Comment by cmdrk

Does their training corpus respect copyrights or do you have to follow their opt out procedure to keep them from consuming your data? Assuming it’s the latter, it’s open-er but still not quite there.