If AI proves every theorem and nothing changes, is theorem-level math a sham?
kev0406 · reddit · 2026-11-21
An honest Reddit discussion: AI is starting to solve problems mathematicians worked on for decades. But if these proofs keep coming with no real breakthroughs—no new technology, no new physics, nothing that changes how we live—does that mean theorem-level mathematics is mainly an academic exercise that keeps departments funded and journals full?
The poster also offers the counterpoint: perhaps theorem-level math was never meant to deliver on a schedule, and its value may surface 50-100 years later, the way number theory eventually underpinned modern encryption. They invite working mathematicians and researchers to weigh in on where the real payoff lies—and how we'd know if it isn't coming.
More from AGI Musings
- Chinese Tech Expats Predicted DeepSeek's Efficiency Push From Chip Export Bans — gandamu_ml · 2026-11-21
- FAccT community criticized for silencing dissent via moral accusations — KLdivergence · 2026-11-21
- Reid Hoffman: AI's Biggest Value Source Is Invisible in GDP — Consumer Surplus — reidhoffman · 2026-11-21
- Cosmologist Brian Keating: Test If AI Can Reinvent 20th-Century Physics From Pre-1900 Data — JosephJacks_ · 2026-11-21
- Seven years of AI skepticism in seven sentences, ending with 'I always said it would work' — c_valenzuelab · 2026-11-21
- Pro-AI vs anti-AI: who's really using AI to mask insecurity? — JHochderffer · 2026-11-21