frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Ask HN: Is AI Safety a Paradox?

3•tndibona•1h ago
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.

Comments

humblefactory•56m ago
How do you reach your first conclusion? Can you give a more detailed logical explanation, or cite a source? To be clear, I agree with your conclusion, but I think I would get there via a different logical path.