Essay predicts AI matching top human AI researchers within a few years
Start here
Dwarkesh Patel argues in an essay that AI systems as capable as the best human AI researchers could arrive within a couple of years, with 2028 as his median estimate and a 25% chance the date slips past 2032. He cites METR's finding that Anthropic's rate of progress on well-scoped tasks doubled in a year, attributing most of that acceleration to AI-assisted R&D, and Anthropic's report that Claude now leads over 26% of its AI R&D work, up from under 1% in February. Patel frames what follows as a 50/50 split: either a software-only singularity or a spark that compresses three to six years of progress into one. These are his projections, not demonstrated outcomes.
Generated from reports · updated 2 h ago
Report timeline
Follow the reports to see the event from different sides.
- Hacker News · AIEssay 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.
Heat for this event
Not enough continuous observations yet to draw a trend.