cross-posted from: https://discuss.tchncs.de/post/21298994
I’m trying to feel more comfortable using random GitHub projects, basically.
Huh. That’s actually kind’ve a clever use case. I hadn’t considered that. I presume the main obstacle would be the token limit of whatever LLM that one is using (presuming that it was an LLM that was used). Analyzing an entire codebase, ofc, depending on the project, would likely require an enormous amount of tokens that an LLM wouldn’t be able to handle, or it would just be prohibitively expensive. To be clear, that’s not to say that I know that such an LLM doesn’t exist — one very well could — but if one doesn’t, then that would be rationale that i would currently stand behind.
I understand, but I wouldn’t be surprised to see some solution out there that could maybe feed the AI chunks of code without context… It may still be able to detect “hey you told me this software is supposed to do X and here it seems to be doing Y”.
I guess we’ll have to wait a couple of years for these tools to be accessible and affordable.
You would first need to define malicious code within the context of that repo. To some people, telemetry is malicious.
Probably not. Obfuscation works, and might even depend on remote code being downloaded at either build time or run time.
There are a lot of heuristics you can use (e.g. disallowing some functions/modules) to check a codebase, but those already exist no AI required. Unless you call static analysis “AI”, who knows.
But an AI can “realise” the code might be downloading something it doesn’t need to. That’s the point.
AI is “smart” and understands that you told it that the library was supposed to do something specific, and it can understand that and look for things that seem not correlated to the purpose of the repo.
If you’re one of those people that think every product is better if there’s “AI” on the box then sure. What you’re describing is static analysis though, it is not new.
