Du är ute och cyklar. Dom åt upp inte alla men flera hundra av dom tusen mest kända matematiska problemen. Yrkesgrupperna du nämner äts upp varje år, grafisk designer, översättare, filmare, programmerare nu senast alla matematiker som inte har ett Claude abonnemang.
Citat:
Ursprungligen postat av
mindstream
Hundratals? Bara 2025 publicerades över 130000 forskningsartiklar inom matematik. Det är människor som använder olika verktyg, inte T-800 som slaktar människor. Redan 1976 blev det uppståndelse kring ett datorstött bevis av fyrfärgssatsen. Oron finns i många yrkesgrupper, men de flesta ser det nog som lite bättre tekniska hjälpmedel.
Citat:
OpenAI has announced three major scientific breakthroughs, spanning mathematics, computer science, and theoretical physics.
Mathematics: A proof of the local smoothing conjecture in all dimensions, which also resolves the Kakeya and Stein restriction conjectures.
Computer science: A disproof of the Exponential Time Hypothesis, a long-standing assumption about the computational difficulty of certain problems.
Physics: Rigorous existence proofs for quantum field theories, including QCD, alongside a completion of the Standard Model. The announcement addresses Yang–Mills existence but does not claim to solve the full Millennium Prize problem.
Citat:
Claude Audits Quantum Gravity
An AI re-derived a published one-loop quantum gravity calculation step by step. It found a sign error and a missing term, and confirmed the physical conclusion.
License: MIT Python 3.10+ checks: SymPy + mpmath paper: PDF
The paper: E. O. Kahya, Auditing a quantum gravity calculation with an AI model: corrected scalar self-mass and unchanged late-time mode functions during inflation (submitted to arXiv, 2026). PDF · LaTeX source
This repository re-derives, step by step, a published one-loop quantum gravity calculation during inflation: the graviton correction to a massless, minimally coupled scalar on de Sitter space (Paper I, Phys. Rev. D 76, 124005 (2007)), and its use in the scalar's late-time field equation (Paper II, arXiv:0710.5282). The re-derivation was carried out with Claude (Anthropic's AI model, Claude Opus 5.5) as a checking tool. It was published by E. O. Kahya, one of the authors of both papers. Each step, from the Feynman rules (I.6)–(I.8) to the renormalized self-mass (I.167) and the late-time analysis of Paper II, was derived again from the inputs stated in the papers, keeping the dimension
D
general wherever the paper does, and cross-checked with independent SymPy/mpmath code. (The bookkeeping tables of Paper I's appendix were audited rather than redone; see Limitations.) Every difference from the printed papers is classified: 4 substantive findings (S1–S4), 6 finite errors absorbed into counterterms (B1–B6) and 23 typographical errors (C1–C23). The complete list of findings gives each one with its evidence.