Browser Agent Learns Tasks by Watching Users, Tops Human Scores
New browser AI agent cotomi Act learns by watching users work, scoring 80.4% on WebArena — above human baseline. It builds shared task boards and wikis from observed browsing patterns.
-
-
AI Eval Costs Surge, Threatening Independent Model Oversight
AI eval costs now rival training expenses — a single benchmark run can hit $320K. Researchers warn this “accountability barrier” is pricing out academic & gov safety bodies, leaving oversight to the labs themselves.
