Essay argues AI could match top human AI researchers within a few years

Dwarkesh Patel argues that AIs as good as the best human AI researchers could arrive within a couple of years, with 2028 as his median estimate and a 25% chance it slips past 2032. He cites METR's finding that Anthropic's rate of progress on well-scoped tasks doubled in a year, with most of that acceleration attributed to AI-assisted R&D, and Anthropic's report that Claude leads over 26% of its AI R&D work, up from under 1% in February. He frames the aftermath as a 50/50 split between a software-only singularity and a spark that compresses three to six years of progress into one.