AnalysisAI AgentsAugust 14, 2026

Sub-1MB blind script matches frontier models on OSWorld

Pierluca D'Oro (Programma Labs) records one successful trajectory per task, then replays those actions blindly — the script never looks at the screen. On deterministic benchmarks like OSWorld, that replay counts as valid and matches or beats the frontier model it was copied from.

Featured · Pierluca D'Oro

1 source

Daily brief

Get tomorrow's AI brief in your inbox

More stories today

Open the live feed