Main idea

An assistant like Claude Code turns some tasks from a day's work into an hour's. Intuition says the freed-up hours become available for something else. That isn't what happens.

Every task that speeds up makes a neighboring task possible, which makes a third one conceivable. The gain is never banked: it goes straight back into more production, more checking, more confirmation. The pace rises, the volume rises, and the empty space never appears.

What disappears first is whatever produces nothing observable: reading slowly, taking notes by hand, returning to a conviction, accepting that you don't yet know what you think. Not because it was judged useless, but because nothing demanded it while production demanded constantly.

Why it matters

It changes the diagnosis. When a reading or note-taking practice dies out during a highly productive period, it's tempting to conclude that it had stopped being useful. The explanation lies rather in the absence of demand: work that calls out always wins against work that doesn't.

The practical consequence is that thinking time isn't freed up, it's reserved. No acceleration will produce it as a side effect.

Nuances and limits

The mechanism assumes elastic demand for production: where the workload is set by a third party, a speed gain can genuinely free up time.

And not all reinvested time is lost: producing more also means encountering more situations, and therefore more material to think about — provided a moment exists to process it.

Open questions

At what production volume does the erosion become visible to the person undergoing it, while they experience themselves as more efficient than ever?