>>25015
If some human writers are willing to compete or maybe improve any of the stories That woud be nice to see.

>>25016
There is not much required to do:

get one click installer from here
https://github.com/oobabooga/text-generation-webui
and select some AI model

The problem is that with 8gb 3050 it will not be able to do much and you may even get comparable performance with CPU inference if slightly larger model especially if you split it between CPU and GPU

all this process is not so much computationally but memory bandwidth limited
since entire model has to be processed every time for each token.