I think this misses a point. There’s no information about the model: how was it trained, on which data. It might be that it can actually “reason” in a human sen

I think this misses a point. There’s no information about the model: how was it trained, on which data. It might be that it can actually “reason” in a human sense, or this particular joke is hardcoded, or… I feel that with OpenAI withholding the details the whole discourse turns into religious flame war.