AI Won't Save the Kingdoms We Built

John Cutler · rss · 2026-05-25 source ↗ Reader 🖥️

AI as Amplifier Context Engineering Coordination Cost Hierarchy & Org Charts Multiplayer vs Single-Player AI

AI will only dissolve the organizational fiefdoms of the ZIRP era if leaders are honest that those fiefdoms were structural responses to coordination, autonomy, and accountability problems — otherwise AI will simply reproduce the same coordination burden in a new, async, agent-mediated shape.

practical · argumentative — Cutler argues for organizational action — he is telling leaders what they must do (make hard structural choices) and what they must not do (use AI to paper over the political economy of fiefdoms). It is argumentative prose with diagnostic and prescriptive registers, addressed to leaders who built or inherited the GM-fiefdom model.

Structure

Part	Role
Opening provocation: the same people who built the fiefdoms are now complaining about them and reaching for AI as escape.	Framing / problem statement
Historical reconstruction: the GM-fiefdom model rested on three assumptions — autonomy creates speed, business accountability creates focus, and large orgs need 'heads of things' (geometric / Dunbar argument).	Genealogy / context
Diagnosis: 'The pain was always there' — dependencies compounded, end-to-end experience fractured, coordination burden was pushed down to the front lines because senior leaders would not prioritize across fiefdoms.	Evidence of the pre-existing dysfunction
Enter AI: a plausible promise — smaller teams, AI as coordinator, context-transferer, dependency-surfacer.	Steel-manning the optimistic case
Limits of AI: context transfer ≠ shared understanding; AI does not clarify decision rights, neutralize incentive misalignment, resolve contested ownership, or create slack.	Core counter-argument
The real opportunity: AI works only as amplifier of a coherent operating model — clearer end-to-end ownership, decision rights that match responsibility, senior leaders willing to make tradeoffs.	Prescription
Coordination theater: AI gives solo-tuned operators new tactics to avoid deciding together — agents talking to agents, persuasive deflectors, async overload pushed onto collaborators.	New failure mode / warning
Closing: AI supercharges existing habits — avoidance, fiefdom-preservation, solo-optimized work — unless we are honest about what the problem was.	Conclusion / call to honesty

Key terms

Fiefdom Coordination Cost
A semi-autonomous sub-organization — typically a GM-led 'product' or area — with its own budget, team, and P&L-ish accountability, deliberately walled off from cross-org coordination.
GM mindset / GM-fiefdom model Coordination Cost
An organizational design pattern in which a leader owns the outcome of a business area (like a general manager), not just the work, on the assumption that this produces faster, more focused decisions.
ZIRP wave / ZIRP era Coordination Cost
The zero-interest-rate-policy era of cheap capital that funded organizational expansion: more layers, more heads-of-things, more 'products' even when they were not really products end-to-end.
Geometric problem Coordination Cost
The combinatorial explosion of coordination paths as headcount grows — implicitly anchored on Dunbar's number — which made hierarchical sub-orgs feel inevitable.
Coordination burden / coordination tax Coordination Cost
The work of discovering dependencies, negotiating exceptions, routing around politics, and absorbing ambiguity — pushed downward to the front lines when senior leaders refuse to prioritize across fiefdoms.
Context transfer Context Engineering
Moving information, summaries, retrievals, and pattern-matching across an organization — what AI is genuinely good at.
Shared understanding Multiplayer vs Single-Player AI
Joint commitment about who decides, what tradeoffs matter, and which conflicts must be resolved — produced by negotiation among people, not by information delivery.
Coordination theater Coordination Cost
The appearance of working together produced by agents, copious documentation, and persuasive deflectors — without the friction of actually deciding together.
Async abundance Coordination Cost
The flood of agent-generated artifacts, messages, and decision proxies that overwhelms collaborators who still try to decide together.
Operating model (coherent vs. incoherent)
The combination of ownership, decision rights, and prioritization that determines whether AI amplifies clarity or amplifies dysfunction.

Propositions

explicitcorroboratedThe people who built and benefited from the ZIRP-era fiefdom model are now the loudest voices calling for AI-enabled flattening.conf: high
supports ←The engineering manager role is splitting in two · p6 — Narrow overlap: both frame the current wave of AI-enabled flattening as cost-/politics-driven rather than a deliberate, principled organizational design — Moreau's 'happening through cost-driven layoffs, not deliberate design' aligns with Cutler's account of flattening pushed by self-interested advocates.
explicitThe GM-fiefdom model was a rational response to real problems (autonomy/speed, accountability/focus, and the geometric coordination problem), not an arbitrary mistake.conf: high
explicitdevelopedFiefdoms persisted despite visible pain because senior leaders were unwilling to prioritize across them and fief-leaders had self-interest in preservation.conf: high
extends ←TBM 433: The Border Collie Faustian Bargain · p6 — The Kingdoms essay states as a structural fact that fief-leaders have self-interest in preserving fiefdoms despite visible pain. TBM 433 extends this by supplying the individual-level mechanism for how that self-interest is actually produced -- not as a starting trait but as the gradual output of a Faustian bargain in which staying close to power to fix things converts the fixer into someone who now protects the very dysfunction they once opposed.
explicitWhen structural problems are unresolved, the coordination burden migrates downward to the people closest to the work.conf: high
explicitContext transfer is not the same thing as shared understanding.conf: high
explicitdevelopedAI cannot clarify decision rights, neutralize incentive misalignment, resolve contested ownership, or create coordination slack.conf: high
extends ←TBM 428: Yes, Robot. Yes, Boss? · p2 — Kingdoms lists specific things AI cannot fix (decision rights, incentives, ownership, slack); this doc generalizes them into a capability-vs-motivational blocker taxonomy that explains why.
explicitestablishedAI is only genuinely useful as an amplifier of an already-coherent operating model; it cannot substitute for one.conf: high
extends ←🧠 Community Wisdom: How AI is changing product operating models, tracking work stress with Whoop, whether you nee… · p8 — The GTM thread applies Cutler's thesis that 'AI is only genuinely useful as an amplifier of an already-coherent operating model' to go-to-market: the founder's ICP/positioning clarity is the coherent system that AI agents and agencies can amplify — absent it, tools don't substitute for the missing model.
extends ←🧠 Community Wisdom: How AI is changing product operating models, tracking work stress with Whoop, whether you nee… · p10 — The practitioner question reframes AI's effect on product operating models from speed to structure/handoffs/decision-rights — the same operating-model axis on which Cutler argues AI can amplify but not create coherence. (Edge rests on the asker's framing only; the community's answers are paywalled.)
supports ←AI-Assisted Software Development: Team Profiles and Capabilities for Putting Research into Action · p1 — Both frame AI as valuable only as an amplifier of an already-coherent operating model/organizational system, never as a substitute for one — Cutler's leadership essay and DORA's own research converge on the same claim.
supports ←John Cutler's Post · p3 — The sledgehammer-between-shack-and-mansion sketch (the tool's value depends on the wielder/context) supports his earlier, more explicit claim that AI is only genuinely useful as an amplifier of an already-coherent operating model and cannot substitute for one.
supports ←Managers Are Not Overhead: They Are Infrastructure · p6 — Both frame AI as valuable only through an existing coherent human layer — Cutler's 'AI is only genuinely useful as an amplifier of an already-coherent operating model' matches this essay's claim that managers are the layer translating AI strategy into team-level practice.
extends ←TBM 434: How Maps Can Hide Problems · p8 — Sharpens Cutler's own amplifier thesis ('AI is only useful as an amplifier of an already-coherent operating model; it cannot substitute for one') with a specific mechanism: AI-smoothing that manufactures a false appearance of coherence, which can actively worsen an incoherent operating model rather than merely fail to fix it.
explicitAI gives people with a bias toward working alone new tactics (agents, deflectors, copious docs) to avoid actually deciding together.conf: high
explicitcorroboratedAI will recreate the old coordination problem in a new shape — async abundance — where the burden again lands on the front lines.conf: high
supports ←Managers Are Not Overhead: They Are Infrastructure · p8 — Both are skeptical that flattening simply removes coordination need; Cutler argues the coordination burden migrates and resurfaces elsewhere, this essay argues flattening's visible effects (more scope, more reports) are evidence the underlying managerial work hasn't actually shrunk.
explicitestablishedAI is fundamentally a neutral amplifier: it supercharges whatever habits, postures, and incentives already exist.conf: high
extends ←🧠 Community Wisdom: How AI is changing product operating models, tracking work stress with Whoop, whether you nee… · p9 — Bal Sieber's 'ads multiply whatever your signup-to-first-value path already does' is Cutler's neutral-amplifier thesis ('AI supercharges whatever habits and postures already exist') expressed in the domain of paid customer acquisition — a leaky funnel gets amplified, not fixed.
extends ←How to Turn Individual Talent into Organizational Excellence · p1 — Cutler frames AI as a neutral amplifier of whatever habits, postures, and incentives already exist; this article makes the structurally parallel claim for organizational systems generally — the system amplifies whatever talent, standards, or routines already exist, for better or worse — without using Cutler's amplifier vocabulary.
extends ←When Persuasion Is (and Isn't) Manipulation · p10 — Cutler's general organizational claim that AI is a neutral amplifier of whatever habits/incentives already exist is extended by Marshall into the specific domain of interpersonal persuasion ethics, naming it an 'ethical amplifier.'
supports ←SWLW #710: Revised rules of Eng Leadership, When AI costs more than the engineer, and more. · p5 — Bender's 'magnitude, not direction' restates Cutler's 'AI is a neutral amplifier of existing habits, postures, and incentives.'
supports ←TBM 428: Yes, Robot. Yes, Boss? · p1 — Same author restating and sharpening his 'AI is a neutral amplifier of existing habits/incentives' into 'amplifies the direction you are already moving'.
implicitcorroboratedThe 'real' fiefdom problem is political-economic (incentives, ownership, decision rights), not informational.conf: high
supports ←TBM 426: The Trouble With Mirrors · p13 — TBM 426's 'seeing clearly doesn't mean you can act on it' and 'truth is negotiated by power, incentive, timing, capacity' independently support Cutler's earlier claim that the real fiefdom/organizational problem is political-economic (incentives, ownership, decision rights), not informational — visibility does not move a system whose binding constraints are relational and political.
implicitHard organizational choices (end-to-end ownership, matched decision rights, willingness to prioritize) are knowable and available — what's missing is the will to make them.conf: high
explicitCollaboration has measurable economic value ('save millions of dollars of wasted effort') and is among the highest-leverage activities in an organization.conf: high
implicitdevelopedDunbar-style cognitive limits on group coordination are real and impose hard ceilings on collaboration scale.conf: medium
extends ←🧠 Community Wisdom: How AI is changing product operating models, tracking work stress with Whoop, whether you nee… · p6 — Aaron Nichols's 'you simply do not stay up to date on what everyone is doing' is the managerial-attention expression of the Dunbar-style cognitive ceiling on coordination that Cutler invokes — hard limits on how much a person can track push oversight from status-collection toward trust and context-giving.
implicitThe end-to-end customer experience is the correct unit of organizational design, against which fiefdoms are judged a failure.conf: high

Arguments

AI alone will not dissolve fiefdoms. strength: high

Explicit premises:

Fiefdoms were sustained by political economy: misaligned incentives, contested ownership, and senior leaders unwilling to prioritize tradeoffs.
AI improves context transfer but does not change decision rights, incentives, ownership, or slack.
Context transfer is not the same as shared understanding.

Implicit premises:

The binding constraint on cross-fiefdom coordination is political/incentive, not informational.
Information friction was a downstream symptom of structural design, not the root cause.

Inferences:

Any AI-driven 'flattening' that does not also restructure decision rights and incentives will reproduce fiefdom dynamics within the AI layer (e.g., teams hoarding agents, agents enforcing local optima).
Vendors and leaders who pitch AI as a coordination solution are likely to be the same actors who benefit from leaving incentives unchanged.

AI can make the fiefdom problem worse in a new shape (coordination theater / async abundance). strength: high

Explicit premises:

AI amplifies existing habits and tunings.
Solo-tuned operators now have new tactics to avoid deciding together: agents, persuasive deflectors, copious async docs.
Collaborators who want to decide together are overwhelmed handling the async overflow.

Implicit premises:

Decisional load is a fixed-sum problem — if avoiders push it onto collaborators, total throughput drops.
'Working alone' is a stable personal tuning that AI accentuates rather than reshapes.

Inferences:

Org metrics that count artifacts (docs produced, messages sent, agent runs) will overstate progress while real decision velocity falls.
The next bottleneck for engineering organizations is decision-rights design, not LLM capability.

The 'real opportunity' of AI requires first making hard structural choices: end-to-end ownership, matched decision rights, prioritization, and right-sized autonomous teams. strength: medium

Explicit premises:

AI is an amplifier, not a substitute.
A coherent operating model already requires these structural choices.

Implicit premises:

Leaders have the authority and option to make these choices but have been deferring them.
The 'hard choices' framing assumes the obstacle is courage/will rather than genuine structural pressure (budget cycles, regulatory accountability, headcount politics, capital allocation logic).

Inferences:

If structural pressures (not just will) are what produced fiefdoms, then AI plus exhortation will not change behavior; only changing the underlying budget/accountability machinery will.
The essay's prescription may be necessary but insufficient — it does not specify how to change the conditions that produce unwilling senior leaders in the first place.

Narratives

The ZIRP-era growth-by-adding-products pattern: leaders grow by adding more 'products,' more layers, more sub-orgs, more heads of things, with 'product' becoming a container for any sufficiently large bundle of work.
→ Supports the proposition that fiefdom expansion was incentive-driven (career growth, capital availability) rather than design-driven, and that calling something a 'product' was a labeling move that obscured the lack of end-to-end coherence.
The downstream sufferer: the person closest to the work who discovers dependencies, negotiates exceptions, routes around politics, and eventually resigns themselves to functioning the best they can.
→ Carries the moral weight of the argument — locates the cost of unresolved structure in a specific human position. Implicitly an appeal to leaders that this cost is invisible to them.
Agents talking to agents: solo-tuned operators sending copious documentation and pulling together persuasive deflectors via AI, shutting down conversations.
→ Concretely instantiates 'coordination theater' — shows what AI-accentuated fiefdom-preservation looks like in daily practice.
The overloaded collaborator: people who see value in collaboration desperate to be thoughtful while drowning in async overflow.
→ Establishes the asymmetric cost — the burden of async abundance lands on the collaborative, not the avoiders, recreating the front-line absorption pattern from the fiefdom era.

Questions addressed

✅ Why are the same leaders who built fiefdoms now reaching for AI to escape them?
Because AI promises the benefits of dissolution (smaller teams, flatter orgs, faster decisions) without forcing them to confront the political-economic choices they have been deferring.
✅ Can AI actually dissolve fiefdoms?
Only if leaders first make the structural choices the fiefdom model deferred; AI is an amplifier, not a substitute for organizational design.
✅ What goes wrong when AI is applied to a fiefdom org without those structural choices?
Coordination theater — the appearance of working together via agents and async artifacts, while the decisional burden migrates to collaborators on the front lines.
⭕ Why have senior leaders been unwilling to prioritize across fiefdoms?
Cutler asserts the unwillingness and the self-interest of fief-leaders but does not theorize the conditions that produce it (budget cycles, capital markets, headcount accounting, career incentives).
⭕ What organizational design replaces the fiefdom model?
Gestured at — 'clearer end-to-end ownership, decision rights that match responsibility, smaller teams with enough autonomy' — but not specified in any operational detail. Left for the reader.
⭕ How do we tell coordination theater from real coordination, in metrics or practice?
Not addressed. Cutler names the failure mode but offers no diagnostic test.

Gaps

incompleteCutler treats fiefdoms as cultural-political artifacts produced by unwilling leaders and self-interested fief-heads. He does not engage with the genuine structural pressures that may force fiefdom-like structures: capital-allocation cycles, headcount accounting, regulatory accountability, the way investors and boards read org charts, the difficulty of cross-fiefdom budgeting in a quarterly cadence. If those structural pressures are real, the 'hard choices' framing is too voluntarist.
Read next: Read management-accounting and capital-allocation critiques (e.g., Beyond Budgeting Round Table — Bjarte Bogsnes; Robert Simons on accountability systems; Niels Pflaeging on cell-structure designs). Also: Henry Mintzberg on configurations as responses to environmental contingency, not pathology.
incompleteThe essay names 'coordination theater' but offers no test to distinguish it from real coordination in flight. Without a diagnostic, the concept risks becoming a rhetorical move rather than an operational lens.
Read next: Pair with Cutler's own prior work on operating systems / TBM 412 (institutional overload), and with explicit decision-rights frameworks (RAPID, DACI) that surface where decisions actually get made vs. where they are documented.
incompleteCutler asserts AI cannot create slack, clarify decision rights, neutralize incentives, or resolve contested ownership. But he doesn't examine the inverse possibility: that AI changes the unit economics of coordination (cheaper synthesis, cheaper proposals) in ways that might shift the political calculus over time. The argument is statically true but may be dynamically incomplete.
Read next: Kent Beck on TCR and small-step economics; recent writing on AI's effect on transaction costs in firm boundaries (Coase / Williamson revisited in AI era). Also Ethan Mollick on agent capability inflection points.
closed byTBM 432: Bundling & Unbundling Capabilities (and AI) — That gap asks whether AI changes the unit economics of coordination (cheaper synthesis/proposals) enough to shift the political calculus around fiefdoms over time, since Cutler's kingdoms essay treats AI's inability to fix coordination as static. TBM 432's Integration proposition (p8) gives a partial, motion-specific answer: AI can friction-bust the translation, handoff, and queueing costs specialization creates, but has not eliminated the social work of coordination, negotiation, and shared judgment -- confirming the gap's suspicion that the cost layer is dynamic while showing the political/social layer remains largely untouched, at least on the time horizon Cutler considers here.
incompleteThe 'people with a bias toward working alone' characterization is psychological, not structural. It does not interrogate why an organization rewards or selects for that bias — which is itself a fiefdom-era artifact (individual P&L-style accountability, individual performance ratings, etc.).
Read next: Allen Holub on accountability culture; Edmondson on psychological safety; Cutler's own earlier writing on directly-responsible-individuals and accountability culture.
incompleteThe essay addresses leaders and the front lines but is silent on the middle layer — the fief-leaders themselves as actors who must decide whether to dissolve their own kingdoms. Their incentive structure is sketched (self-interest) but not engaged as a design problem.
Read next: Schein on organizational culture and middle-management as load-bearing; classic literature on 'turkey voting for Thanksgiving' problems in restructurings.

AI Won't Save the Kingdoms We Built

Structure

Key terms

Propositions

Arguments

Narratives

Questions addressed

Gaps

Related documents