What made it most depressing, upon replaying the conversation afterwards, was that he thought he was describing utopia.

A few months ago, I met a billionaire AI tech bro (to use the accepted terminology). For half an hour this man — our terms of journalistic engagement mean I cannot say who — outlined his vision of the future. It was a future without disease, starvation or want. It was also a future without human dignity, struggle, achievement or purpose. Our conversation was, perhaps, the bleakest half hour of my life.

This week, as every week, another pillar of human exceptionalism crumbled. The International Mathematical Olympiad, the hardest mathematical contest in the world, convened. Alongside the student mathletes, OpenAI and Google entered. Both reached the highest grade: gold. Ten years ago, you wouldn’t have backed AI to do a GCSE. Now it can answer questions so difficult only a few dozen kids take part each year.

I know the riposte. Humans always fear machines. Fretting about this is like worrying that a calculator is better at long division. So, it’ll be fine. Like the Luddites we rage against job losses. When the economy settles, we find we are richer, healthier and have better jobs that better use our uniquely human skills.

AI won’t take your job, goes the mantra, but if you don’t get with the revolution, someone using AI will. My response? You’re not listening.

The tech bros don’t anticipate humans getting the better jobs; they think there won’t be better jobs. In their future there will be nothing humans can do that AI can’t. Bottom wiping and plumbing will hold out a little longer — until the AI designs robots to do those, too.

Now you may think that is ridiculous. You may think our brains do many things that AI can’t. I agree. Sometimes I use AI and it is like magic. Sometimes it is moronic. It is, as critics argue, simply “fancy autocomplete”. But for how much longer?

The brain isn’t mystical; it is material. It can be replicated and exceeded. Maybe that is genuinely too hard in practice. Perhaps there’s something fundamental in cognition we won’t ever achieve. Or maybe even humans are just fancy autocomplete.

So let’s stop with the platitudes. Two weeks ago, the CEO of Meta, Mark Zuckerberg, offered a star AI engineer a $200 million multi-year pay package. These are the hiring decisions of a company in a hurry, a company that thinks humanity is on a precipice and time is ticking to be the first to jump over.

Perhaps they will go splat. Maybe, after this burst of regrettable hubris, in ten years Meta’s HR department will be trying to manage out the programmer on the nine-figure salary. Maybe humans will always be special and true artificial intelligence — whatever intelligence is — will elude us.

But we should at least engage with the fact that the richest, cleverest, best-funded men in the world think otherwise. And we should understand what it means if they succeed.

It’s their utopia; we will just be living in it.