19 ms·
If current AI can barely do maths, decompilation is not something I'd expect it to do well. It will of course try and come up with something plausible, but ofte
by userbinator 3y ago
If current AI can barely do maths, decompilation is not something I'd expect it to do well. It will of course try and come up with something plausible, but often subtly wrong.
- odyssey7 3y agoWould decompilation be closer to arithmetic or translation?
- nanaian 3y agoDepends on your goal. If it's matching decompilation, probably the former. There's been research into the latter but its in early stages. https://github.com/nforest/awesome-decompilation?tab=readme-ov-file#ai-based-decompilation https://github.com/nforest/awesome-decompilation?tab=readme-... decomp.me gives us a large database of C(++) <-> target asm to train a model on ;)
- kevingadd 3y agoIf you want the decompiled code to produce a 1:1 match with the original binary (even if it takes some finessing by hand at the end) you need something rigorous approaching arithmetic. A fuzzy decompiler that just approximates the intent of the original code won't get you there (and this is mostly what you get out of GPT for many tasks), but it could still be useful for something.
- wegfawefgawefg 3y agogpt4 can spit out accurate unoptimized AST of javascript and python. (I just tried it.) Now to test emitting.