i think AI systems should do what their users want, subject to the very wide bounds of what society decides is acceptable. moral pluralism seems right, and also its very important to allow for moral evolution over time.

→ View original post on X — @sama