Disclaimer: The views expressed here are my own and do not represent any employer, past or present. External links are provided for convenience; I’m not responsible for their content or subsequent changes.
When technology increases efficiency, you should be able to do more with less.
More efficient machines, working better, to produce the same level of output.
So maybe you can cut back on the number of total machines producing the thing?
Maybe you can cut back on the total number of people producing the thing.
Anyone, anyone, Bueller?
Just a short while ago it seemed completely intuitive, obvious even, that programmers would be the first to suffer the dreaded AI job apocalypse.
In an ironic twist of fate the programmers built the thing that would replace them.
So- this is odd.
My programmers are busier than they’ve ever been.
A conundrum. A paradox, even?
19th century economist William Stanley Jevons noticed that coal engines were becoming more efficient.
More efficient machines and demand remains constant, less “things” are needed to produce the product
This did not happen.
Demand for coal as an input and coal-powered applications exploded.
Changes in efficiency changes the economics of the activity, which changes behavior.
Hmm- this is pretty good.
In the programming/AI case, the resource becoming more efficient is less obvious than the coal.
AI coding agents, as a technology, broadly increases the efficiency of spinning code into things.
Code into prototypes. Code into automation. Code into products.
(Oprah voice) “You get a prototype! You get an automation! You get a product!”
Everybody wants one, everybody get one, and demand becomes non linear with the changes in behavior.
Even with an assist from Claude Code or Codex, my programmers (AI engineers), have no way to meet the demand of the last months.
I expect this demand to continue.
I was wrong. I thought we would be able to do more with the same/less.
I fell for it.
I can hear Jevons and his big beautiful beard laughing at me now.
We are hiring more programmers.
Build bigger more efficient highways and you get more traffic.
Having capacity makes more possible which increases demand.
The best programmers I ever got to work with were english and philosophy majors.
You might ask why? They were the best at breaking down complex real world business problems into simple declarative sentences and decision trees (requirements and tests).
We still need programmers to do this work for the machines. Each business problem is so far from the mean that predictively solving them is a non trivial problem that requires human insight, intervention and oversight.
So the more programming capacity we have, the more demand there is for creating great requirements that represent a real world model of the business problem, not a simple reductive model.
Hence, the really good developers get busier and thier role becomes more essential.
Maybe?