Dwarkesh Patel changed his mind within three months, from thinking superintelligence was a decade or more away to believing AI could match the best AI researchers within a couple of years.

Asked earlier this year how far away superintelligence was, Dwarkesh would have said at least a decade. Three months later he has changed his answer: a couple of years, possibly a couple of months.
His evidence is mathematics. In three years AI went from failing at four-digit addition to winning a gold medal at the International Mathematical Olympiad to solving a Millennium Prize problem in 88 hours. AI research, like math, has verifiable scoring, so he expects the same avalanche in his own field.
What I find more interesting is his own caveat: this could still slip past 2032. Not many people give a timeline and put a probability on it.



Curated from high-quality sources, with concise summaries and key takeaways.