"It literally did not change. Not one bit." How do you know? Even if the base model didn't change, that doesn't mean they didn't fine tune it in some way ove

"It literally did not change. Not one bit." How do you know? Even if the base model didn't change, that doesn't mean they didn't fine tune it in some way over time. They also might be passing its answers through some other AI or using some other techniques to filter, censor, and/or modify the answers in some way before returning them to the user. I don't know how anyone could confidently say what they're doing unless they work at OpenAI.