Computer use against a five-minute clock beats most coding evals, Astra ties for best
AIFlow_ML · x · 2026-11-18
Argues that computer use under a five-minute clock reveals more about model capability than most coding evals. Notes Astra tying for best on that test, calling it a notable signal.
More from coding & agent
- Don't call every agentic output 'AI slop': a workplace etiquette pushback — MxMnr · 2026-11-18
- Google AI Studio adds custom domains and one-click Android app builds via Play Store — OfficialLoganK · 2026-11-18
- Google Design tethers dependent prototype parameters, exports JSON presets — GoogleDesign · 2026-11-18
- Switching AI assistants feels like moving apartments: context and skills portability pain — No-Obligation-4726 · 2026-11-18
- O'Reilly Ships "Evals for AI Engineers" by Shreya Shankar and Hamel Husain — sh_reya · 2026-11-18
- Open-Source Treg Jobs API Lets AI Agents Run a Full Job Hunt for About $0.03 — rohanpaul_ai · 2026-11-18