Hacker News Reader

This seems to be quite contrarian considering we had this on the front page of HN the other day: "Debian votes to allow "responsible use of generative AI".

I guess this LLM coding wasn't "Responsible" enough. hahaha

Let the AI bubble pop baby

View on HN

Omarchy is all in on AI, if you look at the recent commits and the dev workflows they have set up you can easily tell no human is looking at all the stuff they are merging.

It's not the same thing as allowing some AI contributions under strict guidelines.

Even worse. They don't even have AI review them. I fed the commits that introduced the problem to a few frontier models and they saw several problems, including the aforementioned security problem. Even Mistral saw it. I did have to instruct all models to look for security problems, though, but still.

It's not that we shouldn't use vibecoded distros. It's that we shouldn't use badly vibecoded distros with shitty or non-existent processes.

On Lex Fridman recently DHH was enthusiastically bragging about how he was letting AI generate C++ that he intentionally wasn't even looking at, he was treating it as a pure black box and just submitting the output.

Reply on news.ycombinator.com