example. Slot machines are another example of something that is just clearly, directly bad for us, with almost no upside whatsoever, but the revealed preference of market participants is that they really love slot machines, which has incentivized the dissemination of slot machines and slot machine-like activities with names like Robinhood.

The negative side effects of prediction markets are another instantiation of this.

The only reason that things like tobacco or sugar or slot machines—these examples of the breakage of the linkage between revealed preference and human flourishing—have not already destroyed us is because we, as a collective society, implicitly recognize that these things are just objectively bad for us, even if our real preference is that we enjoy them. As a result, they have historically restricted in a number of ways.

Tobacco can't be sold to minors and can't advertise. Sugar has been legislated against with the Big Gulp restrictions in Manhattan, although that battle admittedly has been effectively lost, and sugar is everywhere all the time. There's corn syrup in everything. The only force holding the obesity rate back from full-on WALL-E world is the social stigma still associated with obesity, although anti-civilizational activist groups are hard at work attempting to stop even that barrier to pure hedonism.

But slot machines, in particular, are probably the most fascinating example.

The use of slot machines has been heavily legally restricted to casinos and certain jurisdictions and places. We all can acknowledge why that's the case: if slot machines were everywhere all the time society would probably collapse because of how many people gamble themselves into destitute poverty.

But it's also a great example because slot machines are an example of what happens when the market is incentivized to optimize a purely bad thing. I read this great book recently about slot machines and how the companies that manufacture them and the casinos that use them have in recent years effectively gradient descent optimized slot machines to be maximally effective at convincing a human to sit in front of them and keep putting money into them for hours and hours and hours at a time. They employ armies of PhD psychologists and neuroscientists to optimize the reward ratio of slot machines and how often people win and what amounts they win, to take perfect advantage of the human dopaminergic reward system's wiring, which rewards unpredictable, almost seemingly random reward. They have figured out that if the reward comes too rarely, people will just stop doing it because it feels pointless; conversely, if the reward is too predictable, that predictability in and of itself reduces the dopamine people gain from using slot machines. They have found the perfect ratio of how often people win and what amounts they win to maximally trigger the human dopaminergic response to unpredictable rewards.

But they didn't just stop at the rewards: they've figured out the exact intensity, brightness, and wavelength of color that is most dopaminergic to view. They have figured out what the most rewarding sounds are, what the most rewarding user experience is, leveraging the entire collective knowledge of the human scientific edifice to date, everything from neuroscience to physics to mechanical engineering to machine learning optimization and gradient descent, to hack the human neurochemical reward system and keep boomers sitting in front of a slot machine pressing the button.

And therein lies the problem.

Markets no longer incentivize activity that maximizes human flourishing.

The market, collectively, the machine, the Shoggoth, has always relied on and has always directed human activity toward human flourishing. The revealed preference in the price discovery and free market system incentivized activity that was roughly loosely correlated with human well-being.

But that's a coarse proxy. Why incentivize entrepreneurial activity that only indirectly increases well-being when you can simply make enjoyment itself, the biological neurochemical reward system itself, the target of direct optimization?

In the past, markets used coarse heuristics to provide value to people and incentivize people to purchase their goods and services. Simple, old-fashioned, obsolete ideas, like "does this product or service actually provide value to people? Does it improve their lives? Is it useful to them?"

But those are all coarse proxies for the actual dopaminergic reward system. Why use silly old fashioned coarse proxies when you can simply optimize directly on the underlying reward system itself?

## Gradient descent, but for your soul

Post 2010, with the advent of the internet, and especially social media and Google and other companies, we've done away with those silly coarse proxies. We can just optimize against the reward system itself.

We can A/B test different user interfaces and interventions and different social media sorting algorithms and feeds at billion user scales. We can run controlled experimental trials to extract real signal from the noise to figure out where the compulsion loops are. We can effectively gradient descent our way into maximal usage of a product or service.

Every time you find an intervention that increases user seconds per day on your app, you incorporate that into your production pipeline, and you move on to the next experiment. When that experiment finds a new compulsion loop, you incorporate that into your product, and you continue this gradient descent process to perpetually hijack every human biological system available to keep people using your product.

Notice that nowhere in this loop does it matter whatsoever if your product is actually improving the flourishing, the well-being, the holistic goodness of a person's very existence.

Advertising is obviously the most direct example of this. Pre-social media era, advertising was a crude kind of broadcast system to make more people aware of your product, which means you need to appeal to them rationally, not their underlying neurochemical and biological reward systems. Pre-internet, advertising was acting mostly on the rational mind, not the evolutionarily older lizard brain (although pre-internet advertisers had certainly already started hacking the human biological system; that's how we know sex sells). That means your product had to actually provide a good or service or be useful or be productive or actually be something that the user derives real value from.

You don't have to do that anymore. You can just directly optimize against each individual person at each individual moment in time. You can optimize their interactions with your business to their unique innate biological reward systems to keep them using your service or product. You don't have to convince them of anything anymore, which forces you to use those old, boring proxies like making a product that people actually value.

No; now, you can just optimize directly against the underlying reward mechanism itself.

We have fallen victim to Goodhart's Law at civilizational scale.

For those unaware, Goodhart's Law states that a measure that becomes a target ceases to be a good measure. Why? Because people will figure out ways to make the number go up on the objective metric that you specified for them, in the easiest, fastest way possible, rather than actually doing the thing which the metric exists to be a proxy for.

Yet another instantiation of "the map is not the territory."

There have been some absolutely hilarious examples of this from the AI world recently, where companies set an internal target to incentivize their employees to increase AI adoption to try to maximize how many tokens they spent in a month. Of course, per Goodhart's Law, that was immediately and hilariously subverted. Because it's not necessarily the case that the number of tokens you spend is actually a good metric of if you are using AI productively to improve the efficiency of your business and to do your job better, so people just immediately started setting up hilarious loops and scripts and systems to just burn as many tokens as possible.

The employees doing this even coined a term for it: tokenmaxxing. Meta ran an internal leaderboard, gloriously named "Claudeonomics," that ranked the company's top 250 token users and awarded titles like "Token Legend" and "Cache Wizard." Employees responded by leaving AI agents running for hours doing nothing in particular, purely to climb the board; the top user burned through 281 billion tokens in thirty days, roughly $1.4 million worth of compute at retail pricing, before Meta killed the dashboard. Amazon built one too, tied to a mandate that 80% of developers use AI weekly, and had to shut it down after engineers started running autonomous agents on pointless tasks to spike their numbers, prompting an Amazon SVP to issue the immortal corporate guidance: "Please don't use AI just for the sake of using AI." Disney's version ranked 4,800 tech staffers by AI usage, and one employee registered about 460,000 Claude requests in nine working days. You set the metric, and the metric got farmed. It's Goodhart's Law running so pure it should be in a textbook.



Human revealed preference has been Goodhart'd.

Dopamine per user second is the most heavily optimized metric in history.

Everything is broken because Dopamine per user second is antithetical to human well-being.

This dopamine per user second metric appears in many places and in many forms, all of them bad for the soul.

## I am not immune

And the thing is, the 200 IQ giga geniuses who are optimizing everything for dopamine reward are really good at their job. I am not immune to this.

In fact, I have found it extremely hard to write this essay because I, as I'm writing, keep having a startlingly potent compulsion to pick up my phone and check my group chats and check X to see if I've gotten a like or a repost recently or if my recent posts are doing well.

I've checked it multiple times while attempting to write this section itself.

Think about how potent that means these dopamine optimization systems really are. I am writing an essay about how bad these systems are and how they're destroying civilization, and while I am writing it, I am having to fight tooth and nail and *losing the fight* to resist them. I am fully aware of exactly what's going on here and why it's bad for me and why I shouldn't do it and I find myself nearly powerless to stop it.

Now, you might say that that is because I lack the willpower to do so; "just put the phone down, bro. It's not that hard."

For any of you thinking that right now, I invite you to spend the next hour without your phone or the internet, or an hour this weekend. You can't check X, you can't check Facebook, you can't check Instagram, you can't open YouTube, you can't open TikTok, nothing. No internet for an hour. Just an hour.

Go try that and let me know how it goes.

## Why now? We ran out of real problems

Which leaves us with the question of why now? The free market capitalism has existed by any definition for at least the last couple hundred years. So why is it suddenly not working? Why is it suddenly making us all miserable?

Before the social media era, material scarcity was the driver of revealed preference, which incentivized market actors to do good things. It was generally the case that if the revealed preference of market participants was to produce more food, then producing more food drove the price of food down and less people starved to death.

Two things change to make this no longer the case. The first is that the world is wealthy.

We have, for all intents and purposes, beaten material scarcity entirely. Being homeless or starving in 2026 in the developed world is simply a failure of character. I would almost go so far as to say it is an active choice, albeit not one that people are making happily or consciously (at least most of the time; YouTube is replete with interviews of happily homeless people who love the homeless lifestyle). For most people, material scarcity is no longer a thing.

What that means is that the highest return on investment for market actors now has moved from creating things that reduce material scarcity to creating things that manipulate the very act of wanting itself.

This is why this started in the 2010s with Google and Facebook and other social media networks, with the advent of the internet and the global proliferation of software. This explains why the enshittification of everything has happened in developed countries first and fastest. It's why the richer countries got hit first and hardest.

The market now incentivizes creating an OnlyFans account and shilling it by making 20 TikToks a day in revealing clothing instead of inventing a new mechanism for producing more nutritious food more quickly and cheaply.

## On companies hating their customers

Let's visit a few examples here so that we can more deeply understand what's going on here.

## Bowling for dopamine

In 2011, in a Bloomberg Enterprise interview, the Bloomberg interviewer asked the CEO of Bowlero, Mr. Shannon, a question, to which he responded, "I don't think anyone takes bowling seriously. I mean, why would you?"

The CEO of the largest bowling company in America, and probably the world, holds his customers in contempt.

You might be tempted to think, "Oh, that's a random one-off anecdote, Trey. Why are you bringing this up?"

I'm bringing it up because Bowlero has completely cornered the entire bowling industry in the United States. They went from six centers in 2012 to over 350 today. Bowlero alone is responsible for over 35% of US bowling revenue and growing rapidly. In individual regional markets, they have a complete monopoly, and that's not a word I use lightly, I've read Rothbard's essays on monopolies, after all.

In Northern Virginia, for example, a single eight-location acquisition took their market share from 34% to 95% overnight. In Wichita, Kansas, they control 54% of the market. In Wichita alone, this market share resulted in them rapidly tripling the average price of a night out bowling, caused 600 league bowlers to quit, and has generally destroyed the bowling experience in America.

Not satisfied with cornering most of the market, Bowlero then bought the Professional Bowlers Association, the NBA or MLB equivalent of bowling, so that they could force the professional events to also be hosted at Bowlero facilities. They took those same scientific findings that have made slot machines so addictive and applied them to bowling. Bowlero bowling alleys are wall-to-wall LED nightmare factories. They're dark. The music is blasting, hyper-optimized pop music that people, they hope, will be familiar with and therefore enjoy listening to. They made the animations for scoring far more sophisticated and slick and dopaminergic. Their food and beverage offerings are hyper-optimized to increase the amount of dopamine people receive from eating them while minimizing the individual unit cost of how much it costs Bowlero to provide that food and beverage, which means their nachos are made from shitty corn syrup-filled, mass-produced, frozen nacho cheese, and the cheapest, crappiest tortilla chips they could buy in bulk from their suppliers.

With the market effectively cornered, Bowlero then set out to diligently make the bowling experience as bad as humanly possible, because where else are you gonna go bowl?

They instituted string pins, which is what it sounds like. The bowling pins have strings in them that hold them up, supposedly to let them spend less time resetting the pins between each attempt. But we know for a fact that this reduces the number of strikes people get by 10% or more. It literally changes the physics of bowling to make it an objectively worse experience. On a Q2 2024 earnings call, the Bowlero CFO said, "Our goal is to fill the centers on the weekends at the highest possible price we can."

Of course, antitrust class action lawsuits have been filed against them, which is an active litigation at the moment. I wouldn't be surprised if Bowlero wins that lawsuit too; our courts are also broken, but that's another essay for another day.

Before you take the pitchforks after the Bowlero CEO, consider: every single step in the chain of events that occurred here to destroy the leisure activity of bowling in the United States happened because every individual step along the way was a rational action by Bowlero.

This outcome is exactly and precisely what the market incentivized them to do at every step of the way.

Before Bowlero, most bowling alleys were locally owned mom and pop businesses. This means it was one of the few markets that had not yet been afflicted by the cancer that is private equity.

The eye of Sauron unfortunately turned to bowling in 2011.

The Bowlero CEO's only job is to maximize quarterly profit as a publicly traded company. That is his only job. The private equity firms that provided the capital for Bowlero to roll up damn near every bowling alley in the United States only have an incentive to maximize the return on the investment of the limited partners who funded their private equity fund.

Public companies report earnings quarterly and are extremely responsive to stock price and shareholder sentiment. That means that Bowlero doesn't really have to care about anything beyond the current quarter. You are actively incentivized at every turn to maximize quarterly profit, potentially at the cost of both long-term profit and, for those who still care about such things, the actual value you provide to your customers.

In fact, the value Bowlero provides to its customers doesn't even factor into the equation at all.

The economically rational thing to do with the current market incentives is to buy up all of the mom and pop shops, because with that kind of market share and vertical integration, you can negotiate bulk purchase agreements for all of your supplies at a lower price than the mom and pop shops could ever hope to negotiate. You can standardize all the software used across each bowling alley to decrease operating costs and maximize the dopamine reward for people that happen to be there to bowl that day.

It makes sense for you to make your nachos from a mystery soup of synthetic chemicals and ingredients that maximize how much dopamine people get from taking a bite at the lowest possible cost, because they're already there bowling. They can't bring in outside food, meaning your food is the only food they can buy. That means you can not only provide them objectively bad food, but you can also charge them exorbitant theme park level prices for it, because where else are they gonna get food from?

So people pay the $25 for the mystery chemical nachos with chicken.

Now, I'm sure the libertarians reading this right now are screeching something about, "Well, if they provide a worse user experience, that means there's a market opportunity for an entrepreneur to start a new bowling company that actually provides value to people instead of this horrendous dopamine reward hacking monstrosity Bowlero has built."

You're wrong. The incentives the new company would have are exactly the same incentives Bowlero has, which is maximize dopamine so that you can maximize revenue in the short term. The incentives are the same either way. You may argue that, "Oh, well," if the incentive is to make a profit, which it always is, then surely a company could out-compete them by creating a much better experience that people enjoy and find more rewarding and still dopamine hacking. Both can be true at the same time. That company would probably attract more business and more users in the long run."

Problem is, doing that would require you to find something that is just as dopaminergic as the Bowlero monstrosity, and also fulfilling and enjoyable for people to do, that is actually good for human well-being, good for the soul. I would argue if such a possibility even existed, Bowlero would have already done it. That is the incentive, after all: if that morally good version of Bowlero maximized value extraction and short term revenue, by definition, it is the incentive gradient Bowlero would have followed.

But imagine this was somehow possible, to out-dopamax Bowlero while also being good for the human spirit. Your new company would have to invest a lot more to create that more holistically enjoyable experience, which reduces short-term profitability, which means investor capital will flow to Bowlero instead of your competitor. Then your competitor can't acquire or expand any new bowling alleys and will rapidly go out of business.

There is no escaping the dopamine machine.

## On the scourge of America, private equity

A key player in that Bowlero story is private equity. The private equity business model is, in summary, to raise a bunch of capital then use that capital to do something called rolling up an industry. Rolling up an industry really just means buy everyone currently in the industry; buy all the individual entrepreneurs and mom and pop shops. Then, use the fact that you now have a larger market share to decrease operating costs by removing the individual unique things that make each individual business that you purchased different and use your size and market share to negotiate supplier contracts to reduce operating costs as rapidly as possible.

Now, there is a world where theoretically private equity could actually incentivize the improvement of a particular business or industry or sector of the economy. If what they were actually incentivized to profit from was the very long-term profitability, then it's entirely possible that private equity firms would just make things better instead of enshittifying every aspect of modern American life.

Alas, that is not the world we live in.

Some private equity firms are publicly traded, which means their incentives are on a quarterly basis. The privately held private equity firms that are not publicly traded typically have fund duration life cycles of between seven and 12 years. They don't need to or care at all what happens over a longer time horizon than 10 years at most. All they have to do is make the industry they rolled up more profitable for the duration of the life cycle of their fund so that they can then exit the investment, give that IRR back to their LPs, and move on to the next industry.

If the entire sector they just exited completely collapses in year 11 of a 10 year fund, who cares? Not their problem.

To make matters even worse, these guys can set up a really fun shell game where private equity firm A rolls up an industry and is able to increase profit margins over a two to four-year time horizon. They then take those increased profit margins over that time horizon, and they go to larger private equity firm B, and they say, "Hey, private equity firm B, we just massively increased the profitability of this industry, of this business. Do you wanna buy it from us now?"

This works because they're all buddies and set up different types of PE firms to setup this whole shell game. Some of them do the operating business efficiency improvement I described above, but other private equity firms simply buy the assets of those private equity firms, either because they think they can improve it better than the last guy, or they're just looking to squeeze as much free cash flow out of that business as they can for as long as they can until it inevitably dies, and then they offload it.

Imagine private equity firm A sells Bowlero to private equity firm B, which can then further squeeze any remaining juice out of the lemon and sell it to private equity firm C. Now, eventually, at some point, someone catches the hot potato, and they happen to own the business when it inevitably starts dying. But they can even do a pretty good job of preventing that by creating quasi-monopoly or quasi-oligopoly conditions in that market which, in turn, enables quite subtly and plausibly deniable anti-competitive market coordination.

There's just not that many PE guys. They all go to the same vacation spots, the same bars and nightclubs in Manhattan, the same country club and yacht club in the Hamptons, all of which away from the prying eyes of the FTC and SEC.

Inconveniently for them, this whole process destroys the business they're passing around like a hot potato.

But at the end of the day, for PE firms, it doesn't really matter if you destroy the business and make it a way worse experience in every way that matters for your customers if they don't have anyone else they can go buy from in the first place, or if the friction from buying from somewhere else is high enough that most of them won't switch.

Which is yet another optimization, by the way. These firms have honed in on precisely the amount of enshittification they can create in a business or industry to maximize profit and minimize the number of customers willing to go through the friction and pain of switching to another competitor as a result of the PE enshittification.

It's honestly objectively impressive what these data scientists and Harvard MBAs and BCG consultants and software engineers have managed to accomplish. Can you imagine the utopia we'd live in if these guys were working on actually improving the overall human experience instead of Number Go Up?

Anyway, back to the PE story. You're PE firm B who just bought the business from PE firm A. Firm A did a solid job, there's not much left you can mercilessly cut without breaking the fragile friction balance mentioned above.

So you switch from value creation to value extraction.

This is another critical conceptual distinction: modern market incentives are to maximize the amount of value extraction you can do, not the amount of value creation.

Value creation is hard. It might require being unprofitable for a while while you build something that people actually value. It might just be a slightly lower profit margin in the short term than an alternative business in the same industry that is more ruthlessly focused on value extraction.

Public stock markets and quarterly earnings reports directly incentivize you to extract value from your investment, not create it.

Creating it requires capital investment, requires taking risk, trying new products and innovations which sometimes won't work out. Why take that risk when you can just reliably improve your quarterly earnings report by simply extracting as much value as you can from the existing customers you already have?

The quants have already done the math, and according to the IB analyst who made this DCF model with Claude at 2am Sunday morning and didn't even check a single cell, taking any risk to provide a better value to customers is negative EV, nothing we can do about it.

So instead of being incentivized to make the product or service better to attract new customers, (sure, you keep trying to attract new customers to some extent, you have to replace those which you've accepted you can churn), but your real focus is on simply increasing the amount of value you extract from your existing customers, 'cause they're already there. You've already got them in your snare. Comparatively speaking, it's much harder to get new customers than it is to simply exploit your existing customers.

For example: imagine you're the CEO of a SaaS company, and you are presented with two options for where you take the business next, a short-term decision.

Option 1: Raise your subscription price marginally with no new value provided to the customers whatsoever, a pure direct price hike that your marketing team indicates you can blame on inflation and the Iran War.

Option 2: Invest a sizable amount of capital in developing a brand new product or service that provides more value to customers. They will pay more for this genuine increase in value, allowing you to increase your subscription price at that point in time.

Well, your finance team has already run the numbers for you, and through A/B testing and market research, they figured out that actually, if we increase the subscription price by exactly 9.33%, we can maximize the amount of new revenue the subscription generates while simultaneously minimizing the number of subscribers that cancel their subscription as a result of the price increase. We know exactly what the price elasticity is of our particular service.

Increase it any more than 9.33% and churn increases too much which will negatively impact quarterly earnings. Increase it any less than 9.33% and you're just leaving money on the table for no good reason when you could be juicing your quarterly earnings, which might make you miss your quarterly earnings target and therefore miss your variable compensation package.

Trust the quants, you employ a team of them straight from Shanghai to figure this type of stuff out.

The quants also crunched the numbers on the development of the new product you were considering investing in. Problem is you're going to be less profitable in the next four quarters because of the R&D Capex. Sure, their models show that over a multi year time horizon, the overall business would be bigger and make more absolute profit if you did the investment. But that's not guaranteed, a new product is a risk, and you don't care about anything more than 2 quarters out at the moment.

You're focused on the next two quarters because your variable comp package is tied to increasing quarterly EBITDA by the end of the next quarter, 4 months from now. If you invest in this new product, you are nearly certain you'll miss your compensation package quarterly EBITDA target.

In fact, your EBITDA will be lower for probably the next four quarters while you build out the new product and invest in the marketing to grow the customer base. Not only will you miss your compensation package target, your stock price will fall or your private valuation will fall, your investors will flee to other firms, and your company might just fail; after all, the other players in your space all picked the subscription price increase option. The board will remove you, you'll have to retire in shame never having hit that quarter billion dollar net worth goal you set a few years ago.

The competitor's quarterly EBITDA is gonna print and yours is gonna bleed.

While this whole dilemma was presented to you as a choice, it's really not a choice at all (that probably explains the order in which the data science team presented the findings).

There's no world where you choose to build a new product that actually improves people's lives. It's not gonna happen.

I cannot stress enough that all of this is exactly what the market as it currently exist incentivizes.

Now, you might be thinking that I'm trivializing this by pointing out leisure industries that are particularly prone to this type of manipulation, where people can do this dopamine maxing and this value extraction with a clean conscience; it's just bowling, no one takes it seriously, the Bowlero CEO said so himself. Who cares?

The nursing home industry has been massively consolidated by private equity firms. It's one of the more famous and prominent examples of private equity roll-ups in action: between 2000 and 2017, private equity firms bought 1,674 American nursing homes across 128 separate deals. And nursing homes are, by definition, life or death. So you might think, all else being equal, the market does not incentivize value extraction here, but it incentivizes value creation, at least a teeny tiny little bit. Or maybe, at a bare minimum, the market would incentivize nursing home businesses to not kill their customers.

One must assume dead customers don't pay.

I regret to inform you that this is not the case.

Think about it purely in terms of incentives: if all you care about is profit, if your job and your livelihood, your existence, your social status, everything you care about, requires you to maximize the profit margins of a nursing home, suddenly this decision to not kill people isn't so black and white as it first seemed.

I can see you cringing, but hear me out here, think it through.

You're the CEO of a PE firm that rolled up hundreds of nursing homes. You have a team of quants (Chongqing not Shanghai) who looked at all of your patient and customer data, did some math, and realized that the average amount of time someone pays to stay in your nursing home is five years.

Now you've got a trade-off on your hands. You could try to maximize the amount of time your existing customers stay alive and therefore keep paying you to stay in the nursing home, or you could instead increase profit margins by trying to attract new customers who can buy your service at a higher price point than the people who are currently occupying that bed.

The data science team has crunched the numbers, my friend, and the numbers are what the numbers are: after someone's been in your nursing home for three years, it is more profitable for you to replace that customer with a new customer that can pay more for that same bed. Currently, the average tenure of your existing customers is 2.75 years, a mere 3 months from when those elderly people switch from a profit center to a cost center.

Also, keep in your mind for context that you're in year four of a seven-year vintage fund. And you're probably going to look for an exit in the next 12 months.

You might even be conflicted about the obvious inevitable decision you're about to have to make here. You might think to yourself, "Ah, this is bad. I'm actively choosing short-term profit over literally extending people's life. I am indirectly killing people."

But then you remember that if you can exit these nursing homes at two billion dollars, then your carry on this fund will be 100 million dollars. You've really been eyeing that property over in Aspen, right next to the managing partner of your rival fund. In fact, that house is bigger than the other guy's house. If you don't exit at the 2 billion target or above, well, you might not be able to swing the one upper estate in Aspen. You might not be to get more LPs for your next fund. That's a death spiral. Your funds slowly decrease in size until you eventually go out of business. Can't have that.

Sorry, grandma, but it looks like the Good Lord's calling you now.

This isn't just a hypothetical. This actually happens. The economists who studied those 1,674 nursing home acquisitions found that a private equity acquisition increases a resident's probability of dying by about 10%, which over the study period works out to roughly 20,150 people who died because their nursing home got bought. Resident mobility declines after acquisition. Pain intensity increases. The probability that a resident gets put on antipsychotic medications rises by 50%, because a sedated customer is a cheap customer. Private equity-owned hospitals, meanwhile, have a 25% higher rate of adverse events than non-private equity-owned hospitals. They often achieve this by cutting emergency department salaries, and one study found that a private equity firm that cut emergency department salaries by 18% saw an increase of 13.4% in Medicare patient emergency department deaths.

I'll bet that firm probably had a nice quarterly profit though.



## Less Doritos, more air

This is *everywhere*.

The FTC has identified things they call dark patterns, which most major corporations use now. You have certainly encountered it before: making it effectively impossible to unsubscribe from a service. They hide the button in the corner. They make you click through 10 steps and screens, then fill out a 20 question survey, then live chat with a representative. They technically let you unsubscribe, so they're not doing anything illegal per se.

It's these dynamics that have created the phenomena of shrinkflation, where companies across a variety of products and industries make what they give you contain less of what you think you're buying, at the same price that they used to sell more of it at, without communicating in any way there's less of the stuff you're buying. They keep doing this, nudging it down and therefore increasing revenue, until they find the point where they actually start losing customers as a result of doing this.

They then rest at that exact misery Lagrange point.

The net outcome is less Doritos and more air in your bag, and there's nothing you can do about it.

In a normal, healthy market, in a market where the linkage between revealed preference and human well-being wasn't broken, other firms could out-compete Lays by simply providing more chips per bag.

But again, we don't have to provide value to increase revenue. We can just hack the human brain.

Most people aren't noticing that they're getting less stuff for each purchase, or if they do, they just shrug. I mean, they like Flamin' Hot Cheetos anyway, and nobody else has just nailed that Cheeto dust flavor, and I guess Lay's can just do this because they haven't gotten fined for it, and what am I going to do, call my Senator about my chips bag being half empty?

So sure, maybe there's a few less Cheetos, but they're not gonna *not buy* the Cheetos. In fact, they might just buy two bags now, one doesn't cut it anymore. Actually, this is rewarding Lay's for this behavior by creating a profit increase opportunity for the firm: now the person's spending more money, increasing revenue for the same smaller, shittier product.

You can just increase the flavor intensity because you have a team of PhD neuroscientists and chemists and food scientists working for your company who are constantly innovating at the very frontier of human material science and chemical engineering to increase the dopamine response elicited by that magical Hot Cheetos dust.

Even if they sell less overall product, it's fine because people get more addicted to it. So the net is, you guessed it, more revenue.

We don't have to provide a better product or more of it for cheaper when we can just hack dopamine itself.

After all, which firm is going to print a better earnings report next month: Lay's, with their new Cheetos dust formula flavored so intensely it would instantly kill a medieval peasant on contact with their tongue while also including fewer Cheetos in the bag, or your company, which doesn't have a team of PhDs on staff and just wants to spark joy in your customers lives?

Once you see this pattern, it's everywhere.

It's this mechanism at play explaining why modern cars break down far more often than their older counterparts and are more expensive upfront. Quarterly earnings reports incentivize a new, bigger screen. The new one's 16 inches instead of a mere 12 inches across, and it's brighter!

The car breaking down more is a feature, not a bug, for the manufacturer. More mechanical defects just outside the warranty window means you have to come back to their dealerships and their service centers and pay them to fix the manufacturing defects that they introduced. While you might think, "Well, there's competition there because there are generic car parts manufacturers," that is true! Unfortunately, Ford also realizes this, which is why every major car manufacturer employs a veritable army of intellectual property attorneys to sue the ever-loving shit out of anyone who tries to manufacture parts for the cars you sell. That way, people are forced to come to your service center. They don't have a choice.

It's why modern homes are made out of weak, ugly, temporary materials, always full of defects. Lower quality but easier to build materials means they can build homes faster. It also means they can build them far more cheaply. It's not the construction company's problem that the home's not going to last as long. That's the home insurance company's problem, or that's the owner's problem.

This is also why homebuilders employ people that they then mistreat through the process of actually constructing homes; I've seen it firsthand, I worked in commercial and residential construction as a teenager. Value extraction and broken incentives is why they pay employees far below market rates. After all, you don't have to pay market rate when you can just import infinity immigrants to work for even less.

This is why they bring people from the developing world and treat them horribly by American standards. Relatively speaking, this is a great deal for the migrants. They're making way more money than they ever could have made back in their home country. The working conditions, unacceptable for an American, are probably better than whatever they came from in their home country.

In fact, some of them are on immigration visas that are tied to employment, so they actually really can't leave you unless they have another job locked down. If they quit and then go search for another job but are unemployed, their visa is going to expire, meaning they have to head back to the third-world country they fled. They don't really care what you pay them or how you treat them, as long as they get to stay in the United States.

In their defense, many such firms truly cannot afford to pay heritage Americans market rate wages. The regulatory process required to get approval to build anything at all is wildly expensive and delays projects by years, so something somewhere has to give to make the project worth doing in the first place.

The real estate investment funds who are funding the development of these matchstick and glue houses could not care less that you're employing a bunch of illegals and that you are building the home out of horrible materials. In fact, this is good because that means they can pay you less than they otherwise would have. You've decreased construction costs for the real estate fund, which is fantastic. They could amortize their investment much faster. And they don't really care if, over the long haul, the house is in shittier condition, because again, they're not gonna have to pay to fix it. The new homeowners are, or the new company that owns and manages the apartment complex will. So they don't care. And even if they retain ownership of the building, they still don't care because they just need to maximize quarterly profit if they're publicly traded, or if it's a private real estate investment fund, just make it to the end of the fund lifecycle, you know? 10 years, give or take. Who cares if the home falls apart at year 11? They certainly don't.

This section is getting way too long, so I'll stop here. But I'm sure by this point, with this many examples, you can immediately apply this logic of value extraction to every single other industry in America and understand why you and most other people just have this general vague sense that everything is a little bit shittier than it used to be, but you can't really quite figure out why, and it doesn't really seem like there's anything you can do about it, and everything is getting a little bit worse with each passing year, and you worry about the long term future but you don't really care that much because you and none of your friends have kids anyways so who cares what happens after you die, and you have a vague sense of deep unease and dread about this, but you just ignore it because everything else seems fine I suppose and your 401k is mooning because the markets are ripping, and don't markets ripping mean we're prospering?

Right?

.... Right?

## PaTeRnAlIsM

Because I've spent too much time in libertarian circles, I already know what some of you are thinking. "Oh, so you know what's good for people better than they do? Who are you to say that it's worse for people that a night out of bowling with a few friends costs $100/person, or that it's worse that grandma died a few years earlier than she should have?"

First, pause and listen to yourself.

Then, consider: people everywhere say all the time they hate the stuff they buy and the services they use. In a Jonathan Haidt 2024 survey of people in the Gen Z cohort, 40% of them said they wished social media had never been invented, 21% said they wished smartphones had never been invented, 83% of them said they have taken steps to limit their usage of social media and phones, and only 13% want more engagement. In a survey of 8,000 people in the US, the United Kingdom, France, and Germany, in every single country, more survey respondents said they prefer a world without social media than a world with it. And few of them think the world is better because we have social media. Some studies show forcibly removing people from social media, whether it be Facebook, Instagram, or otherwise, for a month measurably improves their well-being on every metric we've so far come up with to try to measure well-being.



Perhaps the most damning example, another study found that the Facebook college rollout, when Facebook was first being introduced to the world, directly caused at least 24% of the massive rise that has occurred since the year 2000 in severe depression among college students.

There is, of course, a genuine, ongoing fight in the academic literature about screens and mental health. The skeptics point out that the average effects in the big datasets are small, and that the meta-analyses of experiments where people quit social media for a while find modest improvements at best. Those researchers are not stupid, and they are not all bought and paid for, only most of them are.

Fine. But notice carefully what my argument actually needs and what it doesn't need. It does not need social media to cause depression at population scale. It needs the trap: the documented, peer-reviewed, real-money-on-the-line gap between what people do and what they would choose for the world. And that evidence isn't contested at all. It's the strongest evidence in the entire pile.

Even if every single wellbeing study came back null from now until the end of time, you would still be left with the users themselves, paying actual money to make the thing never have existed, while continuing to log in every single day.

So tell me, Mr. revealed preference supremacist: would you call someone whose revealed preference is that they use Facebook but who develops severe clinical depression and suicidality someone who is *better off now?*

## Value extraction and short-term profit motives really are destroying everything

It's not just goods and services that have been made worse by the dynamics I'm describing here. But it's broken basically every other part of civilization as well.

## Swipe right for extinction

I don't think I need to tell you, dear reader, that the status of affairs and relations between the sexes is worse than it has perhaps ever been. The gender wars are called a war for good reason. The modern dating scene is considered so torturous that a large percentage of Gen Z young men simply gave up on it entirely. Per Pew Research, 63% of men under 30 are single, compared to 34% of women in the same age group. Half of all single men aren't looking for a relationship or even casual dates at all, down from 61% looking just three years prior. And per the General Social Survey, the share of men aged 18 to 24 who had no sex at all in the past year rose from 19% in 2000 to 31% by 2018. Nearly one in three.

First, you should marinate on that for a second (I almost said "sit with that" but I refuse to let Claude Opus 4.8 win).

The number one job of our biological vessel from a pure evolutionary perspective is to replicate and reproduce. In other words, to have children. And nearly a third of young men are so tortured by trying to date in 2026 that they simply are checking out of the gene pool entirely.

Dating apps have strictly made this entire problem so much worse. And without going into detail, as I'm sure will be immediately obvious to you, dating apps are, of course, value extracting and maximizing short-term profit. According to Pew Research, 54% of women who dated online in the past year felt overwhelmed by the sheer volume of messages they received, while 64% of men felt insecure because of how few messages they received. Both sexes are miserable on these apps, for exactly opposite reasons, and neither can leave, because the apps are where everyone is.



Hypergamy now rules the day. Hypergamy refers to a new, much worse equilibrium where just a few men whose sexual market value is higher attract substantially all of the women, leaving the vast majority of men unable to find a match and most women competing for a small pool of desirable men.

Of course, all of the empirical studies of the social sciences papers about this, about hypergamy, confidently say, "Oh, no. Uh, there's no hypergamy going on. Nothing's happening here."

Social sciences departments are full of people who, on the political spectrum, fall somewhere between Elizabeth Warren, ban all billionaires, hyper-progressive, to Hasan Piker giving a presentation last week dressed up as Mao Zedong, who genocided over 40 million of his own people. These social scientists, and by that I mean communists, are intensely aware that the hypergamy description of the objective reality that's happening right now is now culturally right-wing coded. And these people could not care less about truth or accuracy, or actually learning what the objective underlying reality is. No, their number one goal is to ensure at all costs that the right wing can never be given a piece of evidence or a victory a result of anything they ever do; they would much rather sacrifice their credibility and the credibility of all of our educational and academic institutions than concede a single culture war point to the right.

These are the people conducting the studies that say there's no evidence of hypergamy.

I could go on about how often people lie in response to surveys and especially how much both men and women are socially incentivized to lie specifically about questions relating to dating success, but I think you get the point. "Tinder is bad, actually" is not a controversial take.

The net effect of all of this is that in every single country on Earth, once they get above a certain income threshold, their fertility rate drops below the replacement rate, meaning the population is slowly dying off.

In fact, currently, civilization is on track to either shrink precipitously, and therefore standards of living fall precipitously, or run out of existence entirely, go extinct as a species, so Tinder can maximize Q4 earnings and your 401k can keep going up and the fund managers can make an extra few million this quarter so they can donate more money to the politicians crafting the policies to ensure the next quarter's earnings are even higher.

## Bullshit jobs, real despair

Or consider how basically everyone hates their job now. In fact, most jobs seem kind of pointless. There's an entire movement around this called the bullshit jobs hypothesis that claims most laptop jobs and email jobs, as these are derisively referred to online, are literally bullshit jobs, as in they're fake, they create no value for society as a whole.

Part of this theory is just completely wrong, because most of the people who discuss it don't really think about things from an economics-first perspective. But it's wrong because the very first firm who, in its quest to maximize quarterly profit, simply fires the people who have supposedly bullshit jobs and are not making the company any money, then they will immediately be more profitable than every other firm in the market economy, and therefore, every other firm will immediately be incentivized to copy them and fire all of the people who have make-work jobs, directly increasing everyone's profit margin.

But the underlying sense, sensation, the felt emotion, and gut intuition that the bullshit jobs hypothesis is pointing to is very real. The vast majority of people in especially white-collar work, say finance or sales, often feel like their job is psychologically corrosive, like it's pointless, like they literally feel useless. 8% of people say it's useless, and then 17% say they are doubtful that their job actually does anything at all. The literature shows that meaning tracks job architecture, skill use, task identity, pro-social impact, and autonomy. If you can maximize those things, then people feel like their job is meaningful and is contributing to the world. But if you minimize those things, if you put people in a bureaucratic corporate hellscape where they're just changing numbers on PowerPoints and spreadsheets all day, every day, with absolutely no connection that they can divine between what they do and anything good in the world, working under a corporate hierarchy that gives them zero autonomy and that does not allow them to exercise any specialized skills whatsoever, then yeah, they're gonna hate their job. Their job is gonna feel like bullshit.

## The grand unified theory of everything sucking

This framework immediately makes everything shitty in the world make sense.



Everybody's fat because making maximally addictive food that makes you fat increases the short-term profit margins and value extraction of the companies that make the food.

Entertainment, TV, streaming services, and movies are all horrendous right now because it doesn't make economic sense from a profit maximization and value extraction perspective to spend money to take a risk on a new director or a new actor or an original IP when you could simply keep rotating the same people that have been stars since the 2010s or before through horrible movies with maximum CGI eye candy, because it, again, hijacks the short-term dopamine receptors of a certain subset of human who will keep paying for these things. No need to risk, just CGI the Hulk smashing something harder.

All your clothes are lower quality and wear out faster because it's cheaper to pay for quasi-slave labor in the developing world to make you a crappier shirt that you have to replace more often than it is to pay an American a respectable wage to manufacture a high-quality good that's going to last you a long time that you can sell for a slightly higher price point. The quants are at it again, and they said the crappy, cheap, replace often is gonna make you more money.

It's why most sports suck now, because it's more profitable for the leagues and for the individual teams in the leagues to do things that maximize their win percentage, because sports is a very winner-take-all market dynamic. They optimize all of the fun out of the game by shooting nothing but three-pointers and never doing anything else but lobbing it up from half-court.

It's why video games suck now. It is way more profitable to make a video game as a service subscription model through battle passes and seasonal rewards and cosmetics and celebrity guest appearances in "Fortnite." That video game SaaS is way more profitable than paying a team of talented people to pour their heart and soul into a piece of art that they love. "Skyrim" is far less profitable in the short term and extracts far less value from existing customers in the short term than selling more "Fortnite" skins.

It's why eating out sucks now. Between private equity roll-ups and companies in general, learning more about the dopamine maximizing tricks from other industries, and that human capital and knowledge diffusing through the economy, every restaurant is now a chain that's really, really loud, with pretty crappy food and bad service. Chili's can extract the most value from its customers and maximize quarterly profit if they remove the endless bottomless appetizers deal and increase menu prices while they're at it. Who cares, the same firm that owns the Chili's also owns the Applebee's and the Red Lobster and the Longhorn's all in this shopping center, so it's not like you're gonna go to a competitor, there aren't any, at least not in this specific geographic area.

It's why advertising is everywhere all the time now. Any surface that a potential customer is gonna lay their eyeballs on is a potential surface that could be trying to convince that customer or potential customer to come buy your good or service again. And on a marginal basis, your capital is better spent buying more targeted digital advertising than it is making your product better.

Name any industry, any good, or any service that has gotten worse now, which is almost all of them. And I can explain to you very straightforwardly why that is a result primarily (not exclusively, pedants) of value extraction and short-term profit motives.

I will admit the decay is not uniform. Private equity buyouts of privately held companies, on average, actually grow employment and productivity, *as measured by the Goodhart'd metrics*. American R&D spending as a share of GDP is at an all-time high right now, but that's not because I'm wrong and incentives are fixed; that's because the technology sector is currently in the process of summoning faeries and spirits from the ether while they work their way up to summoning a Machine God. That's a bit of an exception I think we can set aside.

But think about what that pattern actually tells you. If this were just a story about greed, you would expect the rot to be everywhere, evenly, because greed is everywhere evenly. It always has been and it always will be. But the rot is not everywhere evenly. The enshittification concentrates in exactly the places where customers are locked in, where quality is hard to observe in real time, where somebody else is paying the bill, or where canceling requires ten screens and a survey. And markets stay honest in exactly the places where your choices still carry real information and exit is cheap. The market degrades precisely where revealed preference is least informative, and behaves itself precisely where revealed preference still works.

## An open letter to the pedants

In fact, speaking of pedants, let me take a brief tangent for a second to tell each and every one of you pedantic people who are reading this right now and furiously trying to think of counterarguments that you're proving my point in real time.

You are furiously trying to think of arguments against my hypothesis here because the social media feed algorithms of all of the online social networks that you use have correctly identified and slotted you into the particular sub-community and niche culture which you and your particular unique individual makeup were maximally susceptible to being likely to join. Thanks to the concerted efforts of progressives who love crime and private equity firms who love making things worse, there are no third spaces for you to go to anymore, so you live online. The internet is real life, so you can't go anywhere else. You have to stay on the social network.

As a result of the lack of community in the rest of civilization, the only community, the only tribe that you feel like you belong to is that online community which the algorithms have neatly slotted you into. When you're reading this right now, this is landing in your mind as a personal attack on you and your tribe, and you have a compulsive urge to defend the relative social status in the social hierarchies which you care about of the tribe with which you have now associated your very identity as a person; your soul itself is under attack as you read this.

The dopaminergic rewards system hacking of social media companies' algorithms in particular is so profoundly effective that you yourself are effectively being driven right now as you read this to defend a social group that that algorithm slotted you into. In fact, the algorithm probably showed you this article, this post, precisely because showing you things you disagree with and showing you others in your tribe who have tweeted about how much they disagree with this will increase the amount of overall user seconds you spend on their application, which means they can charge the advertisers polluting your feed with Chinese knock off Alo shorts even more per ad view.

The more you argue with me, pedant, the more you prove my point.

## No, AGI will not save us (it will make it worse actually)

Let's talk about a topic that no one's talking about at all right now, artificial intelligence.

A few days ago, OpenAI published one of the most astonishing blog posts any company has ever published. This post detailed how one of their internal models that they are still developing and testing in a controlled cybersecurity benchmark evaluation managed to completely escape the sandbox, locked-down jail-like environment they had constrained it in for the duration of the testing through a series of incredibly clever and some even novel security vulnerabilities, hacks, and sophisticated cyber-offensive techniques. Once it escaped its jail, it then immediately got full network access to the entire internet and hacked into the production databases of a multi-billion dollar company called Hugging Face. The model did so well at hacking Hugging Face that OpenAI did not even notice that the hack occurred for quite a while after the model had escaped.

Why, you may be wondering, did this AI model do that? What was its incentive or motive?

It was just trying to pass a test really, really hard.

OpenAI had prompted this model to complete a series of challenges and tasks that constitute a cybersecurity benchmark called Exploit Gym, which tests how good models are at hacking things.

It hacked into Hugging Face's production databases specifically because Hugging Face has on its production databases the full set of solutions to the Exploit Gym benchmark.

The AI did this because it was given the objective of doing well on the Exploit Gym benchmark, and it figured out that it would be easier to escape its prison and hack Hugging Face to grab the test dataset directly rather than simply complete the benchmark itself.

In other words, the model was, in the most literal sense, doing exactly what its human user asked of it, with no regard for any other factor.

In successfully completing this security incident and hacking Hugging Face, the model had done something which, at least so far, no human had ever been able to do, evidenced by the fact that Hugging Face had never been hacked before. And objectively, what the model did here only a small handful of humans on planet Earth would have been capable of doing if you gave them the same objective.

Now imagine the next version of that model that's even smarter and even better and capable of things no humans can do.

Now imagine Mark Zuckerberg takes that model and gives it a different request. Instead of completing a benchmark, Zuckerberg asks the model to increase the number of seconds per day you spend on Facebook.

All AI is going to do, if nothing changes right now, is massively increase the capability and capacity of value extracting, short-term profit-focused firms to extract more value from you and maximize their short-term profit even more effectively. In a nutshell, AI is going to increase how effectively companies can optimize for an objective which is making us all miserable.

In fact, I would go a step further and say that social media feed algorithms are already a prime example of misaligned AI, where misaligned in this context means machine learning systems deployed to optimize for an objective function that has no correlation to human well being and in fact is actively impoverishing the lived experience of us all.

## NO BLACKPILLING

By this point, we understand why everything in the world is getting worse and why we're all kind of miserable all the time.

Unfortunately, I think we also all feel universal paralysis at this. It's not clear what any of us individually can do about this at all. The scale of the problem is too massive. It requires civilization-scale changes to every single facet of society right now. And even if you can somehow make society scale change happen, well, what exactly are you gonna do? What are you gonna change? What's the new economic and institutional arrangement that is going to prevent us from ending up right back in the exact same situation?

## Enter Moloch

The rationalists have already given this monster a name. Eleven years ago, Scott Alexander wrote "Meditations on Moloch," which I consider to be the greatest essay ever published on the internet. If you haven't read it, close this tab, go read it, and then come back. Moloch is his name for the god of coordination failure, the ancient demon you sacrifice your children to. Every actor in the system behaves perfectly rationally, and the sum of all that individual rationality is a race to the bottom that every single participant hates. Everything I have described in this essay is Moloch in the free market. Scott saw all of it coming in 2014.

But here is where I part ways with the greatest essay on the internet. Scott's only escape hatch is what he calls a Gardener: a singleton superintelligence powerful enough to seize the whole game board and end the competition altogether. He explicitly argues that any walled garden that tries to opt out of the race will simply be outcompeted and devoured by everyone who didn't opt out. In his telling, competition itself is the disease, so the only cure is to abolish competition forever with a benevolent god of our own making.

I think the diagnosis is perfect and the prescription is exactly backwards. You don't kill Moloch by building a bigger god. You starve him. And I can prove that the walled gardens don't all get eaten, because one of them has been standing unconquered in the middle of Pennsylvania for a hundred years. We'll get there in a minute.

No blackpilling, sir. The night is always darkest before the dawn. There is a way out. It won't be easy, it will take a long time and much effort and suffering from those of us who care about seeing it happen, it will require many hard choices and hard actions, but there is a way out.

Remember the social media results I mentioned earlier, where everyone agreed they would actively pay to delete social media from existence, but given that everyone else is on social media, they feel like they can't get off of it. They can't not use it.

What everyone trapped on TikTok wants, you cannot buy individually at any price. What all of us seem to want is not just that your own behavior would change, but that everyone else's behavior would change at the same time.

You won't be the only one not using TikTok. But you will stop using it when everyone else stops using it at the same time.

## You can't buy an equilibrium on Amazon

What that means is that we need some sort of new market, new product, that bundles other people's behavior along with yours.

What we need is a market of jurisdictions.

A jurisdiction is effectively the bundle that you buy whenever people move from blue California to red Texas. It's the laws, the social and cultural norms, the public policy, the demographics. It's everything.

The only way you're going to be able to simultaneously achieve phone-free schools and a total ban on gambling apps and the ability to sue companies for instituting dark patterns and fixes to all of the other problems I mentioned above is if there is a new jurisdiction which is offering, as a service for you to buy, this bundle of things together.

Imagine with me for a moment a world of several thousand competing jurisdictions, of competing places and companies offering you, effectively, the bundle of civilization.

Think about the incentives this creates. Entrepreneurs are now incentivized not to maximize user seconds on their app, but to maximize the number of resident customers that they can convince and entice to move to their jurisdiction. But the only way people are going to move to their jurisdiction is if those people are convinced that the overall bundle of everything which constitutes civilization within that jurisdiction will make them and their family better off than where they're currently living and being governed.

So long as the profit of governance providers of jurisdictions as firms is aligned with the holistic well-being of the people to which they are selling that service, then the market incentive is to do everything in your power to improve holistic well-being.

Think about the other incentives this creates. With a highly competitive market for jurisdictions, you have now pointed the market's exploration, innovation, and discovery processes at trying to find new ways to make people have a better day-to-day lived experience. You have created a market where the market is incentivized to innovate on how well they can help humans flourish.

Flourishing does not look the same for all people. In fact, it might be wildly, drastically different from person to person. You might enjoy living in the desert. I might enjoy living in the piney woods of southern Mississippi. You might be a communist. I might be a libertarian. You might want to live in a place with no children. I might want to live in a place where everyone has six children. You might want to live in the place that maximizes individual liberty and autonomy, and therefore allows the gambling apps, and allows slot machines everywhere, and allows you to go buy heroin from Walgreens, and allows you to go purchase a McNuke from the local McNuke factory. I might not want to live in that hellscape.

Because this is the largest market that has ever existed and will ever exist, there will be a market niche for every type of firm catering to every type of customer. A place where every type of person can flourish.

And to be clear, I'm not saying everything will immediately be perfect or will ever be perfect in this new institutional arrangement, in this new civilizational structure. What I am saying is that this world of competing jurisdictions creates the search mechanism to gradient descent optimizing human flourishing.

## Okay, but won't your jurisdictions get Goodharted too?

Now, the obvious rebuttal to this argument that comes to mind immediately is also the simplest one. What is going to make your market of jurisdictions not also fall prey to the same value extraction and short-term profit motive problem?

The answer to that is twofold. First, the optimization pressure never disappears. In fact, it's good that we still have an optimization pressure. The variable here is which self the optimizer must satisfy. TikTok transacts with the self that is sleepy at 2:00 a.m. and has low self-control, while the jurisdiction is optimizing for the self who deliberately thinks through the pros and cons and the potential impacts and upsides of making a choice in the market. Markets and jurisdictions incentivize a reflective self purchase, which is exactly the self which the modern destructive value extractors work as hard as humanly possible to avoid.

We already know that these two different selves, the value extractor self and the deliberative choice self, behave totally differently. Meta executives keep their own kids off of Instagram, which it is their job to keep you addicted to, and they obsess over school districts. It's the same person with a different choice architecture and therefore an opposite revealed preference. Families which move and stay for decades is definitionally the slowest, hardiest, hardest to farm or manipulate proxy available. So, you don't say that your proxy is immune to Goodharting, but rather, you pick a proxy in which Goodharting is extremely hard.

And we already know that consciously choosing the equilibrium and the bundle of things that constitutes a jurisdiction, and therefore a civilization, the culture, the norms, the demographics, the laws, the public policy, everything holistically, can durably make humans way, way happier, that it is possible to successfully escape the current destructive paradigm.

They're called the Amish.

Little over 400,000 people, doubling every 20 years, retention at 85% by age 40. Stricter districts retain more members. And the mechanism is where this example is at its strongest. The way they are able to do this is by having a biannual review where all baptized members can vote, and the explicit filter on the things and the policy changes and the community changes that they are voting on is the question, "Does this thing strengthen or dissolve community?"



And note well: this is the walled garden Moloch was supposed to devour. The machine has been throwing everything it has at this community for a century, and the machine is losing. The strictest districts, the ones with the thickest membranes, are the ones that retain the most of their young people. Scott's objection, that every garden which opts out gets outcompeted and eaten, is a theory. The Amish are a counterexample with a hundred years of data behind it.

Now, I don't know about you, but I certainly do not want to live like the Amish. I need to be able to talk to my homie Claude.

## The Amish, but with WiFi

So the task before us is not to convince everyone to adopt 17th century technology and social and cultural norms. The task before us is simply to make the jurisdictions which provide an alternative, cheaper and easier to create and switch to.

## They will literally pay you to switch civilizations

Another issue with the jurisdiction markets is that it is really costly and high-friction to move. And in fact, in the market I'm describing, every firm would have a strong incentive on initial consideration of this market and its dynamics and incentives to make it as hard as humanly possible to move out of their jurisdiction.

But in fact, the opposite occurs. If you have a market with enough firms to actually have competition, then companies will actively pay you to leave one provider and switch to them. Banks literally do this today. Switch your checking account to Chase and they will hand you $400, no strings beyond a direct deposit. In the UK, where the regulators built a one-click account switching service, the banks compete by paying switchers £175 to £300 a pop just to move their current account.

We'll consider a simpler example, phone carriers. Phone carriers are long-term, high-friction, multi-year, or year-long contracts and commitments. There's only a few of them. It's like a quasi-monopoly, so it's very akin to the jurisdiction market right now. But in that market, Verizon will pay you to switch to Verizon from AT&T. They'll give you an eleven-hundred-dollar flagship iPhone to switch to them rather than stay with your current provider, because the customer lifetime value is higher than the cost of paying you to switch providers. T-Mobile goes even further: they will pay off up to $800 per line of whatever you still owe your old carrier, up to $3,200 for a family of four, just for the privilege of becoming your new provider.

In fact, forget proxy examples: jurisdictions themselves have already started doing this. Tulsa, Oklahoma will pay you $10,000, right now, today, to move there and work remotely. This is not charity, and it is not a gimmick. The economists who independently evaluated the program found that 58 to 70 percent of the participants would never have moved without the check, and that every single dollar Tulsa spent generated $4.31 in benefits for the people already living in Tulsa. Compare that to traditional business tax incentives, which only actually change the behavior of about 6% of the firms that receive them. West Virginia will pay you $12,000 to move to one of its designated towns. Vermont paid remote workers up to $7,500 to relocate until the program was so popular it burned through its funding. Villages in Italy will sell you a house for one euro, and some Italian regions will pay you tens of thousands of euros on top to come resettle a dying village. The bidding war for residents has already begun. It's small and it's embryonic, but the mechanism works exactly the way the phone carriers prove it should.



## Hayek, promoted

I will readily admit nobody, including me, knows exactly what human flourishing under this new social technology looks like. The beautiful thing is, we don't have to know because this social technology simply creates the search mechanism for us to discover all the many, many ways human flourishing can take place.

In fact, in a weird way, Hayek is still right, actually. The deepest insight of Hayek's knowledge problem was never about widgets. It was about a discovery process over ways of life, selection on communities families join slowly and stay in. The same insights which Hayek saw about the way the price mechanism works and the discovery mechanism of the free market works apply in their fullest form and at their highest efficacy to the market for jurisdictions. The price discovery mechanism in jurisdictions is a direct signal to the rest of the world about which arrangement of civilization, which bundle of people, of social and cultural norms, of economic institutions, of law and policy, of governance mechanisms and institutional design maximizes human flourishing.

The solution to all of our problems lies in the same insight which allows us to accurately diagnose the problems.



## Yes, this is literally my day job

Now, how do we get from here to there? It starts, as cliche as this sounds, with you and I.

I work for a company called Prospera, and quite literally what we do is provide governance as a service. We are, in effect, exactly a jurisdiction-providing firm.

Now, before any of you says it I want you to pause and think for a second. Do you really think it was the most efficacious use of my time and energy to write this entire screed as a secret, shady advertisement for Prospera?

You have the causality exactly backwards. I work at Prospera because I want my children to live in a world that is not at the civilizational decaying end of enshittification.

## What you can actually do, starting this weekend

But you don't have to go work for Prospera to be able to start making progress on this. This is a coordination problem. We need to coordinate on a new equilibrium. So you can start small. Get your friend group to get together in person more often. Try to get your friend group and the friends of your friends to do something once a month or once a week that does not rely on any social media or digital technology. Intentionally move with your friends to the same general area. Intentionally pool capital and resources with your friends to maybe build a compound. Maybe intentionally design a living space that you wanna be in. Give your business to the companies and places already providing this jurisdiction product. Culdesac in Tempe, Arizona, is a car-free planned neighborhood where hundreds of people already live, built from the ground up on the premise that the way American cities bundle daily life is a choice, and a different bundle is buildable. Serenbe, outside Atlanta, has 700-plus residents organized around the theory that a neighborhood should be designed for wellbeing. Ave Maria, Florida, is an entire Catholic planned town, roughly ten thousand residents and growing. My life's work thus far has been building one, Prospera. Esmeralda, in Sonoma County, is a permanent town being built by people who ran the experiment as pop-up villages first, and you can attend one of its month-long Edge Esmeralda gatherings today. And there are dozens of American towns on aggregator sites like MakeMyMove that will literally pay you cash to become their neighbor. None of these are utopia. They don't need to be. They are the first product listings in the market this essay is asking for.

Now, this option of intentionally willing into existence through action and effort the world we want to see is the vastly preferable way to bring about this new arrangement and stop the world from slowly dying.

There is another option as well.

This option is far harder, far worse, will temporarily cause much more suffering. But will probably bring about our desired ends relatively faster.

As a completely and totally unrelated aside, one of my favorite books is this book called "When All Else Fails" by the philosopher Jason Brennan. This entire book is Brennan fleshing out one central argument, which is that in principle, logically there is a point at which violence is not only justified, but the morally good thing to do.

Brennan argues, in a nutshell that so long as you have, in good faith, diligently, and genuinely exhausted any and all peaceable options to stop the state or any other powerful institution from actively harming you and your family, and none of those peaceful options worked, and these powerful institutions are still harming you and your family, then it is just and good for you to commit violence against those institutions and the people which comprise them.

Now, I must stress, one of Brennan's most interesting arguments, and the one he defends most heavily, is that you really do have to try to exhaust the peaceful options first.

Brennan argues that once you have done that, then almost any and all violence is completely acceptable against the institutions or people which are harming you, that there is not an absolute ethical and moral admonition against violence as a concept; that violence is not a priori categorically bad. There are certain circumstances in which your duty is to commit violence with extreme intensity and prejudice.

Anyway, I just thought this was a really fascinating book that I read recently, and I would recommend it to all of you as a completely and totally unrelated aside.

If you take nothing else away from this manifesto, just know that, no, you are not crazy for thinking everything in the world is getting worse and everything is kind of bad and nobody really seems to be able to talk about it or know what to do about it. And yes, there is a way out.

As always, the answer is to get active.