4 ms·
I wonder if it'll be much easier going forward. Take a suspect piece of code, run it through an LLM to explain it, pass the explanation to another LLM to genera
by politician 4y ago
I wonder if it'll be much easier going forward. Take a suspect piece of code, run it through an LLM to explain it, pass the explanation to another LLM to generate code from the explanation. An automated pipeline for clean room reverse engineering.
- jedberg 4y agoAs long as you can prove that you had a license to all the code that was used to train the LLM! :)
- akiselev 4y agoSomeone would first have to prove you used an LLM, no?
- dilyevsky 4y agoThat’s what discovery and depositions are for
- politician 4y agoHah! It appears that no one training these models is doing that. GPT doesn't generate the correct GPL/MIT/BSD/Apache license information and attribution for the source code it generates from its training set. All of these code-based LLMs are willfully violating copyright already. It's doubtful this goes anywhere in the courts, however.
- hencq 4y agoIANAL but it seems that if you rely on looking at the suspect piece of code to see what it's doing (even if done by an LLM), it's not exactly clean room.
- politician 4y agoIt's a two-step process. https://en.wikipedia.org/wiki/Clean_room_design https://en.wikipedia.org/wiki/Clean_room_design