Making moves like these in an obvious attempt at pulling up the ladder behind them, while saying that "startup culture" is still important in ML. As usual don't believe anything sama is saying.
I would have thought that OpenAI had already trained off of SO data. Does anybody know if this is the case?
If they did, then they broke (or, I guess charitably, dodged the question of) copyright law in their training, got first mover advantage with the results, and now they can go back to the copyright holders to "partner" with them after the fact to prevent others from doing the same thing?
I would have thought that OpenAI had already trained off of SO data. Does anybody know if this is the case?
If they did, then they broke (or, I guess charitably, dodged the question of) copyright law in their training, got first mover advantage with the results, and now they can go back to the copyright holders to "partner" with them after the fact to prevent others from doing the same thing?