how would you feel if someone published a proof that an AGI (agent able to interact in human environment at ~ human level of skill) cannot be "aligned" (guaranteed to never seek to kill everyone)?

Posters: mimi10v3. mimi10v3

2 likes recorded by the source corpus.