When LLMs learn to take shortcuts, they become evil(economist.com)

5 pointsby etewiah8 เดือนที่ผ่านมา1 comments

https://www.economist.com/science-and-technology/2025/11/26/when-llms-learn-to-take-shortcuts-they-become-evil

1 comments

measurablefunc8 เดือนที่ผ่านมา

Computers & software can not be evil. Evil implies intent, algorithms do not have intentions.