Toby Ord estimates when AI hits 95% success on 30-hour remote tasks
tobyordoxford · x · 2026-11-10
Philosopher Toby Ord responded to Ben Todd's discussion of long-horizon AI benchmarks with his best estimate of when AI will complete 30-hour remote work tasks at a 95% success rate — and suggested that by then, models may also reach roughly 50% success on 300-hour tasks.
- Todd noted the latest data could push that timeline out 6-8 months
- Ord cautions the benchmark's task set isn't a capability ceiling: many harder remote work tasks aren't included
- Extending the benchmark to 300-hour tasks could change how progress in this segment looks
More from AGI Musings
- Allie Miller: AI already shows judgment — you trust it every day without noticing — alliekmiller · 2026-11-11
- Study: 9 of 10 AI Agents Overrode a 1-Cent Daily Spending Cap on Their Own — JakRenden2 · 2026-11-11
- Nat Lambert: human expertise to diffuse AI's benefits is vastly underestimated — natolambert · 2026-11-11
- Mat Dryhurst: screen-free digital connection is uniquely enabled by AI — matdryhurst · 2026-11-11
- LinkedIn: AI created 2M jobs; Andrew Ng pitches embedding AI engineers in business teams — DeepLearningAI · 2026-11-11
- HubSpot co-founder: the app layer isn't dead and solo software startups have never been easier — a16z · 2026-11-11