We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies. We continue to believe that iteratively putting great tools in the hands of people leads to great, broadly-distributed outcomes.
We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word. We love our current products, but we are here for the glorious future. With superintelligence, we can do anything else. Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity.
What’s the threshold where you’re going to say, “OK, we’ve achieved AGI now”? The very rough way I try to think about it is when an AI system can do what very skilled humans in important jobs can do—I’d call that AGI. There’s then a bunch of follow-on questions like, well, is it the full job or only part of it? Can it start as a computer program and decide it wants to become a doctor? Can it do what the best people in the field can do or the 98th percentile? How autonomous is it? I don’t have deep, precise answers there yet, but if you could hire an AI as a remote employee to be a great software engineer, I think a lot of people would say, “OK, that’s AGI-ish.” Now we’re going to move the goalposts, always, which is why this is hard, but I’ll stick with that as an answer. And then when I think about superintelligence, the key thing to me is, can this system rapidly increase the rate of scientific discovery that happens on planet Earth?
“We know how to build AGI”—Sam Altman
Link post
More context in this Bloomberg piece.