Who needs static analysis when you can put your trust in a magic robot?
(This is one thing that baffles me about the “let’s use LLMs to code” movement; a lot of the proponents don’t seem to be just adding it as a tool (I don’t think it’s a terribly _useful_ tool, but whatever, tastes differ), but using it as the only tool, discarding 50 years worth of progress.)
(This is one thing that baffles me about the “let’s use LLMs to code” movement; a lot of the proponents don’t seem to be just adding it as a tool (I don’t think it’s a terribly _useful_ tool, but whatever, tastes differ), but using it as the only tool, discarding 50 years worth of progress.)