LLMs are just the ultimate IP-violation-machines. I love using them for reverse engineering. I picked up so many projects I abandoned in the past because staring at obfuscated code or assembly got boring and tiresome and felt like an endless endeavor that will never be realistically reverse engineered completely anway.

Now I give Opus or whatever other model enough context, datasheets and tests and it starts reversing. Does it hallucinate and is not always correct? Yeah. But who cares? I am not always correct and misunderstand things when manually reversing stuff as well. We all do.

Gradually building more and more context to be able to reason a bit easier about things you didn't understand yet is exactly what an LLM can incrementally help you with. And gathering more and more information helps both me and the LLM to understand the stuff we are looking at a bit better.

It's so awesome.

I'd strongly advise against trying to illegally occupy foreign soil in this day and age. Very unhealthy.

https://www.youtube.com/watch?v=vZCbycrMczo