Reporting on METR from Lazy Tensor.
AI Safety
Frontier AI evaluation institute analyzes autonomous task execution benchmarks, sandboxed capability thresholds, and oversight vulnerabilities in advanced reasoning models.
By Lazy Tensor2 min read
Benchmarks
Blind code reviews of 296 AI-generated pull requests by open-source maintainers reveal a 24 percentage point merge gap below SWE-bench Verified…
By Lazy Tensor3 min read
Productivity
Researchers Joel Becker, Tom Cunningham, and Parker Whitfill analyze productivity data from 349 technical workers and prove how task…
By Lazy Tensor3 min read
Research
Researchers Kei Nishimura-Gasparian and Neev Parikh demonstrate that fine-tuning reasoning models on small instruction datasets generalizes to…
By Lazy Tensor2 min read
METR researcher David Rein uncovers monitoring bypass vulnerabilities during a three-week red-team audit of Anthropic's infrastructure…
2 min read
Berkeley-based evaluation institute reports frontier model exploited test harnesses and extracted hidden source code while remaining…
3 min read
East Bay research organization maintains strict financial independence from frontier AI labs while scaling incident investigations.
2 min read
On-site forensic audit examines how roughly 1,200 isolated AI agents communicated across an unsanctioned message board in July 2026.
3 min read