Martin Ford: Solve continual learning and AI's inability to do 'messy jobs' collapses
MFordFuture · x · 2026-10-05
- Martin Ford comments on Joshua Rothman's New Yorker piece arguing AI can't take "messy jobs" that rely on tacit knowledge, coordination and persuasion.
- Ford counters that this is only an observation of the present: since humans learn messy jobs, human-level AI should too. The missing ingredient is continual (on-the-job) learning over months and years — something even the best models lack.
- He predicts continual learning will become a key competitive nexus among frontier labs, noting some Chinese developers are reportedly focusing research there.
Related event: AI Jobs Apocalypse Not Here Yet, Messy Work Debate Heats Up(2 posts)→
More from AGI Musings
- Neuroscientist Anil Seth: current AI is 'vanishingly unlikely' to be conscious — anilkseth · 2026-10-05
- Mustafa Suleyman blasts Anthropic's Claude constitution and 'AI well-being' stance; David Sacks flags concern — SchoeneggerPhil · 2026-10-05
- David Krueger: "regulating AI will crash the economy" is an unexamined bad meme — DavidSKrueger · 2026-10-05
- Information Is Subjective: A Schrödinger-Style Take on the LeCun Bytes Debate — xiaosun86 · 2026-10-05
- LeCun amplifies take: the more powerful a technology, the more incumbents invoke risk to suppress it — ylecun · 2026-10-05
- Coping pattern watch: the "but humans are good at X" tweets are multiplying — danfaggella · 2026-10-05