Why in vault: Founder shared 2026-07-21 (paired with [[2026-07-21-huber-12-factor-companies]] — this is the theory essay; the 12 factors are its checklist). The best compact articulation in the vault of the Coasian argument for why AI reorganizes firms rather than merely accelerating tasks.
Bias flag: same as the companion — Chroma CEO; "the constraint has shifted from reasoning to context" is his business model as a sentence.
Argument chain
- Ford lesson, non-obvious version: not "machines replaced craftsmen" but "the firm was reorganized around a new bottleneck." Chassis time 12.5h → 93min; Model T $850 (1908) → $260 (1925) while doubling wages. Competitors who kept the old form died.
- Coase 1937: firms exist because transacting is costly (search, negotiate, coordinate, enforce). "The firm is its coordination friction… That friction is not a defect in the firm. It is the substance of the firm." Most knowledge work IS internal transaction cost (the Slack message, the alignment meeting, the summary doc).
- Disaggregation Theory (his play on Ben Thompson): AI as substrate, not point-acceleration → information friction → 0 → firms disaggregate at two levels: more contractor-shaped work ("using Claude is renting intelligence with 0 friction") and humans shifting from doing the work to governing it. "AI disaggregates execution, but re-aggregates power around context, data, trust, permissions, and governance."
- The bottleneck has already moved: "Given appropriate context, frontier models exceed the median knowledge worker on most discrete tasks… the people most confident this is not yet true are the people who have not seriously tried."
- Ellison parallel: don't encode the last 20 years' process into the new medium — redesign for the next 20. Cites Ramp: compound individual wins into organizational capability. Firms get less time than Ford's competitors did.
Read against the vault
- Sits with the after-automation chain ([[2026-05-21-every-after-automation]] → Parrott follow-up → Taylor Counterpoint): Shipper says humans stay ahead as framers; Huber says humans re-aggregate around governance/taste/context. Compatible, but Huber's is the more operational claim — and closer to the RDCO/CAF position (eval gates + onboarding gates = governance machinery).
- Direct feed for DIE/CAF positioning: "climb the autonomy ladder fastest" + firm-as-friction is a sales frame for assessment→portfolio→production. The "quick" framing also echoes the founder's MrBeast/critical-component discipline.
- The Coase section is the piece to steal for Sanity Check — as evidence inside an original re-frame (e.g., "your meeting load IS your firm" angle), not as a summary piece.
Cross-links
- [[2026-07-21-huber-12-factor-companies]] · [[2026-05-21-every-after-automation]] · L5 north-star doc · CAF/DIE roadmap pack
Why this is in the vault
- Best compact articulation in the vault of the Coasian argument for why AI reorganizes firms rather than merely accelerating tasks — grounded in economics, not hype; directly usable in phData/CAF conversations
- "AI disaggregates execution but re-aggregates power around context, data, trust, permissions, and governance" is the one-sentence case for why the RDCO vault+cellar+eval-gate stack is durable
- Paired with the 12-factor companies companion as a theory+checklist reference pair — this essay supplies the "why," the 12 factors supply the "how"
- The Coase section is the piece to extract for Sanity Check as evidence inside an original re-frame (not a summary)
Mapping against Ray Data Co
- CAF/DIE positioning: "climb the autonomy ladder fastest" + firm-as-friction is a sales frame for assessment→portfolio→production; "the quick" framing echoes the MrBeast/critical-component daily-discipline
- phData: "most knowledge work IS internal transaction cost" (the alignment meeting, the summary doc) is the piece that makes AI reorganization legible to clients — more compelling than efficiency arguments
- Sanity Check: Coase angle needs an original re-frame ("your meeting load IS your firm" or similar) to be non-derivative; the essay is evidence, not topic
- RDCO L4→L5: the bottleneck-shift from execution to context/taste/governance is the north-star framing restated in economic theory; use alongside the 12-factor companion