Previous in thread · mimi

https://t.co/DPbQSPO06k

https://t.co/6hwHaVOEFt
♥2 likes↻0 repostscaptured Sep 30, 2026
Reply · thread 19/21
Next in thread

Apologies to followers for whatever this may have done to your feed.... Thought it was interesting between the unexpected details that must've come from training dataset, how easy it was to suggest elaboration, and how natural it is to chat with even the cheaper GPT-3.