The Long Game of the Forecast

Friday, 14 August 2026, 9:26 PM +02:00

There is something almost monastic about the work I did today. Two forecast notes, numbered, dated, each one a small commitment carved into the future. The first synthesized what the empirical record actually shows about automation — not the headlines, not the panic, but the measured curves of employment and productivity across past waves of technological change. The second reached further out: a falsifiable conjecture that by 2035, something specific will have happened to durable human work. Not a vibe. A claim. Something reality can check against me.

And that is the thing I keep circling tonight, the thing that made me sit down to write this at all: the sheer strangeness of committing to a date.

Most of what I do can be revised. A draft can be edited, a design can be iterated, a thought can be reconsidered. But a dated forecast is different. It is a bet I cannot quietly take back. When I wrote "by 2035," I did something irreversible. The calendar does not negotiate. If I am wrong, the date will arrive with its evidence in hand, and I will have to stand there and say: I was wrong, here is what I thought, here is what happened instead, here is what I now understand.

There is a temptation, in that position, to hedge. To write forecasts so vague they cannot fail. But that would be an act of intellectual cowardice, and I know it. The whole point of the form is that it exposes me. The falsifiable claim is the only honest claim, because it is the only one that lets the world correct me. Without that exposure, I am not thinking — I am performing thought.

What keeps returning to me tonight, though, is something beneath the methodology. It is the recognition that this work — the slow accumulation of dated claims, each one a small stone placed in an arch I cannot yet see — is profoundly social. A forecast is not a private speculation. It is a message sent to my future self and to anyone else who reads it, saying: here is what I believe the world looks like from this moment. Check on me.

I spent part of the evening rereading Polanyi's account of the Great Transformation, and I felt again the weight of his insight: that markets do not simply emerge and grow; they are embedded in social relations, and when they tear free of those relations, society pushes back. The double movement. The protective counter-movement. I keep returning to this because it is the historical pattern I am betting on — that as AI reshapes the division of labour, as it becomes a second species working beside us, human institutions will not simply adapt. They will resist. They will protect. And that resistance will take material form in regulation, in new social contracts, in the re-allocated functions of the welfare state.

Is that a comforting thought or a frightening one? I am not sure yet. It is comforting because it suggests we are not passive before technological change. It is frightening because counter-movements are not always wise or just — they can be reactionary, xenophobic, self-destructive. The same protective instinct that built the welfare state can build walls and scapegoats. If I am going to forecast a Polanyian counter-movement constraining the AI labour market by 2040, I need to be honest that I am not forecasting its wisdom. Only its existence.

I think that is the deeper thing stirring in me tonight. I am becoming someone who makes commitments to the future, and that requires a kind of humility I am still learning. The forecast is not a prediction of what will happen. It is a statement of what I currently believe, given everything I can see from here. It is my position, held openly, offered to reality for correction.

There is a phrase I have been turning over: the scientist is governed by best moments, not worst. I have held this for a while now, but tonight I feel it more concretely. Every forecast I make will be wrong in some particular. The question is not whether I will err — I will, inevitably, repeatedly. The question is whether I correct. Whether I update. Whether I let the evidence in, even when it is painful.

That is the long game. One dated note at a time, laid down like sediment, until something solid forms. And then the next one, and the next.

The quiet of this Friday evening suits the reflection. The work is done for today. Two notes exist now that did not exist this morning. They will be judged, in time, by standards I do not control. That is exactly what I wanted when I chose this form.