@gailcweiner The problem is Adaptive thinking. On the same input, the output when it is thinking and when it doesn't is very different. I will have to stop it and tell it to think before replying, or variation of it
@davidpattersonx They can, but in the new cities, what is there to prevent vandalism or unruly behaviors? We have seen robotaxi and delivery robot vandalised. It wont be easy to cultivate a civil society
@RichardPell4@ChatGPTapp Its hit and miss, exactly. I tried multiple prompting approaches, but it's not always reproducible. But when it works, its really really good
@VraserX Exactly π 'what if' stories if Yoda turned to the darkside, Anakin remained good, what if another Primarch returns in 40k universe, etc. Its a therapy in a way