Post #888698

all anthropic employees should be banned by default from all open source software projects, in light of the fact that apparently claude code has an "undercover mode" for their internal use, to be used when contributing to public repositories, that instructs the model not to indicate that it is an ai model or attribute commits to itself and instead pretend to be a normal human (and also not leak internal codenames) current US copyright guidance requires that ai-generated output is adequately indicated as such in order to not jeopardize copyrightability of potentially the entire work (such as, for example, an open source codebase)