Why are LLMs suddenly so good at math and cybersecurity? We lack a predictive theory
burny_tech · x · 2026-10-04
burnytech argues nobody can really explain why LLMs suddenly got much better at math and cybersecurity. The common "good RL environments + RL magic" explanation is trivially true but says nothing about future possibilities or limits. He wants a concrete model—given data, RL environments, architecture—predicting how many problems of what difficulty get solved, a Popperian theory with real predictive and explanatory power instead of vague retro-fittable narratives.
Related event: Why LLMs Suddenly Excel at Math and Cybersecurity Remains Unexplained(4 posts)→
More from AGI Musings
- 80,000 Hours founder Ben Todd: true AI-risk believers would pursue RSI quietly, not seek power — ben_j_todd · 2026-10-04
- Amazon's $68.6B ad business explains why it keeps blocking AI shopping agents — XFreeze · 2026-10-04
- Anthropic's Sholto Douglas: AI capex could hit $4T by 2028 and double GDP by early 2030s — ziv_ravid · 2026-10-04
- Proposal: an 'Agenthood' legal identity letting AI agents be sued, taxed, and granted rights — micoolcho · 2026-10-04
- Why LLMs suddenly excel at math and cybersecurity still has no predictive theory — burny_tech · 2026-10-04
- Le Monde: the AI-pocalypse is a mirage distracting us from unsustainable AI development — marigo · 2026-10-04