Google's Milan Far: Automation speed inversely tracks domain entropy — math and coding fall first

docmilanfar · x · 2026-10-07

Google researcher Milan Far argues that the speed of automation is inversely proportional to the entropy of a domain, and ranks fields by difficulty:

Math and coding are the most structured, which is why AI does them so well — they'll be automated first.
Hard sciences come next: governed by strict physical laws but complicated by noisy measurements and lab experimentation.
Financial markets are harder still — complex adaptive systems driven by human psychology, with abundant data but non-stationary statistics.
Clinical medicine is far harder than most realize: high-stakes, fragmented signals, idiosyncratic biology and behavior — "a nightmare of incomplete information."
Law builds on structured text and precedent but leans on ambiguous language, evolving norms, rhetoric and human judgement — not automated anytime soon.
Culture, art and strategic leadership are safest: tacit knowledge, intuition, emotional intelligence and moral trade-offs are "the pinnacle of disorder and human creativity."

His takeaway: we'll have artists, lawyers, doctors, and CEOs for a long time yet.

More from AGI Musings