Mathematicians debate AI superhuman reasoning: even solving Navier-Stokes wouldn't end human expertise
littmath · x · 2026-09-16
- A debate sparked by an essay imagining an AI solution to Navier-Stokes: Tony Feng argues that even if AI could solve any precisely formulated problem faster than humans, it wouldn't be "robustly" superhuman, since asking good questions is core to mathematics.
- Commenters agree human mathematicians remain essential, and AI actually makes it easier to become an expert by explaining concepts and walking through proofs.
- Consensus: academia should still start preparing for the future the essay depicts.
Related event: Mathematicians Debate Whether AI Can Truly Surpass Human Reasoning(2 posts)→
More from AGI Musings
- Aligning AI through empathy: Rorty's sentimental-stories argument revisited — yeastsplainer · 2026-09-16
- AI slop pushes veteran reviewer's conference service to 10-year low — anmarasovic · 2026-09-16
- Debate: building AI doesn't make you the right kind of expert on its risks — binarybits · 2026-09-16
- Researcher proposes desk-rejecting Pangram-flagged AI slop papers unread intro checks — anmarasovic · 2026-09-16
- Opinion: Empathy Toward Humans May Be the More Reliable Path to AI Alignment — yeastsplainer · 2026-09-16
- The Singularity Has Already Started, Just Wildly Unevenly Distributed — ns123abc · 2026-09-16