Why Crypto Failed to Free Us — and the One Lesson Worth Keeping

If you want to understand why crypto failed to free us, start with the promise, because the promise was genuinely beautiful. Money without a central bank. Payments without a bank taking its cut. Contracts that enforce themselves. A financial system that answered to nobody in particular, run by everybody at once. For a while it looked like the internet had finally found a way to hand power back to ordinary people — to route around the middlemen who sit between us and our own money and skim on the way past. I watched that dream take shape, believed a lot of it, and I still think the engineering deserves respect. But the world it built is not the world it promised, and the gap between the two is the most useful thing crypto ever taught us.

This is not a hit piece. Crypto proved something real, and I’ll get to it. But the honest story is that a technology designed to distribute power ended up concentrating it — often faster, and with less accountability, than the system it set out to replace. That’s not a bug in the story. It’s the whole lesson.

The promise: cut out the middlemen

The founding idea was elegant. Instead of trusting a bank to keep an honest ledger, you’d let thousands of independent computers keep the ledger together, each one checking the others. No single party could quietly edit the record, freeze your account, inflate the currency, or decide you weren’t allowed to transact. Trust would come from math and open code rather than from an institution you had to take on faith.

Around that core idea grew a much bigger dream. If money could be decentralised, why not everything money touches? Lending without banks. Trading without brokers. Organisations without executives, governed by token-holders voting on-chain. Art and identity and reputation, all owned by users instead of platforms. The word that carried all of it was decentralisation, and it was doing an enormous amount of work — because it quietly meant two very different things at once. It meant the technical fact that no single server ran the network. And it meant the political hope that no single group would end up in charge. Everyone heard the second when the technology only ever promised the first.

What actually happened: the plumbing spread, the wealth pooled

Here is the uncomfortable part. The code really did distribute. Anyone could run the software, read the ledger, verify the rules. That much of the promise was kept. But almost everything that mattered around the code concentrated, and it concentrated fast.

Coins concentrated. Early holders and insiders accumulated enormous positions before most people had heard the word, so a system meant to spread ownership started life with ownership already pooled at the top — often more unequally than the old economy it was meant to fix. Mining concentrated. What began as something you could do on a home computer became an industrial arms race, and the power to validate transactions gathered into a handful of large operations and pools with the cheapest electricity and the best hardware. Exchanges concentrated. Because raw blockchains are hard to use, most people reached crypto through a few big companies that held their coins for them — which is to say, through new intermediaries that looked a lot like banks, except without the deposit insurance, the regulators, or the obligation to give your money back. When some of those intermediaries collapsed, people learned the hard way that “not your keys, not your coins” had teeth.

A technology designed to remove the middleman was mostly used to build new middlemen — with less accountability than the ones it replaced.

Influence concentrated too, and this is the part that gets missed. Even in projects governed by on-chain voting, votes were weighted by tokens, and tokens were held unequally — so “one member, one vote” quietly became “one coin, one vote,” which is just wealth voting, dressed in new clothes. The venture funds that bankrolled the early networks held large allocations and shaped which projects lived or died. The influencers who could move a market with a single post accumulated a kind of power no central bank ever had. The system was leaderless in theory and steered by a small crowd in practice.

None of this happened because the builders were uniquely greedy. It happened because distributing the plumbing of a system does nothing, on its own, to distribute its rewards. Open the ledger to everyone and the person with the most capital, the cheapest power, and the biggest audience still wins — sometimes more decisively than before, because now there’s no regulator to tell them no. This is a pattern I keep running into everywhere technology touches power, and I’ve written about how technology gets captured more generally: a tool built to open something up gets quietly enclosed by whoever was already best positioned to own it. Crypto didn’t escape that pattern. It ran the experiment at high speed and produced a textbook case.

Be fair: what crypto actually proved

If I stopped there, I’d be telling you half the story, and the missing half matters. Because crypto did prove something that serious people thought was impossible.

It showed that a trust-minimised system can run at scale. A network of mutually distrusting strangers, with no central operator, kept a consistent global ledger for years, moved enormous value across it, and settled transactions without anyone in charge — and it mostly held. Whatever you think of the speculation stacked on top, that base-layer achievement is real computer science, and it answered a genuine open question. You can build a system where the rules are enforced by code and consensus rather than by a trusted authority. That’s not nothing. That’s a milestone.

It raised useful questions, too. Why does moving money across a border cost so much and take so long? Why do a handful of payment companies get to decide who is allowed to transact? Why is the plumbing of our financial lives so opaque, so slow, and so profitable for the people who own it? Crypto forced those questions into the open and pushed the incumbents to answer them. Some of the ideas it popularised — programmable money, transparent public ledgers, self-custody, settlement that doesn’t wait for banking hours — are good ideas that will outlive the speculative frenzy that introduced them. Being clear-eyed about the failure doesn’t require pretending the achievement away. Both things are true at once.

The one lesson worth keeping

So what do we actually carry forward? One sentence, and it’s the most valuable thing the whole episode produced:

Technical decentralisation is not political decentralisation. You can distribute a system’s plumbing and still centralise its wealth.

Spreading the machinery of a system across many computers tells you nothing about who ends up owning the value that machinery produces. Those are two separate questions, and crypto conflated them so thoroughly that a lot of smart, well-meaning people genuinely believed that decentralising the technology would decentralise the power. It didn’t. The technology went to everyone; the ownership went to a few. That’s not a detail. That is the story.

Once you see it, you can’t unsee it, and it changes the questions you ask about any new technology. Not “is this decentralised?” — that question is too easy to answer yes to while the wealth quietly pools. The real questions are simpler and harder. Who owns it? Who captures the upside when it succeeds? Who carries the risk when it fails? Who gets to change the rules, and who just has to live with them? A system can be technically distributed across ten thousand machines and still be politically owned by ten people. If you only check the plumbing, you’ll be fooled every time.

Now point that lesson at AI

I care about this less for crypto’s sake than for what comes next — because the exact same word, democratise, is now doing the exact same work in the pitch for artificial intelligence. “AI will democratise expertise.” “AI will democratise creativity, coding, medicine, education.” It’s a lovely sentence, and we should treat it with precisely the suspicion crypto taught us to have.

Apply the lesson directly. Yes, the tools are spreading to everyone — anyone with a phone can now use models that would have been science fiction a decade ago. That’s the plumbing, and it’s genuinely being distributed. But ask the ownership question and the picture changes fast. The largest models are trained at a cost only a few companies on earth can afford. The chips, the data centres, the electricity, the capital, and the scarce talent all pool at the top, in fewer hands than crypto ever managed. The people using the tools mostly don’t own them, don’t govern them, and can’t see inside them. If a handful of firms own the models everyone depends on, then “democratised AI” describes who gets to use the technology, not who gets to own it — and those are, once again, two completely different things.

This isn’t a reason for despair; it’s a reason for precision. When the value that used to flow to workers starts flowing to whoever owns the machines, the question of what everyone else lives on becomes unavoidable — which is why the argument for universal basic income is really an argument about ownership and distribution, not charity. It’s why the drift toward what some now call technofeudalism — a world where most of us are tenants on platforms and infrastructure we’ll never own — is the pattern to watch. And it’s why the most interesting counter-models are the ones that build ownership in from the start, like platform cooperatives where the people who create the value actually hold a share of it. Those experiments matter precisely because they answer the question crypto dodged: not just how do we distribute the technology, but how do we distribute the ownership.

Judge by who ends up owning it

So here is the rule I’d hand anyone the next time a technology promises to set them free. Ignore the adjective. “Decentralised,” “democratised,” “open,” “for the people” — these are marketing until proven otherwise, and crypto is the proof of how far the gap between the word and the world can stretch. Don’t ask whether a system is distributed. Ask who ends up owning it, who takes the profit, who pays the cost, and who can fight back when it goes wrong. Follow the ownership, because ownership is where the power actually lives.

Crypto didn’t fail because the technology didn’t work — much of it worked beautifully. It failed to free us because freedom was never a property of the code. It was always a question of who owns the thing the code produces, and on that question crypto gave the same old answer, just faster. That’s the lesson worth keeping, and we’re about to need it more than ever. The next technology that promises to hand you the future will be judged not by how cleverly it spreads its machinery, but by who’s holding the deed when the dust settles. Keep your eyes on the deed.