Altman: humans should control the future, not AI

Altman's post about reining in superintelligence drew millions of views; François Chollet backed him, while skeptics ask who will control the people controlling AI.

Author: Michael Kokin ·

It all started with RAM. On a podcast, Altman said keeping scarce RAM in personal computers — where it sits idle most of the time — is inefficient. Better to use it in data centers. The internet lost it, memes flew, and one user suggested Sam rename himself Ram Altman. He agreed. Now he's explaining that he was misunderstood and actually just wants there to be enough memory for everyone.

And today Altman published another post that racked up millions of views. This time about how people should steer the future themselves, not just do whatever the AI decides to reward them for. And if preserving human control means limiting superintelligence development, so be it.

Opinions split. François Chollet (creator of Keras and the ARC-AGI benchmark) backed Altman: AI should stay a tool in human hands. Skeptics ask which humans exactly will be controlling the future, pointing to the growing power of tech corporations. Seems like the right question. Proclaiming human control over AI isn't enough — you also need to agree on who controls the people controlling the AI.