An essay argues AI is entering a brute-force loop for math, software and drugs
bennstancil · x · 2026-07-27
This essay argues that a new kind of AI progress is emerging in verifiable domains like mathematics, software, and drug discovery: long-running loops where the system keeps generating hypotheses, testing them, and asking itself to “continue the search.”
It opens with Terence Tao and an Anthropic researcher using a frontier model to find a counterexample to the Jacobian conjecture, then contrasts that with a startup founder using just a few terse prompts to get ChatGPT to disprove another conjecture. The author’s point is that some problems no longer require a human to “operate” the model in a sophisticated way; they may only need a machine to relentlessly improve its own attempts. That could reshape domains with clear, checkable answers.
Related event: AI Pushes Into Mathematical Search, but Deep Theory Remains Hard(3 posts)→
More from AGI Musings
- 'Refusing to code with AI in 2026 is like refusing Google in 2005' — prasenx · 2026-10-16
- Bessemer survey of 226 healthcare execs: back-office AI returns 3-4x, clinical AI lags — AllThingsApx · 2026-10-16
- SAIR Launches Open Math OS, Offering Free Compute and 100+ Tools for Community-Owned Math AI — Thom_Wolf · 2026-10-16
- 'Human in the loop' called one of AI governance's biggest lies: presence isn't authority — glenbeer · 2026-10-16
- Interpretability debate: 'all concepts are arbitrary' dismissal called the default doomer prior — basedjensen · 2026-10-16
- "Flyers or Liars": the Wright Brothers' 1908 European skeptics as a lens on AI ambition today — bennash · 2026-10-16