OpenAI says an AI-generated, Lean-checked construction establishes finite-time blowup for a forced three-dimensional Navier–Stokes flow. This explainer breaks down the equation, the precise Clay alternatives, the limits of the claim, and why data provenance matters.
A Discovery Loop automates the propose → run → evaluate → learn cycle that drives scientific and engineering progress. Instead of manual, sequential iteration, it runs many experiments in parallel and learns from outcomes to propose better next steps.
A new preprint argues that many AI unicorn startups contribute almost nothing to the formal scientific record. Over half reportedly produce no citable paper or preprint, reducing reproducibility and making safety evaluation harder—especially when model details stay opaque.