Hacker News Reader

The rate at which AI is capable of producing code is intractable for humans to deal with. Right now the bottleneck is human reviewers. If AI ever becomes effective at generating provably correct code, it's joever.

View on HN

The good news is there are several thousand newly unemployed software engineers on the market…

It's also likely that there are thousands who aren't hustle-culture oriented who have been looking for so long that they don't even count as unemployed anymore.

Or maybe I'm just projecting.

But will humans understand the proofs?

Oh…

The proofs are not meant for human consumption. It's for the AI to know to try again rather than spit out hallucinations. Of course there's a leap of faith somewhere here.

The proof either passes the SAT solver in a reasonable amount of time, or it doesn't.

Somebody will have to write proof verifier, and that in many ways will be harder than writing some CRUD app that they want proof verifier to validate.

We might even end up increasing the demand and pay for devs if this happens to pass.

We have many of those that are perfectly fine. Writing proofs is still quite hard, especially proofs that actually say something about your program.

Proving something is correct is a far harder exercise, than writing a broken but acceptable version of that thing.

Reply on news.ycombinator.com