The Second Decoupling

Start with a specific, quiet thing that is happening right now, because it tells the whole story before any theory does.

For decades, the deal at the bottom of the white-collar world worked like this. You graduated, you got a junior job — paralegal, junior analyst, associate developer, assistant, the entry-level seat — and that job was bad on purpose. It was the grunt work: the document review, the data cleanup, the first draft nobody respected, the spreadsheet nobody wanted to build. You did it badly, then less badly, then well, and in the doing you learned the trade. The bottom rung wasn’t supposed to be glamorous. It was supposed to be a rung. You stood on it to reach the next one.

That rung is quietly disappearing.

The grunt work that used to train the junior was the most automatable work in the building. It was routine, it was structured, it was exactly the kind of thing the new tools do in seconds. So firms are doing the rational, short-term thing: they’re handing the grunt work to the machine and not hiring the junior. The numbers are no longer anecdotes. Researchers at Stanford found that workers between twenty-two and twenty-five in the most AI-exposed jobs have seen a measurable decline in employment since generative AI spread — while older, more experienced workers in the very same occupations have not. Entry-level postings in fields like software and data analysis are down sharply from a few years ago, in some slices by more than half. The damage isn’t spread evenly across the workforce. It’s concentrated, with eerie precision, on the first rung.

Here’s the cruelty that makes this different from a normal downturn: you cannot climb a ladder you can’t get onto. The whole system assumed people would start at the bottom and learn their way up. If the bottom is gone, the question isn’t just “what happens to this year’s graduates.” It’s “where does the next generation of experienced people come from, if no one is allowed to be inexperienced first?”

That is the small, sharp edge of a very large thing. And the large thing is the second subject of this newsletter — the shift I’ve named twice and never sat down with properly. Money is changing shape; we spent five letters on that. But work is changing value too, and this is where it starts: with a rung that isn’t there anymore.

Every generation thinks the machines are coming for the jobs, and every generation, so far, has been wrong in the way that matters. The hand-loom weavers, the farm hands, the switchboard operators, the typesetters — each was displaced by a machine, and each time the doomsayers were proven wrong over the long run. Not because the displacement wasn’t real. It was brutally real for the people living it. But because the economy kept doing the same trick, over and over, for two hundred years.

The trick was this: machines took the work that was routine and physical, and humans moved up — into work that required more judgment, more dexterity, more thought. The plow replaced the digging, and the digger’s grandchildren did something cognitively richer. The assembly robot replaced the welder’s hands, and the welder’s children worked with their heads instead. Every wave of automation pushed humans up the ladder of skill, toward the thing machines couldn’t do: think.

That is the assumption underneath two centuries of “don’t worry, automation always creates more jobs than it destroys.” And it has been true. But read the sentence carefully and you’ll see the trapdoor in it: it was true because there was always an “up” to move to. The whole optimistic story depends on a higher rung existing — a more cognitive, more human kind of work that the machine couldn’t reach, waiting for the displaced worker to climb into.

So here is the honest question this letter is actually about, the one I can’t answer and neither can anyone else: when the machine automates cognition itself — the thinking, the analyzing, the writing, the drafting, the very rung we’ve spent two hundred years climbing toward — where is “up”?

Maybe there’s still an up. Maybe it’s judgment, taste, responsibility, the human-to-human work of care and trust and persuasion — rungs the machine still can’t reach. I think that’s likely, and I’ll come back to it. But notice that this is the first time in the entire history of automation where we have to ask the question seriously, because it’s the first time the machine has come for the high ground instead of the low. Every prior wave automated the work we wanted to leave behind. This one automates the work we were climbing toward. That is not the usual panic. That is a genuinely new shape of problem, and pretending it’s just the loom again is its own kind of denial.

Longtime readers know we’ve been here before, one shift over.

Three letters back, in the 1971 story, we watched the first great decoupling. For two centuries, when workers produced more per hour, they got paid more — productivity and wages rose together, locked in step, the engine of the entire middle class. Then around 1971 the two lines split. Productivity kept climbing. Wages flattened. The value workers produced kept rising; the share of it they took home did not. That was the first decoupling: the link between how much you produce and how much you earn came apart.

What’s happening now is the second decoupling, and it’s deeper. The first one severed the worker’s output from the worker’s pay. This one threatens to sever human labor itself from economic value — to make it possible, for the first time, to produce enormous economic output with very little human work at all.

Think about what that does to the oldest assumption in economics. For all of history, to make more stuff you needed more labor, or more productive labor — either way, humans were in the loop, and that necessity is what gave human work its value and its bargaining power. You had to be paid, because the work couldn’t happen without you. The second decoupling is the slow removal of that “had to.” When capital — the models, the compute, the systems — can do more and more of the cognitive work that used to require a person, the person stops being necessary to the output. And value that doesn’t need you is value that doesn’t have to pay you.

That is the mechanism under the missing rung. The junior didn’t lose the job because times are hard. The junior lost the job because, for the specific routine-cognitive work the junior used to do, human labor and economic output came unbundled. The work still happens. It just happens without the human. Multiply that across every category of routine knowledge work and you have the shape of the shift: not a sudden robot apocalypse, but a steady, category-by-category unbundling of human effort from economic production.

This is the letter where the honesty section matters most, because the doom version of this story is easy to write and mostly wrong, and I refuse to write it.

Where I’m confident:

Where it genuinely cuts the other way:

If the optimists are right, this letter is a chronicle of a hard transition that ended well, like the looms. If the pessimists are right, it’s the early description of something genuinely new. I don’t know which. What I know is that “it’s obviously fine” and “it’s obviously catastrophe” are both postures, not analysis — and the people who’ll navigate this best are the ones holding the uncertainty honestly instead of resolving it prematurely to feel better.

The point of naming a shift is never to frighten you with it. It’s to let you position before it’s obvious. And the second decoupling, even held as an open question, already points at a few calm conclusions — the same ones the playbook from last week was quietly built around.

If human labor and economic value are slowly unbundling, then the side of that equation you want to be on is the side that owns the output, not only the side that performs it. That’s not a moral claim, it’s a structural one: when value flows increasingly to whoever owns the productive capital — the models, the systems, the equity in the companies building them — then owning a small slice of that capital is how an ordinary person stays attached to the value, even as the labor link frays. Last week’s “Upside” layer wasn’t really about Bitcoin. It was about ownership, in an age when ownership is becoming the thing that pays.

And on the labor side, the move isn’t to outrun the machine — you can’t — but to do the work that sits above it: directing it, judging its output, taking responsibility for the call, doing the human-to-human work it can’t. The most valuable person in the new arrangement isn’t the one who does what the machine does, slightly worse. It’s the one who does what the machine can’t, and uses the machine for everything else. That’s not a comforting platitude; it’s the actual shape of where the higher rung might be.

We’ll spend the next two letters there: where the value actually flows now that labor and output are unbundling, and what it means to become an owner rather than only an earner in an economy that’s quietly changing who it pays. The missing rung is the diagnosis. The next letters are the response.

For six letters the three types of people were a frame for the money shift. Here they are again, wearing the work shift’s clothes.

The Blind assume the ladder still works the way it did when they got on it — that the rung will be there for their kids because it was there for them, that “get the degree, get the job” is still the whole map. They’re not stupid. They’re using a map of a country that’s being redrawn.

The Scared read a letter like this and spiral — conclude the machines win, that effort is pointless, that there’s no move to make. That’s just doom wearing the mask of realism, and it’s as paralyzing as denial.

The Prepared do the harder, calmer thing: they take the shift seriously without taking it as fate. They assume the rung is moving and position to stand somewhere higher — building skills that sit above the machine, owning a slice of the capital the value is flowing to, staying fluent enough in the new tools to direct them rather than be replaced by them. They don’t know if it’s the loom again or something new. They build for both.

This opens the second movement. We spent five letters on money changing shape, because it was the clearest of the three forces. This is the first real letter on the second force — work changing value — and it begins where you can already feel it: the first rung of the ladder, quietly going missing.

Next Sunday: where the value actually goes when labor and output come unbundled — who captures it, and why the answer rewrites the old advice about how to build a life.

If this letter was useful, three things:

One — send it to someone early in their career, or someone raising a kid who’s about to enter it. Not to scare them. To let them plan with the real map instead of the old one.

Two — notice your own reaction. If it was “this is doom,” or if it was “this is hype, nothing changes” — both of those are the easy exits. The harder, more honest place is the uncertainty in the middle, and that’s where the useful thinking happens.

Three — if you think I’ve got the diagnosis wrong, reply and tell me. The work shift is newer and less settled than the money shift, which means I’m more likely to be wrong about it — and the readers who push back have made every letter sharper.

Understand. Position. Don’t panic — especially on this one, where panic is the least useful response of all.

— Bill2Billion

P.S. Nothing here is financial or career advice for your specific situation — it’s one writer thinking out loud about a shift that’s still unfolding and genuinely contested. I’ve tried to represent the uncertainty honestly, including the real possibility that the optimists are right and this turns out like every prior wave of automation: disruptive, then fine. The data on entry-level hiring is real but still being argued over by serious people, and reasonable readers will weigh it differently. As always: for decisions that affect your actual livelihood, think for yourself, and talk to people who can see your specific situation. I’m here to widen the lens, not to make the call for you.