Google DeepMind researcher on math, research, and meaning in the age of AI
AndrewLampinen · x · 2026-09-21
Andrew Lampinen reflects on AI's rapid math progress — OpenAI's claimed Millennium Prize solution came just 1.5 years after a survey put 50% odds on such a solve by 2053. He argues the dueling mathematician essays (Fields Medalist critiques and counter-critiques) actually grapple with the same questions: is math about results, understanding, or process; what's the human mathematician's role if AI builds proofs strictly better; how to handle proofs that resist understanding; how to share partial results without being scooped by whoever has the most AI access; and how to assign credit. He expects these themes to hit other fields, mixing grief and excitement.
More from AGI Musings
- 'LLMs Are Just Vector Multiplication'? So Are Humans Just Atoms Obeys Physics — burny_tech · 2026-09-21
- Building Meaningful AI Software Will Feel Like Growing an Organism, Not a Factory — manosaie · 2026-09-21
- "LLMs Are Great, But Is Anyone Else Noticing the Death of Flow States?" — _AustinCalvert_ · 2026-09-21
- Richard Sutton flags a proposal to rescue Grokipedia via an open two-sided market for scientific data — RichardSSutton · 2026-09-21
- Drawing out the inner/outer loop of judgement: autonomy vs accountability — _AustinCalvert_ · 2026-09-21
- When AI proves theorems faster than mathematicians can absorb them, what happens to math? — burny_tech · 2026-09-21