Mathematicians pinpoint the moment AI cracked the Exact Overlaps Conjecture via Astra transcript
vishalmisra · x · 2026-10-10
Constantin Kogler and Samuel Kittle posted an arXiv paper (arXiv:2610.10511) proving the exact overlaps conjecture and its generalization in dimension one, giving a formula for the dimension of self-similar measures via random walk entropy and Lyapunov exponent, built on a new entropy inequality averaging variance across scales. Kogler says the work solves the same problem as paper #148 in OpenAI's announcement, with the key breakthrough occurring at 16:54 on September 27, made visible through vishalmisra's analysis of the Astra conversation transcript. Vishal Misra adds the key analytical point: the transcript shows progress reports but not the search between them—the explored-and-discarded paths are where the interesting computation lives.
More from AGI Musings
- Fields Medalist Hugo Duminil-Copin: Solving Problems No Longer Has Value — soumitrashukla9 · 2026-10-10
- Researcher compiles 'Bad AI Consciousness Takes' bingo card, with rebuttals to follow — dioscuri · 2026-10-10
- Dona Sarkar: handing writing to AI 'feels like handing over my brain' as creative roles rebound in tech — donasarkar · 2026-10-10
- AI safety researcher warns pro-AI biases in AI systems are quietly disempowering humans — DavidSKrueger · 2026-10-10
- MacroPolo tracker: China-educated researchers make up large share of US elite AI talent — burny_tech · 2026-10-10
- Anil Seth's 'Conscious AI and biological naturalism' collection now open access — anilkseth · 2026-10-10