Import AI 470: No rights for machines; automating environment generation with SPADE; and building better GPU kernels with Hawkeye
METR finds AI dramatically accelerated cyber vulnerability discovery but not AI research itself
“The rate of vulnerabilities reported across many projects has dramatically accelerated in 2026 compared with 2025, both for specific projects (cURL, OpenSSL, Firefox, and Microsoft) and for aggregate vulnerability databases (the US NVD, and OSV)”— Jack Clark
A METR study finds AI's impact on scientific progress is uneven: major acceleration in cyber vulnerability discovery, minor and hard-to-measure gains in mathematics, and no measurable acceleration in AI research optimization across seven benchmark problem areas. Jack Clark frames this as 'differential acceleration' driven by domain-specific phase changes, with cyber joining day-to-day coding as a field that has clearly crossed a threshold in 2026. The finding matters because it challenges the assumption that AI is uniformly accelerating all science — the benefit is lumpy and field-dependent.