disappointing to see people run to copyright and throw the idea of fair use under the bus in response to GitHub's codegen tool having been trained on open-source code.
I'm disappointed to see folks putting out arguments that statistical models derived from mass data harvesting are "fair use". The same is happening with people's personal photos, writings, biometrics, etc. and we need to see it as the injustice it is.
in this specific instance the fair use argument seems second-order to me because the GitHub terms of service grant them a broad license to your code already
Nope. Github ToS are irrelevant when the party using Github is not the copyright holder and does not have license to sublicence to Github with additional permissions.