Theoretically a PM could say "the code is disposable and obsoleted by the next deployment. let's just document our prompts."

I don't know if that's a good idea but a lot of people are going to try it.

This is an obviously terrible idea that will lead to regressions on every deployment.

How long should we wait after promotions before prompting to add logging back to the service?

Who needs logging when we can just ask the LLM to confabulate the logs?

I'm afraid you're going to have to spell it out for the kids in the back who aren't paying attention.

It's like NixOS but instead of the much-maligned Nix language, you can use the English language!

What could go wrong?

You mean like COBOL ?

Wait shit brb, getting millions in VC funding.

The prompts are the source code in that case.

The LLM is some sort of transpiler.

Somebody will have to add some syntax to these LLM prompting systems to include text that doesn’t get converted. So the next round of PMs can ask you to documents your prompts).

> As you say, by the time you specify everything, you've written the code

Sadly not when it’s a 2000 page word document with a million formatting quirks and so many copy-paste versions you don’t know if you should trust “reqs_new25”, “reqs_new25v1” or the x number of other copies floating around.

Someone should come up with a tex2fortran conversion LLM.