While I’m at the SIAI house, I’m trying to orient towards the local priorities so as to be useful. Among the priorities is building community via Less Wrong, specifically by writing posts. Historically, the limiting factor on how much I post has been a desire not to flood the place—if I started posting as fast as I can write up my ideas, I’d get three or four posts out a week with (I think) no discernible decrease in quality. I have the following questions about this course of action:
Will it annoy people? Building community by being annoying seems very unlikely to work.
Will it affect voting behavior noticeably? I rely on my post’s karma scores to determine what to do and not do in the future, and SIAI people who decide whether I’m useful enough to keep use it as a rough metric too. I’d rather post one post that gets 40 karma in a week than two that get 20, and so on.
As your goal is to build community, I would time new posts based on posting and commenting activity. For example, whenever there is a lull, this would be an excellent time to make a new post. (I noticed over the weekend there were some times when 45 minutes would pass between subsequent comments and wished for a new post to jazz things up.)
On the other hand, if there are several new posts already, then it would be nice to wait until their activity has waned a bit.
I think that it is optimal to have 1 or 2 posts ‘going on’ at a time. I prefer the second post when one of them is technical and/or of focused interest to a smaller subset of Less Wrongers.
I’d say damn the torpedoes, full speed ahead. If people are annoyed, let them downvote. If posts start getting downvoted, slow down.
Your posts have generally been voted up. If now is the golden moment of time where you can get everything said, then for the love of Cthulhu, say it now!
I don’t anticipate being so obnoxiously prolific that people collectively start voting my posts negative such that they stay that way. But people already sometimes register individual downvotes on posts that I make, and I don’t want that to happen on a larger fraction of posts due to increased frequency, because I can’t reliably distinguish between “you must have had an off day, this post is not up to scratch” and “please, please, please shut up”.
The best signal to anticipate from the audience in this case is “how many votes in total do I expect if I post at full speed vs how many votes I expect if I posted less frequently and so ended up writing less posts overall”. Increased frequency may give you less votes per post. Frequent posts from the same author may be less desired and if you post less you may only be giving the best posts. But if the net expectation is higher for more prolific posting then that can be interpreted as “the lesswrong.com community would prefer you to post faster than a spambot”.
Even if you expected a total of less karma for more posts I wouldn’t say that means you ought not post more. So long as your posts are still breaking the 10 mark we clearly don’t mind your contribution. There are probably other benefits to you from posting than maximising the benefit to lesswrong. I find writing helps clarify my thinking for example. So as long as you are still being received somewhat positively you are free to type away.
one obvious idea that I didn’t notice anyone else mention:
Another option is to go ahead and write the posts as fast as you think is optimal, but if you think this is too fast to actually post the stuff you’ve written, then you can wait a few days after you wrote it before posting.
LW has a handy “drafts” feature that you can use for that.
This also has the advantage that you have more time to improve the article before you post it, but the disadvantage that you may be tempted to spend too much time making minor, unimportant improvements. Another disadvantage is that feedback gets delayed.
A related question: If I have a large topic to cover, should I cover it in one post, or split it up along convenient cleavage planes and make it a sequence? (If I make sequences, I think I’ll learn my lesson from the last one I tried and write it all before posting anything, so I don’t post 2⁄3 of it and then stop.)
Long posts are more offputting than short ones, and individual steps are more likely to be correct than entire theorems—both of these points would suggest posting sequences preferentially.
An inquiry regarding my posting frequency:
While I’m at the SIAI house, I’m trying to orient towards the local priorities so as to be useful. Among the priorities is building community via Less Wrong, specifically by writing posts. Historically, the limiting factor on how much I post has been a desire not to flood the place—if I started posting as fast as I can write up my ideas, I’d get three or four posts out a week with (I think) no discernible decrease in quality. I have the following questions about this course of action:
Will it annoy people? Building community by being annoying seems very unlikely to work.
Will it affect voting behavior noticeably? I rely on my post’s karma scores to determine what to do and not do in the future, and SIAI people who decide whether I’m useful enough to keep use it as a rough metric too. I’d rather post one post that gets 40 karma in a week than two that get 20, and so on.
As your goal is to build community, I would time new posts based on posting and commenting activity. For example, whenever there is a lull, this would be an excellent time to make a new post. (I noticed over the weekend there were some times when 45 minutes would pass between subsequent comments and wished for a new post to jazz things up.)
On the other hand, if there are several new posts already, then it would be nice to wait until their activity has waned a bit.
I think that it is optimal to have 1 or 2 posts ‘going on’ at a time. I prefer the second post when one of them is technical and/or of focused interest to a smaller subset of Less Wrongers.
(But otherwise no limit on the rate of posts.)
I’d say damn the torpedoes, full speed ahead. If people are annoyed, let them downvote. If posts start getting downvoted, slow down.
Your posts have generally been voted up. If now is the golden moment of time where you can get everything said, then for the love of Cthulhu, say it now!
I don’t anticipate being so obnoxiously prolific that people collectively start voting my posts negative such that they stay that way. But people already sometimes register individual downvotes on posts that I make, and I don’t want that to happen on a larger fraction of posts due to increased frequency, because I can’t reliably distinguish between “you must have had an off day, this post is not up to scratch” and “please, please, please shut up”.
Post away.
The best signal to anticipate from the audience in this case is “how many votes in total do I expect if I post at full speed vs how many votes I expect if I posted less frequently and so ended up writing less posts overall”. Increased frequency may give you less votes per post. Frequent posts from the same author may be less desired and if you post less you may only be giving the best posts. But if the net expectation is higher for more prolific posting then that can be interpreted as “the lesswrong.com community would prefer you to post faster than a spambot”.
Even if you expected a total of less karma for more posts I wouldn’t say that means you ought not post more. So long as your posts are still breaking the 10 mark we clearly don’t mind your contribution. There are probably other benefits to you from posting than maximising the benefit to lesswrong. I find writing helps clarify my thinking for example. So as long as you are still being received somewhat positively you are free to type away.
Post as much as you like, if you think it’s good quality; I promise to say if I start to think slowing down would be a good idea.
I don’t mean “downvoted negative” just “downvoted relative to other posters”.
one obvious idea that I didn’t notice anyone else mention:
Another option is to go ahead and write the posts as fast as you think is optimal, but if you think this is too fast to actually post the stuff you’ve written, then you can wait a few days after you wrote it before posting.
LW has a handy “drafts” feature that you can use for that.
This also has the advantage that you have more time to improve the article before you post it, but the disadvantage that you may be tempted to spend too much time making minor, unimportant improvements. Another disadvantage is that feedback gets delayed.
If I sit on posts for too long, I start second-guessing myself and often wind up deleting them.
A related question: If I have a large topic to cover, should I cover it in one post, or split it up along convenient cleavage planes and make it a sequence? (If I make sequences, I think I’ll learn my lesson from the last one I tried and write it all before posting anything, so I don’t post 2⁄3 of it and then stop.)
I really like the “sequences” approach—it’s easier to read and digest a chunk at a time, and it focusses discussion well, too.
Long posts are more offputting than short ones, and individual steps are more likely to be correct than entire theorems—both of these points would suggest posting sequences preferentially.
As for a specific reference on length: thirty-three hundred words sharply focused on a single, vivid subject is pushing the upper limit of what I find comfortable to attack in a single sitting.
Posting 2⁄3 of a sequence and stopping is fine if people turn out not to be interested. I recommend fast posting and fast feedback.
It seems to me that any strategy that does not end up with three posts by you on “Recent Posts” seems fine, as a rule of thumb.
Respondents, please upvote thomblake’s comment if this seems like an acceptable rule of thumb.
Edit: And likewise for other things people say if those seem like good ideas.