Gavin Leech finds autonomous AI outperformed human-collaborative teams and historical baselines at resolving Erdős conjectures
Autonomous AI found mathematical counterexamples at a 45% rate.
——0——
@gleech @nc_znc I'm not sure this is the best analysis to do, though: one of the recent human disproved conjectures was disproved using an LLM-inspired technical (human, but causally downstream of the machines). So it may be better to do a purely temporal plot.
gavin leech (Non-Reasoning)@gleech@geoffreyirving seems you're right (Erdos problems, human vs teaming vs AI): +17% diff, nearly significant at 5% thanks to @nc_znc
2:08 PM · May 29, 2026 · 777 Views
9:33 PM · May 29, 2026 · 56 Views