x
The Algorithm Rewards Engagement — LessWrong