METR reports AI capabilities scaling exponentially but reveals significant weakness in judgment tasks

According to AI evaluation nonprofit METR, an updated capabilities scaling report shows AI models continue improving along exponential trajectories for technical tasks but reveal a substantial capability gap in judgment-based work. Technology Review reported METR's findings suggest models perform significantly better at narrow technical benchmarks than on broader reasoning requirements.

Topics

Agent observability

Sources

This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.