Post

And I know people hate being told that, but I'll take the punches: I believe it would be genuinely unethical for me to spend so much time researching this and then NOT share what I've learned

use cases/problem domains/results vary, obviously. If I was trying to write web front ends in react my life would considerably easier with many gen ai tools, but in the plumbing of some types of code the subtle bugs can blunt many efficiency gains.

0 reposts0 likes

That's true: LLMs are more helpful for some types of code than others, and there are definite risks of subtle bugs sneaking through That's what I mean by the LLM learning curve though: part of that is learning what they're good at, what they're bad at and processes you can use to reduce that risk