I love this change, for most of the same reasons as Ben. Thanks, LessWrong team! Some ideas for further ways to empower finer-grained epistemics at the community level:
(additional metrics) I think it’d be nice to have a drop-down or hover-over to see more fine-grained statistics on a post or comment, such as:
total_upvotes := the the total number of upvotes (ignoring downvotes)
I’m sure you guys look at variables like these internally, so I bet you already understand their value; the only thing missing is to trust and empower users to see and make use of them as well.
(custom sorting order) To make the community more robust against various kinds of Goodharting, I think it would be good to empower users to sort by any of the above metrics, so we enable a variety of different perspectives on the state of discourse. This degrades common-knowledge-of-what’s-important somewhat, but I think that’s probably worth it for helping users to prioritize their attention in the way that they want, for getting broader attentional coverage across topics (because different users will choose different metrics to prioritize), and for preventing one-particular-metric from becoming a Goodhart target. You could even collect data on which kinds of sorts users tend to choose, so you have your finger on the pulse of how the community is prioritizing attention.
(custom sorting functions) In the additional metrics dropdown, you could even allow a user to enter a custom function that they want to sort by, such as “voting_activity + agreement+controversy”. This adds even more robustness to the community to protect itself against goodharting in various ways.
Overall, big hearts for adding this finer-grained collective-thinking-tool :)
I love this change, for most of the same reasons as Ben. Thanks, LessWrong team! Some ideas for further ways to empower finer-grained epistemics at the community level:
(additional metrics) I think it’d be nice to have a drop-down or hover-over to see more fine-grained statistics on a post or comment, such as:
total_upvotes := the the total number of upvotes (ignoring downvotes)
total_downvotes
voting_activity := total_upvotes + total_downvotes
voting_controversy := min(total_upvotes, total_downvotes)
total_agreement
total_disagreement
agreement_activity := total_upvotes + total_downvotes)
agreement_controversy := min(total_upvotes, total_downvotes)
I’m sure you guys look at variables like these internally, so I bet you already understand their value; the only thing missing is to trust and empower users to see and make use of them as well.
(custom sorting order) To make the community more robust against various kinds of Goodharting, I think it would be good to empower users to sort by any of the above metrics, so we enable a variety of different perspectives on the state of discourse. This degrades common-knowledge-of-what’s-important somewhat, but I think that’s probably worth it for helping users to prioritize their attention in the way that they want, for getting broader attentional coverage across topics (because different users will choose different metrics to prioritize), and for preventing one-particular-metric from becoming a Goodhart target. You could even collect data on which kinds of sorts users tend to choose, so you have your finger on the pulse of how the community is prioritizing attention.
(custom sorting functions) In the additional metrics dropdown, you could even allow a user to enter a custom function that they want to sort by, such as “voting_activity + agreement+controversy”. This adds even more robustness to the community to protect itself against goodharting in various ways.
Overall, big hearts for adding this finer-grained collective-thinking-tool :)