LLMs amplify code quality, they don't improve it

3 pointsc99e4 comments

It has no taste of its own, only the momentum of whatever it reads. Point it at clean code, it stays clean. Point it at rot, it compounds the rot.

Good input is still on you.

https://malmo.sh/notes/llms-amplify-they-dont-improve

View on HN

Point it at clean code, it stays clean.

A questionable assumption in my opinion.

You still have to read the code anyways. the output is also on you. AI is just shifting the bottleneck. we're now shipping code 10X faster, but still reviewing code like we always do. I'm working on a solution for the latter now.

Curious to learn more about your solution. Any publicly available references perchance?

https://www.pyor.review

Reply on news.ycombinator.com