I’m split on which is worse tbh.
The Tulpa thing takes more effort than just talking to an AI, but on the other hand, it’s has to be much harder to snap back to reality when the voices are literally in your head…