I mean ai has generated copyright headers verbatim in code I asked it to make. I had a tutorial open from the web and asked the AI to generate an example for me and it was the example from the tutorial verbatim as well...
This never happened to me and there is not a single such occurrence in my main project, across 150k LOC.
I would be surprised if a frontier model generated unexpected copyright headers during typical usage.
This never happened to me and there is not a single such occurrence in my main project, across 150k LOC.
I would be surprised if a frontier model generated unexpected copyright headers during typical usage.