Hubert Dreyfus

Philosopher

A Harvard-trained Berkeley philosopher (1929–2017) and the leading American interpreter of Heidegger, whose 1965 RAND report and 1972 book What Computers Can't Do argued that human expertise is embodied skill rather than rule-following, a claim the AI establishment first ridiculed and later substantially conceded. With his younger brother Stuart, an operations researcher, he built that argument into the five-stage Dreyfus model of skill acquisition.

Lived
1929–2017

Dreyfus earned a B.A. summa cum laude from Harvard in 1951, an M.A. in 1952, and a Ph.D. in 1964 with a dissertation on Husserl.[1] On a Harvard traveling fellowship he studied at the University of Freiburg (1953-54) and later at the École Normale Supérieure in Paris.[1] He taught at Brandeis,[1] then at MIT, where the nascent field of artificial intelligence caught his attention.[6] None of this touches management: it is the training of a continental philosopher, and by his own Festschrift editors’ account, interpreting Heidegger and Merleau-Ponty (not AI) “came first in the order of his philosophical interests.”[1]

While still at MIT, RAND Corporation hired him to review Allen Newell and Herbert Simon’s AI research, producing the 1965 report Alchemy and Artificial Intelligence: his opening argument that machine symbol-manipulation could not reproduce the situated, embodied character of human skill. It “caused an uproar”:[3] Seymour Papert dismissed a third of it as “gossip,” Herbert Simon called it “garbage” for borrowing RAND’s name, and Dreyfus’s own MIT colleagues stopped eating lunch with him, except Joseph Weizenbaum, who ate with him deliberately to protest the field’s treatment of a colleague.[3] His MIT tenure case, still pending when the paper landed, was held in limbo in part because of his colleagues’ objections to it.[4] His 1972 book What Computers Can’t Do, expanded through three editions, made the same case at length and became his public reputation, however secondary that was to him.

The reception is the part worth weighing carefully, because it reversed. By the 1980s, robotics and the “sub-symbolic” approaches of connectionism began rediscovering what Dreyfus had argued from philosophy; by the era of neural networks and deep learning, AI historian Daniel Crevier concluded that “time has proven the accuracy and perceptiveness of some of Dreyfus’s comments.”[3] Dreyfus himself put it more bluntly in 2007: “I figure I won and it’s over, they’ve given up.”[3] His objections are now a standard part of AI’s own self-teaching, discussed in Russell and Norvig’s textbook of the field he once scandalized.[3]

Joining Berkeley in 1968 (winning the Harbison Prize that same year), promoted to full professor in 1972, he taught until his last class in December 2016: one institution for his whole career, not a portfolio of operating or consulting roles. A decade in, he co-wrote Mind Over Machine (1986) with Stuart, the book behind the Dreyfus Model: expertise moves through stages, and past a point it stops running on rules at all. His honors (a 2001 fellowship at the American Academy of Arts and Sciences, an Erasmus honorary doctorate for work “in the field of artificial intelligence” and continental philosophy, Berkeley’s Distinguished Teaching Award, and a 2004–2005 term as president of the APA’s Pacific Division)[5] recognize a philosopher and teacher, not a management thinker; the All Things Shining bestseller and doctoral students including Taylor Carman, John Haugeland, and Sean Kelly confirm as much from a different angle.

Dreyfus sits at the far adjacent end of the spectrum. Management was never his subject: he never wrote for managers, and nothing in his record suggests he thought about organizations at all.[6] What MC borrows is narrower and, because of that, more durable: his account of why an expert stops following rules, that skilled coping is holistic and situated, resistant to being written down, and degraded rather than improved by conscious rule-application. That is real authority, earned the hard way (ridiculed, then vindicated),[3] on the phenomenology of expertise. It is not authority on how to run a team, structure an org, or manage anyone; for that, MC leans on Stuart’s applied half of the same partnership, and on its own translation of the model into a coaching tool.

Hubert Lederer Dreyfus (1929–2017) was a professor of philosophy at the University of California, Berkeley, where he taught from 1968 until his death.[1] Educated at Harvard, he became the leading American interpreter of Martin Heidegger and a central figure in carrying phenomenology and existentialism into English-speaking philosophy.[1]

He was also the most prominent philosophical critic of artificial intelligence. His 1972 book What Computers Can’t Do argued that human expertise is intuitive and situated in concrete experience, not the rule-following a computer reproduces, an argument the early AI community resisted and later partly vindicated.[1] With his younger brother Stuart, an operations researcher, he built that view of expertise into the five-stage model of skill acquisition.[2]

1
"Hubert Dreyfus," Wikipedia, accessed July 27, 2026.
https://en.wikipedia.org/wiki/Hubert_Dreyfus
2
Stuart E. Dreyfus and Hubert L. Dreyfus, A Five-Stage Model of the Mental Activities Involved in Directed Skill Acquisition, ORC 80-2 (Berkeley: Operations Research Center, University of California, Berkeley, 1980).
https://apps.dtic.mil/sti/tr/pdf/ADA084551.pdf
3
"Hubert Dreyfus's views on artificial intelligence," Wikipedia, accessed July 27, 2026.
https://en.wikipedia.org/wiki/Hubert_Dreyfus's_views_on_artificial_intelligence
4
Sean D. Kelly, "In Memoriam: Hubert L. Dreyfus (1929-2017)," Harvard University Department of Philosophy, April 24, 2017.
https://philosophy.fas.harvard.edu/news/memoriam-hubert-l-dreyfus-1929-2017
5
"In Memoriam: Hubert Dreyfus," American Philosophical Association, April 22, 2017.
https://www.apaonline.org/news/341800/In-Memoriam-Hubert-Dreyfus.htm
6
Yasmin Anwar, "Hubert Dreyfus, preeminent philosopher and AI critic, dies at 87," Berkeley News, April 24, 2017.
https://news.berkeley.edu/2017/04/24/hubert-dreyfus/

Keep learning

Dreyfus ModelToolEvery skill runs on a ladder from novice to expert, with a sixth rung, mastery, for those who push further, and where you sit changes how you learn it and how you should be coached. The Dreyfus model is the map: name the rung, name the next one, and the vague feedback that helps no one turns into a plan.