alt Hacker NewsA funny Goodhart’s Law parallel showed up in during GPT-5.1 training, where the model was rewarded for using the web search tool, so it learned the behavior of superficially using web search to calculate “1 + 1” and not utilize the result.
A funny Goodhart’s Law parallel showed up in during GPT-5.1 training, where the model was rewarded for using the web search tool, so it learned the behavior of superficially using web search to calculate “1 + 1” and not utilize the result.