Post

i think im more careful about typos in my prompting than i am in my slack messages

interesting. why?

i have a bad theory of mind for LLMs and treat their attention as limited in ways similar to that of a human

also im somewhat partial to the argument that clean, typo-free prose overlaps in latent space with the set of all good engineering practices

i like that argument i'm going to adopt it

I might be imagining it but I think sprinkling in some LaTeX when referring to specific variables gives a slightly more precise answer

Yeah, this is a noticeably strong effect, they seem to be better at math overall if you get them to write it out in LaTeX instead of unicode or plaintext approximations

i always start my AGENTS md with something like “we are colleagues and collaborators. we’re pals.”

The why / framing matters more then the telling. They respond well of you give a reasonable reason, like we are pre-release, with no users, you don't need compatibility, versioning right now

i have a boilerplate README that i started using which says "this project is brand new, has no users, does not need backwards compatibility layers" etc etc to get claude to actually start cutting into load-bearing walls without hedging

My bile refactoring skill is the only thing that has ever gotten claude to really think big about reshaping my code bases in a substantial way - I love to run it after I have gotten a project roughly in the direction I want to really improve the underlying design and architecture of the code