Hacker News new | ask | show | jobs
by AndrewKemendo 3171 days ago
It's honestly an impossible goal to make AGI safe for humans.

If an AGI cannot be independent in it's goals, meaning it could potentially harm humans, then it's not really AGI.