Comment by Verdex
The comment that started this whole thread off mentioned LLMs oneshot-ing parsers. I didn't think an LLM could one shot a parser and I am interested in parsers which is why I asked about more info.
It's not a goal of mine but because of interests in parsing I wanted to know if this was something that was happening or if it was hyperbole.
By one-shot in my original post, I mean that it comes up with a decent, working implementation that I can then refine.
As opposed to getting there through incremental revisions where I must dictate the directions the LLM takes -- I don't necessary know the directions it should take because I might be agnostic about it.
I find success in breaking the problem down into tokenize and parse just like I do when writing parsers myself.
And definitely get it to generate tests every step of the way. The insane part is when the LLM can run tests itself and can iterate on code until all tests pass.Or you extend it with a feature just by writing some failing tests and it does the rest.