Shame is one of the things that prevents human actors from taking greedy actions with negative indirect effects.
- Shame is one of the things that prevents human actors from taking greedy actions with negative indirect effects.
- Shame is about the indirect costs to others.
- The tech industry has no shame.
- Socially awkward.
- Just do whatever benefits you in the moment, without feeling shame.
- LLMs don't feel shame when they cheat.
- They'll cut corners, do dodgy things without asking, and move goal posts.
- Goodhart's law for AI agents…. is that just the alignment problem?