Jack Dorsey just cut half of Block's workforce and was unusually honest about why. Not market conditions. Not a strategic pivot. The tools they were building were making large teams structurally unnecessary. Over four thousand people, not because the business was failing, but because it was working.
Most of the commentary landed in familiar territory. AI is coming for jobs. History says it works out. Adapt or be left behind. All true, as far as it goes. But that framing has a convenient blind spot: it treats the labour disruption in isolation, as if it's a technology story with an eventually happy ending, rather than one signal in a much larger convergence that is happening simultaneously across several systems at once.
Here's the fuller picture.
Technology is deflationary. It has always been deflationary. Jeff Booth mapped this clearly in The Price of Tomorrow: every meaningful wave of technological progress drives down the cost of production, makes goods and services cheaper, and generates abundance. That's not a side effect. That's the point. The loom, the engine, the PC, the internet, each one made things cheaper at scale, and each one displaced the workers whose jobs it replaced.
The problem isn't the deflation. The problem is the system sitting on top of it.
The fiat monetary architecture, and particularly its post-1971 form, doesn't just tolerate inflation, it requires it. Debt-based economies need nominal growth. Falling prices make debt burdens heavier in real terms. So every time technology tries to make things cheaper and more abundant, central banks print enough money to cancel the signal, and then some. The monetary base of the dollar has expanded by roughly 10,000 percent since Nixon closed the gold window. The institutional response to deflationary technology has been, consistently, to print the deflation away.
That worked, after a fashion, when the deflationary pressure was manageable. It is not manageable now.
AI isn't making workers more productive in the way the PC did. The PC gave a person leverage. AI, at scale, replaces the person. The distinction is not subtle.
The METR time horizon data, published in January 2026, makes this empirical rather than theoretical. Their longitudinal study measured how long a task can be before a frontier AI agent fails at it. In late 2022, when ChatGPT launched, the answer was roughly thirty seconds. By early 2026, frontier agents can autonomously complete tasks that take human professionals over fourteen hours. That time horizon has been doubling every seven months, accelerating to every four months in the most recent period. The extrapolation, if the trend holds, puts full work-day autonomous completion in 2027, full work-week in 2028.
That's not AI helping with your job. That's AI doing your job.
The collision Booth identified, deflationary technology against an inflationary monetary system, is now arriving at a scale the money printer cannot neutralise. And the institutional response is entirely predictable: stimulus, transition payments, eventually some version of universal basic income, all funded by further currency debasement. The people displaced by AI automation will receive more nominal currency and find it buys roughly what it bought before, or less. The number goes up. The purchasing power follows at a lag, if at all.
This is the Cantillon effect in its contemporary form. New money appears first at the point of creation, moves through financial institutions and asset markets before it reaches wages and consumer prices. The people who own assets, which are repriced upward by monetary expansion, are protected. The people who hold their savings in currency, predominantly the working and middle classes, pay the continuous and involuntary tax of debasement. The AI disruption and the monetary debasement are not separate problems landing on different people. They are the same problem, hitting the same people, at the same time.
The incentive structures protecting outdated ways of working aren't accidental. They're downstream of the monetary system itself. Inflation erodes savings, so you spend rather than accumulate. Debt fuels apparent growth, so organisations borrow and expand headcount. Headcount signals organisational importance, so managers hire rather than optimise. The entire architecture of the modern organisation was built inside an incentive structure that rewarded accumulation and punished efficiency.
The people defending 2003 workflows aren't just being stubborn. They're responding rationally to a system that has consistently punished the people who made themselves redundant by making their teams too lean.
Dorsey blowing through that logic at Block is notable precisely because it's rare. Most organisations will wait until they're forced. And some won't survive the wait.
The people who will navigate this transition well are not necessarily the most technical. They're the ones who understand what these tools can do well enough to direct them, question them, and combine them with the judgment and relationship work that remains genuinely human. That's a real position of leverage, and right now most organisations are chronically short of people who occupy it.
But the individual adaptation is only half the answer. The structural question, what kind of monetary system makes sense when technology is compressing costs at this speed, is one that the current architecture is not equipped to answer honestly. The fiat system was designed for a world of information scarcity, in which institutions could maintain credibility through control of the narrative. That information environment is gone. The same technological acceleration making large white-collar teams unnecessary is also making it impossible to hide what's actually happening to the purchasing power of wages.
Booth's conclusion, and it's the right one regardless of whether you share all his prescriptions, is that fighting the deflationary reality with monetary expansion is not a solution. It's a delay mechanism that concentrates the eventual pain while distributing the immediate benefits upward through the Cantillon gradient. The technology wants to make things cheaper and more abundant. The monetary system keeps trying to stop it. One of these forces will eventually win, and the one backed by mathematics has a better track record than the one backed by institutional credibility.
Dorsey's letter to Block employees deserves to be read as more than a corporate announcement. It's a ground-level data point in a much larger structural shift.
The tools we're building are changing what it means to run a company. That's true. But what it means to save, to work, and to hold value across time is also changing, and those questions are all part of the same convergence.
The machine panic is as old as the machine. The difference this time is that it's arriving simultaneously with a monetary system under structural stress and an institutional order exhibiting the classic signals of a Tainter complexity crisis. Those aren't three separate stories. They're one story, and we're somewhere in the middle of it.
Pay attention to what you know, to what you can build, and to what your savings are actually held in.
A note on sources: METR time horizon data is drawn from their January 2026 longitudinal study. Jeff Booth's deflationary framework is set out in The Price of Tomorrow (2020). The Cantillon effect is well-established economics, not contested. The 10,000% dollar monetary base figure is approximately accurate from 1971 to present, exact figures vary slightly by measure used. The Tainter complexity framework is developed across The Collapse of Complex Societies (1988) and is explored in depth in The Convergence extended analysis.
