There are several ways to approach AI Avatars, so I wanted to compare experiences.
I am particularly interested in evaluation and security. I would like to compare approaches that people have actually used rather than just theoretical suggestions.
What trade-offs have you found between...