No — he's reversed the take. He once said agentic AI was "robotic process automation on the cloud" and that VCs were overhyping it; now "I don't agree with it anymore. I think the opposite's now true." He thinks people are "quite flippant about the current capabilities" and are undervaluing "what's coming."
More from this episode
He concedes the original take "was probably true at the time." Now "I completely disagree. I think we've gone the other way." He thinks people, "especially as like a broader people outside of tech," are "undervaluing" what's possible — "assuming there's no like mass extinction event in our lifetimes," there are "probably multiple futures where you could say like the only limit is really our imagination," starting "in like computer space" and later extending to "world models but for physical reality."