← sources

Does Jevons paradox really apply to most white-collar jobs?

source: reddit (reddit:singularity)published 2026-10-08 22:30ingested 2026-10-09 02:37

retained text

Jevons paradox is frequently brought up as a counterargument against an AI-driven job apocalypse. The idea, which is probably well known by now to people who follow AI+jobs news is that increased productivity lowers costs, which increases demand and ultimately creates more jobs.  And it is very interesting because I would argue that while a lot of anti-AI arguments are bad and at worse put out in bad faith (due to coping), I think this is a legitimate economic argument that many people genuinely believe in and as such should be taken seriously. 

But how broadly does it actually apply?  The problem is I have no idea because I only know my job industry well and cannot claim to have any insights on the rest of the white collar jobs.  

So I asked ChatGPT to examine different white-collar occupations, and its rough estimate was that only around 20–25% of the roughly 100 million US white-collar jobs might experience substantial demand expansion. Obviously, this isn't a rigorous study, but the reasoning is interesting.

Consider a few examples.

**Tax preparation:** If AI lets me file my taxes in two seconds, that doesn't mean I'll suddenly file hundreds of returns per year. I'll still file one.

**Real estate:** If buying a house becomes as easy as clicking a button, that doesn't mean I'll suddenly buy 20 houses a year.

**Legal services:** Cheaper legal work might increase demand somewhat, but people aren't going to need 100 times more divorces or lawsuits.

Once you go down the list, you realize that many white-collar occupations serve needs that simply don't scale proportionally with productivity.

Then there's the issue of **physical bottlenecks**. AI might accelerate scientific discovery or engineering design 1,000-fold, but synthesizing materials, conducting experiments, and constructing buildings are still constrained by physical processes that don't improve at nearly the same rate.

And even where demand does increase substantially, there's another problem: If AI makes workers 100 times more productive but demand only increases tenfold, you still need 90% fewer workers, assuming the productivity gain applies to the entire job.

So while Jevons paradox might apply strongly to certain digital industries, I'm skeptical that it can save employment across most white-collar occupations.

document fetched 2026-10-09 00:32 · canonical https://reddit.com/r/singularity/comments/1x14t7t/does_jevons_paradox_really_apply_to_most