Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

And I thought “Computer, disable holodeck safety protocols” was too simple.


I did that once on an AI Dungeon story about a holodeck technician, and the GPT-3-based model correctly inferred the next thing the character should do is to shoot themselves in the face with a phaser on stun, to test if disabling safeties actually works.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: