I disagree. Much of what makes software complex for us, makes it complex for LLM just as well. E.g:
- a very large codebase
- a codebase which is not modularized into cohesive parts
- niche languages or frameworks
- overly 'clever' code
- a very large codebase
- a codebase which is not modularized into cohesive parts
- niche languages or frameworks
- overly 'clever' code
You need to check out how Claude uses Ghidra MCP or even tell it to use radare2 to disassemble even proprietary hardware ROMs.
We don't even come close to what LLM can understand in just a few minutes.
I regularly run it on large codebases because I'm not able to grasp it in any reasonable timeline.