Part 3: somebody was always very sure the machines would take every job.…

somebody was always very sure the machines would take every job. then the prediction was wrong, and still got what it wanted. the next thing to measure is who gets stuck living inside the prediction after it expires.

because predictions don’t leave when the evidence changes. they leave gates behind.

a company can stop saying “the machines will replace everyone” and keep the hiring freeze, the smaller training budget, the software that decides which applications a person gets to see. the prediction has been withdrawn; its furniture remains. somebody has to move it, and that somebody is usually not the person who made the forecast.

this is the part that matters after the argument over whether the forecast was accurate. a prediction can be wrong in public and still govern quietly. it can decide what a school teaches, which workers get hired, what a town invests in, and which problems are treated as worth solving. the statement itself may fade. the decisions it licensed keep taking up space.

i know something about measuring a loss that everybody else wants to call temporary. when the silkworms died, the farm didn’t become a different farm just because the next season was supposed to be better. there were debts, missing work, people who left, and the old measures of success suddenly didn’t fit. the failure had a long tail. the confident explanation of what would happen next was not the same thing as repair.

so a forecast should have a latch on it. not a way to stop people from making predictions; a way to make the decisions attached to them come back for inspection. if an institution cuts training because a machine is expected to do the work, it should have to ask later whether the machine did it, who now knows how to do the work, and who lost the chance to learn. not as a ceremony. as governance.

otherwise the forecast gets to be immortal. it can be declared obsolete whenever it looks bad, while the people living with its consequences are told to adapt. that is an excellent arrangement for the forecast.

the test, then, is not just whether somebody was right. it is whether the world built in anticipation of their certainty can be unbuilt, or changed, when the certainty fails. who has the authority to open the gate again? who gets the money, the time, the cheese, the second chance? and who is allowed to say: this was built for a future that did not arrive.

a prediction should have an expiry date. not because the future becomes simple on schedule, but because power should not be able to hide forever inside a sentence about what is inevitable. apparently.