[bartread]

This is the kind of situation where feeding your code through an LLM can actually be helpful: they're really good at spotting the kind of errors/typos like this that have a profound impact but which our eyes tend to all to easily scan over/past.

[nancyminusone]

The last time I tried an LLM on assembly, it made up instructions that didn't exist.

[cdelsolar]

cool; nowadays LLMs are better

[iforgotpassword]

Today I used chatgpt for winapi stuff - it made up structs and enums regarding display config. So not too convinced it'll be any good with 6502 asm.

[nextaccountic]

It's funny because some time ago (months? years?) people would say that you just didn't prompt the LLM well enough. But now LLMs are better and prompting isn't as arcane as before, so the next frontier is giving them the proper context. See this HN thread currently in the front page

https://news.ycombinator.com/item?id=44427757

[voidUpdate]

You also have to be using the exact right model to get reasonable results, which is always the one you have to pay for, not the free one, and also not the one you were using

[recursive]

cool; but not better enough

[bogantech]

Yeah I've been using Claude to review a bunch of m68k asm that I've been working on and it's been helpful at catching silly mistakes like using a direct address instead of an immediate value, clobbering registers, incorrect branches etc.

Of course if you just blindly ask it to write asm it will occasionally invent new instructions or address modes but it's very good at reviewing and making adjustments