← home

The business of privatizing intelligence

Intelligence shouldn't be a metered service. There's no solution, but there are alternatives.

Sam Altman, Dario Amodei, and company have been selling the idea for a while that, in the near future, humanity will have access to PhD-level intelligence at any time. What they basically propose is that intelligence becomes a paid service, like electricity or water: you pay, and they provide it to you. To me, it sounds like a dystopia, but they keep repeating it over and over as if it were an inevitable future. Unlike Thanos, who carried the trauma of seeing his planet go extinct, these billionaires—whatever their own ghosts may be—seem fascinated with the idea of putting humanity’s progress behind a paywall.

If today a group of researchers or PhD students decided not to publish their work and keep it just for themselves, everyone would call them selfish and reject them. However, when the leaders of big AI corporations openly talk about limiting access to intelligence, they get applause when they should get rejection.

If you’re thinking about comparing artificial intelligence to a new programming language or a trendy framework, think twice. To understand the real impact of AI, we have to look back at the invention of the steam engine. Before that, production was slow, expensive, and purely dependent on manual labor. The invention of the steam engine revolutionized the world because it multiplied human power to levels never seen before and transformed society in just a few decades. AI is about to do exactly the same thing.

A more dominant bourgeoisie

The Industrial Revolution brought with it the rise of a new bourgeoisie. As the economic and social power of big entrepreneurs grew stronger, new key players emerged who ended up consolidating the system: a model defined by private ownership of the means of production and the self-regulation of prices through the market.

On the other hand, the future of AI, if intelligence becomes a centralized service like electricity or gas, looks much darker. Sooner or later, users will have to pay the real price per token. Although companies are currently operating at huge losses to subsidize prices, this won’t last forever. When the subsidy ends, only a few will be able to afford the real cost. The rest of the people won’t have access to the equivalent of a PhD, but, at best, to a teenager who finished high school as the top of their class.

Therefore, the most likely scenario is that the current tech bourgeoisie will consolidate an even greater advantage. This isn’t like the internet boom, when access was relatively flat: anyone, from their home, could launch a service, an app, or a game that would shake entire industries. All you needed was a computer in a room (or even in a public library) to destabilize entire sectors of the economy.

On the other hand, if access to PhD-level intelligence is reserved only for those who can afford it, the gap between the average person and those groups will only continue to grow exponentially.

There’s no solution, but there are alternatives

The reality is that the Silicon Valley elites are going to win no matter what. Whether AI turns out to be a bubble or not doesn’t matter to them at all; they don’t play by our rules. Trying to regulate AI through governments isn’t the solution because laws usually come too late and, even worse, end up stifling local development while protecting the monopolies of the usual players.

The ironic thing is that, in this scenario, the best thing that could happen is for the open-source community to have total freedom to experiment and research. I’d rather live in a world where knowledge flows and we all benefit—though some might take a bigger slice—than in a world where only the megacorporations hold the keys to technology. And that’s what really worries me: if intelligence becomes purely something you rent, the average person will be left completely defenseless.

The alternative is to buy the equipment and install it yourself at home. I’m not saying it’s easy, but it is simple. In the end, not everyone always buys new; second-hand markets exist for a reason. Finding good parts at a good price just takes a bit of patience. The process involves learning about AI, but it also gives you the power to teach it later to friends, family, and anyone who’s interested.

Dystopia is inevitable

Eventually it’s going to happen: you’ll be able to rent that PhD-level intelligence that Sam Altman talks about. It’s inevitable, but that doesn’t mean we should just sit idly by. If today you can run the simplest of models in your own home, consider yourself lucky; you’re on the right track. The truth is, it’s addictive to see an intelligent agent at work, especially when it solves tasks perfectly.

So there you have it, get up and buy a GPU.