The "disagreement" was never dealt with as far as I can tell --- OpenAI's safety approach hasn't become more conservative --- which means that the only effect of bringing it to the CEO beforehand was to try to have it suppressed.