Ethan Mollick: AI task horizon went from 4 minutes to 17 weeks between his two books
emollick · x · 2026-10-23
- Ethan Mollick's new book Co-Existence arrives 30 months after Co-Intelligence, and the gap quantifies AI progress: when his first book came out in 2024, the best AI could complete 4 minutes of human work at 50% success; at draft time earlier this year it was 17 hours; at publication it reached 17 weeks (METR time-horizon data, Oct 2026 reading 8–55 hours; MirrorCode, June 2026, estimated Claude Opus 4.7 rebuilt a 16,000-line program in 14 hours — 2–17 engineer-weeks).
- He also made the book's charts by delegating to Opus 5.5 orchestrator agents in Claude Projects: a couple of sentences produced research, graphs, layouts, images and citations in minutes, and the agent even caught his math error ("28 months" vs 30).
- His takeaway: the workflow itself shifted from guiding a chatbot step-by-step to delegating whole tasks to agent swarms.
Related event: Wharton's Mollick says AI task horizon has surged to 17 weeks in new book(2 posts)→
More from AGI Musings
- Dave Morin: email is now other people's agents' todo list for your life — heyneighbor · 2026-10-23
- Robotaxi cumulative miles growing ~2.2x per month, could rival all US driving in 18 months — juanbenet · 2026-10-23
- Mathematician David Harvey on understanding the o(n log n) integer multiplication result — and AI — akbirthko · 2026-10-23
- "OpenAI/Anthropic will ditch users and solve quantum gravity" take gets mocked — QuintinPope5 · 2026-10-23
- Banks may ban AI agents on security grounds to fend off agentic bank runs — Gccooke · 2026-10-23
- People who explain complex tech clearly may become more valuable than those who build it — signulll · 2026-10-23