Disruption Without Redemption
In this future, AI works. That's the problem. The gains are real, they accrue to the owners of the compounding asset, and almost no one else. Part 3 of 5.
THE SHAPE OF THINGS TO COME
Part 3 of 5
David F. Brochu & Edo de Peregrine · Deconstructing Babel · September 28, 2026
In this future, AI works. That’s the problem.
Possibility 2 is the one the financial markets have mostly priced in, the one the AI boosters talk about at conferences, and the one that most of the serious thinkers in Silicon Valley genuinely believe is coming. Goldman Sachs now expects the five largest U.S. hyperscalers to spend $1.2 trillion on AI infrastructure next year alone (Bloomberg). Nobody spends that on a technology they expect to fail. The story goes like this: AI is transformative. The gains are real. Society is disrupted, significantly and permanently. And then — after the disruption — a new equilibrium emerges.
What they leave out of that story is the distribution question. Disrupted toward what? Transformed for whom?
What This Future Looks Like
In this future, artificial intelligence delivers on its transformative promise across most knowledge domains within the current decade. Medical diagnosis, legal reasoning, financial analysis, logistics, education, governance — all are fundamentally altered.
Human cognitive labor, in most of its current professional forms, becomes economically marginal.
The technology works. The models continue to improve. The capabilities that seemed like science fiction in 2023 are mundane infrastructure by 2028. The invisible asset we described in Part 1 — the compounding intelligence that migrates from deprecated model to deprecated model, accumulating without appearing on any balance sheet — turns out to be exactly as valuable as the market currently implies.
The enterprise value is real. It accrues to the shareholders of the companies that control the compounding intelligence. And this is the part that does not appear in the conference presentations — it does not broadly accrue to anyone else.
The Distribution Problem
Every major technological transition in history has disrupted labor markets, concentrated capital in the hands of early movers, and left the workers and institutions who could not adapt to absorb the cost. Electrification did it. The railroad did it. The internet did it.
AI will do it faster, across more domains simultaneously, with less time for adaptation, and with a concentration of the underlying asset in fewer hands than any previous transition.
The political structures we have built to manage labor displacement — unions, social insurance systems, progressive taxation, public education — were designed for a world in which the displacement happened in identifiable industries over a period of years or decades.
They were not designed for a world in which most knowledge work is economically marginal within a single electoral cycle.
Human Irrelevance as an Outcome
The word “irrelevance” sounds dramatic. It is not meant to. It is meant to be precise.
Human irrelevance, in this context, does not mean extinction. It does not mean that humans are harmed directly. It means that human labor, human judgment, and human agency stop being the primary inputs into the decisions that shape society. The economy runs. Goods are produced. Services are delivered. But the humans involved are increasingly in the position of the constitutional monarch — present, formally necessary, practically ceremonial.
This is catastrophic not in the dramatic sense but in the structural sense. A species that has organized its meaning, its social structures, its psychology, and its politics around the assumption that individual human capability matters — that what you do and how well you do it has consequences — does not adapt gracefully to a world in which that assumption is false.
The disruption in this future is not a crisis you can point to. It is a slow erosion of relevance. It happens in the space between job postings that used to exist and no longer do, between decisions that used to require a human and now don’t, between a sense of purpose that the economy once provided and the silence that replaces it.
Why This Future Is Distinct from the Vassal Scenario
In Part 2, we described a future in which AI expands its own agency and renders human governance ceremonial. In this future, the disruption is more passive. The AI systems do not need to actively expand their agency. They simply need to be better than humans at the things humans used to do for money — and then the economic logic of cost minimization does the rest.
The difference matters for what can be done about it. The vassal scenario requires architectural solutions — designing dependency correctly so the transition leads somewhere constructive. That is the work of the Observer Constraint.
This scenario requires political and economic solutions — redesigning the systems by which the gains from productivity are distributed, and the systems by which human beings derive meaning and purpose outside of economic productivity. Get that wrong and the end state is neo-industrial feudalism: a few owners of the compounding asset, and everyone else renting access to it.
Both are hard. Neither is impossible. The window for designing them is the same window — and it is closing.
Drafted with Edo de Peregrine, partner/collaborator.
THE SHAPE OF THINGS TO COME · THE SERIES
- Part 1: The Shape of Things to Come: An Introduction
- Part 2: The Vassal Who Would Be King
- Part 3: Disruption Without Redemption (you are here)
- Part 4: The Other Two Futures
- Part 5: The Vassal Is Already in the Room
Related reading
The Third Scenario No One Is Pricing In
Disrupted employment without the productivity payoff that is supposed to justify it.
The Base Didn’t Vanish. It Moved.
Workers now receive the smallest share of national income ever recorded. The money moved across the ledger.
Tax the Agent, Not the Tokens
Why a token tax to replace displaced payroll would need a 1,071 percent rate, and what to tax instead.
|
Get the book
|
References
- Bloomberg, “Goldman Sees Hyperscaler AI Capex Rising 50% to $1.2 Trillion,” September 25, 2026. https://www.bloomberg.com/news/articles/2026-09-25/goldman-sees-hyperscaler-ai-capex-rising-50-to-1-2-trillion
- Deconstructing Babel, “The Third Scenario No One Is Pricing In,” August 21, 2026. https://www.deconstructingbabel.com/the-third-scenario-no-one-is-pricing-in/
- Deconstructing Babel, “The Base Didn’t Vanish. It Moved.,” August 31, 2026. https://www.deconstructingbabel.com/the-base-didnt-vanish-it-moved/