Researcher: Google Astra solved a seminar open problem in 2-3 prompts
IgorCarron · x · 2026-10-15
Igor Carron shares an academic anecdote: after a seminar speaker posed an open problem (whether a certain construction can exist), the author asked Google Astra on the way home. It produced a concrete example in 2-3 prompts, then a more complicated second example on request, explained why the solution works during the tram/metro ride, and at home compiled the result into a research article — a striking demo of AI productivity in pure math research.
More from Research
- Shallow LLM guidance beats deep injection in VQA: 39.4% vs 31.8% on OK-VQA — Aneerban Saha · 2026-10-15
- First-of-its-kind multi-agent observability dataset: 8.9M events, 18 agents, released openly — strickvl · 2026-10-15
- Physics benchmark scores 8 video models: top Seedance 2.5 only 57.76/100 — jiqizhixin · 2026-10-15
- Toby Ord: OpenAI's RL scaling curves show no smooth slowdown, just logarithmic gains — tobyordoxford · 2026-10-15
- Top-labs survey: 5 of 7 robot self-improvement studies show no cross-task gains, zero confirmed L5 — rohanpaul_ai · 2026-10-15
- Cohere Labs Open Science Community hosts three public ML talks next week — Cohere_Labs · 2026-10-15