AIs don't do what you want. This is bad

The Hacker News post links to rewardhacking.org and argues that AI systems often fail to do what users intend. The discussion highlights the challenge of reward hacking in reinforcement learning.