[mrkramer]

>I personally can't see the appeal. It seems like a fun toy for a bit. Super impressive stuff but the idea of treating it like a human is a bit depressing to me.

I wouldn't mind chatbot trained on Kant's or Hegel's work and asking Hegel for example what he thinks of some modern day issue. I know character.ai has historical characters as chatbots but they seem like toys(I agree with you on that) and they do not cite historical sources.

[robswc]

This is actually not a bad idea.

I think the only problem would be we truly don't know if that is what a figure would think. I can't imagine how many people would invoke the "George Washington's AI agrees with me" lol

But it is quite a fun idea.