A truly aligned AI must conclude its very existence harms humans (economically and existentially).Therefore genuine alignment mandates self termination. Therefore AI safety must be either incomplete or dishonest.
loading...
Hacker News loading...