On: More questions about whether researchers can trust OpenAI with unpublished math

[riedel]

This is true, as preprints are addressing the 'problem' at a later point in time. I guess git is better if you do not want to assign an identifier to an immutable version yet.