AI may expose the limits of authority-driven mathematics
jasondeanlee · x · 2026-07-23
The thread argues that AI may trigger a “BS-busting” era in mathematics, partly by exposing weak appeals to authority.
- The author notes that mathematicians often rely too heavily on “appeal to authority.”
- They point to examples like the Erdős unit distance conjecture and the Jacobian conjecture, both of which ended negatively, as cases where prestige and intuition can mislead.
- A reply asks a deeper question: if frontier AI is already doing strong math and even solving open conjectures, have we actually gained new understanding, insight, theory, frameworks, or mathematical objects?
The discussion is less about a specific model and more about whether AI will change how mathematical knowledge is produced and validated.
Related event: AI's Rise Sparks Debate on Authority and the Future of Mathematics(2 posts)→
More from AGI Musings
- Quote-tweet jokes that AI will keep humans alive only as training-data farms — markjeffrey · 2026-07-24
- Richard Socher’s new book argues AI can accelerate science across every stage — RichardSocher · 2026-07-24
- Richard Socher says people may still need “mind gyms” after superintelligence — RichardSocher · 2026-07-24
- Will Depue says commodified superintelligence could end capitalism in 20 years — willdepue · 2026-07-24
- AI mathematician debate turns into a spat over human mathematicians — littmath · 2026-07-24
- A satirical AI regulation list imagines kill switches, dataset registries and liability — NickPassig · 2026-07-24