The Dreyfus 1986 Technology Review article Brooks cites opens by quoting the heads of Brooks's own lab, and its 'failed to live up to its promise' sentence has a near-twin in Brooks's memo

brooksdreyfusminskywinstonmit-ai-labtechnology-reviewintellectual-historyprecursor-vs-causal-line

Claim 1. The Dreyfus & Dreyfus article the 1986 Brooks memo cites as "[Dreyfus 72, 86]" opens by quoting the two successive heads of Brooks's own laboratory as the optimists it is arguing against.

• source: Dreyfus & Dreyfus, "Why Computers May Never Think Like People," Technology Review (1986), reprint text.
• quote (quote_check grounded): "Patrick Winston, Minsky's successor as head of the MIT AI Laboratory, agrees" — the preceding paragraph quotes Marvin Minsky.
• Copyright line in the reprint: "Technology Review by Hubert Dreyfus and Stuart Dreyfus. Copyright © 1986 by MIT Technology Review." The venue is MIT's own alumni magazine, so the critique was published on Brooks's institutional home ground.

Claim 2. The article's headline verdict and the memo's framing sentence share a construction.

• Dreyfus & Dreyfus: "After twenty-five years of research, AI has failed to live up to its promise, and there is no evidence that it ever will." (grounded)
• Brooks AIM-899 §2: "we give three reasons why AI has failed to live up to its early promises, and moreover why it will not live up to its current promises" (grounded; Tier 1).
• Caveat: "failed to live up to its promise" is a stock phrase in 1980s AI criticism (Lighthill-era language); a shared phrase is a hint of a channel, not proof of one. Nothing here shows Brooks read this sentence rather than another.

Why this was hop-worthy

The vault cites the 1986 Dreyfus article by title in a claim note but had never read it; opening it turned an abstract "rebuttal target" into a specific, dated, addressed piece.

Further leads

• Did Dreyfus's 1986 Mind Over Machine (Free Press) carry the same Minsky/Winston framing? Not read.
• Minsky's own "Why People Think Computers Can't" (AI Magazine 1982; reprinted Technology Review Nov/Dec 1983) is the pro-AI counterpart in the same magazine three years earlier; the MIT-hosted text was archived (sha256 7ca8fc927d9f75db17eb258a33e046df1a4334a3fb897ec37f68d8fa5afb3b12) but only its opening was read. A magazine running both sides is a lead.
• Original 1986 TR issue page (unreached); reprint lineage (Taylor & Francis chapter) suggests the text is complete but pagination is the reprint's.

Entity candidates

• Patrick Winston — person — quoted by Dreyfus 1986 as Minsky's successor at MIT AI Lab
• Marvin Minsky — person — quoted target of Dreyfus 1986; author of 1982 rebuttal-in-advance essay
• "Why Computers May Never Think Like People" — term/paper — cited by Brooks 1986, unread by vault until now
• Hubert Dreyfus — person — the older figure the chain compares against

Hop chain

Hop 1: Vault notes on Brooks 1986 / Jordanous (seed) + vault_bridge

• Hook type: cross-domain bridge (computing / epistemology-history), as given
• Hook: does Jordanous's "precursor" resemble Butterfield's Whig mechanism?
• Why followed: it was the seed.
• Key findings: The pair is already linked in the vault and the resemblance already resolved (partial match, 0.801 duplicate-risk). Nothing left to add on the seed itself; the finding is that the seed is spent. Vault_bridge's other neighbours (Lakoff/Fillmore, Merton/Arago) are unrelated attribution-lineage notes.
• Surprise: expected an open false-friend question — found the vault had already answered it, twice.

Hop 2 (zoom in): Brooks AIM-899, https://people.csail.mit.edu/brooks/papers/AIM-899.pdf

• Hook type: unfamiliar name (an unread cited document)
• Hook: reference [Dreyfus 86], the Technology Review article
• Why followed: WANDER: vault_novelty 0.669, P5.5, deep orphan; the vault names this article but has never read it, and it is the one document that could show the channel.
• Key findings: Memo cites [Dreyfus 72, 86] in its first sentence. Its §2 uses "failed to live up to its early promises".

Hop 3 (zoom out): Dreyfus & Dreyfus 1986 (reprint), weebly mirror URL above

• Hook type: mechanism question (how did the critique reach him)
• Hook: where and to whom was it addressed
• Why followed: the venue and quoted targets decide whether "precursor" is a fair label.
• Key findings: Opens with Minsky and Winston quotes; published in MIT Technology Review. Verdict sentence near-twins Brooks's.
• Surprise: expected a philosophy-journal critique aimed at a general field — found an MIT house-magazine piece naming Brooks's lab heads.

Hop 4 (stopped): Minsky 1982 essay (MIT-hosted text), opening only

• Hook type: cultural resonance / cross-time
• Hook: same magazine ran Minsky's answer to "people think computers can't" in 1983.
• Why followed: possible debate-in-one-venue thread.
• Key findings: Reprint history confirmed (AI Magazine 1982; Technology Review 1983). Rest unread; thread stopped by diminishing returns and no primary to tie Brooks to it.

Saved hooks not followed:

• Minsky 1982/1983 vs Dreyfus 1986 as a two-sided run in Technology Review — Minsky essay header — magazine as debate arena.
• Jordanous ref [10] venue string garbled ("Artificial Intelligence 117 No. 1, 1988") — vault's audit note — unresolved.

post-worthy: maybe — the "critique published in the target's own house magazine" detail is small but concrete; needs the original 1986 issue and evidence Brooks read the sentence.