As the cost of something goes down, demand for it goes up. Total spending can end up higher than before, because projects grow in scale, in complexity and in number.
The idea comes from William Stanley Jevons. In The Coal Question (1865) he pointed out that more efficient steam engines did not cut Britain’s coal use. They made coal power cheap enough to use in far more places, so total consumption went up.
I’ve been thinking about it in the context of this industry and how AI relates to it. Of the courses I’m taking this semester, software engineering is the one it lines up with most closely. If AI makes code cheaper to produce, Jevons would predict that we end up wanting a lot more code.
One of the concerns is that more demand for code does not have to mean more demand for coders. That makes it look closer to horses and automobiles. Demand for transport exploded once cars arrived, and horses were not the ones who benefited.
There is another side to it. A model can generate code faster than a team can review it. Code accepted without enough review turns into maintenance problems and technical debt that cost more down the line. So the bottleneck may shift from writing code to reviewing it. If that holds, the scarce skill becomes reading, judging and maintaining code. Demand for that kind of work could still follow Jevons even if demand for typing code out does not.
Thoughts?
Originally posted on my software engineering course’s discussion board. Tidied up and expanded a little here.