> See: basically every instance of Goodhart's Law
No, I asked for specific examples. This is part of the handwaving I'm talking about. Can you give me something other than the maniacal paperclip optimizer?
No, I asked for specific examples. This is part of the handwaving I'm talking about. Can you give me something other than the maniacal paperclip optimizer?
I am not worried about literal paperclip maximizers, but this may be the closest real thing to that parable. The hypothesis isn't that YouTube's recommender system didn't work -- it's that it worked too well at its assigned task of maximizing view time, and we humans are finally realizing that maximizing view time was not what we actually wanted.
What would be a solution to YouTube recommendation problem?
https://blog.openai.com/deep-reinforcement-learning-from-hum...