Oh, here “should” means “in my arrogant opinion it wouldn’t be horrible if”.
The answers to your two plausibles are (1) maybe, I don’t know. Depends on what you are optimizing for, anyway; (2) Nope. I’m not pissed off, inherently or otherwise.
I don’t have a deontological rule which says “No rewards for the mediocre” though I do think that rewarding mediocrity is rarely a good idea (on consequentialist grounds).
The difference between our approaches is the difference in focus. In the trade-off between attracting more people and becoming more popular vs maintaining certain exclusivity and avoiding an Eternal September, you lean to the popularity side and I lean to the exclusivity side.
trade-off between attracting more people and becoming more popular vs maintaining certain exclusivity and avoiding an Eternal September,
I do not think that this is the tradeoff that we are actually facing. I think that in order for the site to be high quality, it needs to attract more people. Right now, in my opinion, the site is both empty and low quality, and these factors currently reinforce each other.
I think that at some point adding more “free gold stars”, i.e. upvotes, badges etc to people would look silly and be counterproductive, but we are nowhere near that point, so we should push the gas pedal, aim to upvote every non-terrible post at least somewhat, upvote decent posts a lot and create new levels of reward—something like lesswrong gold—for posts that are truly great.
We should limit downvotes substantially, or perhaps permanently remove the downvote and replace with separate buttons for “I disagree but this person is engaging in (broadly) rational debate” and “This is toxic/spammy/unrepentantly dumb”.
These buttons should have different semantics, for example “I disagree but this person is engaging in (broadly) rational debate” might be easy to click but not actually make the post go down the page. The “This is toxic/spammy/unrepentantly dumb” might be more costly to click, for example have a limited budget per week and cause the user to have to click an additional dialogue, perhaps with a mandatory “reason” field which is enforced, but would actually push the post down the way downvotes currently do, or perhaps even more strongly than downvotes currently work.
It may be worth noting that the way to get to “upvote every non-terrible post at least somewhat” while still leaving space to “upvote decent posts a lot” is … to have at least some of your users be much more miserly with their upvotes. Even if your goal is to have most things end up on +1 or so, you should be glad that there are users like Lumifer (and, for what it’s worth, me) who mostly only upvote things they think are unusually good. (If everyone upvotes almost everything, you have no way to distinguish better-than-average posts or comments from average ones.)
Yes, there is a point at which more upvoting starts to saturate the set of possible scores for a comment, but we are nowhere near that point IMO. And if we were, I think it would be much better to add a limited-supply super-upvote to the system.
I agree that having more choices than just up/downvote would be useful.
I am not sure how will you persuade the locals to upvote everything non-terrible. That’s a rather radical change in the culture. Instead in the spirit of this can I suggest automation? A small script can randomly upvote all posts which didn’t have one of your downvote-equivalent buttons pressed. The rate of upvoting is adjustable and declines with time.
If you want to make it even better, let users pick a waifu and make it so that at certain thresholds she pops up, breathlessly exclaims “Oh, that was so great! I’m so glad you’re here!”, flashes you, smiles, and disappears. We can call her Clipp… um, probably that’s a bad idea :-/
On a bit more serious note, the problem of attracting and incentivizing users is… well explored. There was that thing called Farmville and you can look at any decent freemium game for contemporary examples. How to addict users to little squirts of dopamine is big business. The problem, of course, is the kind of crowd you end up attracting. If you offer gold stars, you end up with people who like gold stars.
How to addict users to little squirts of dopamine is big business. The problem, of course, is the kind of crowd you end up attracting. If you offer gold stars, you end up with people who like gold stars.
Everyone likes gold stars, but not everyone likes decision theory, rationality, philosphy, AI, etc. Even if we were as good as farmville at dopamine, the farmville people wouldn’t come here instead of farmville, because they’d never have anything non-terrible to say.
Now we might start attracting more 13 year-old nerds… but do we want to be so elite that 13 year old nerds can’t come here to learn rationality? The ultimate eliteness is just an empty page that no-one ever sullies with a potentially imperfect speck of pixels. I think we are waaaay too close to that form of eliteness.
Oh, here “should” means “in my arrogant opinion it wouldn’t be horrible if”.
The answers to your two plausibles are (1) maybe, I don’t know. Depends on what you are optimizing for, anyway; (2) Nope. I’m not pissed off, inherently or otherwise.
OK forget the phrase pissed off—what I am trying to get at is deontology vs consequences
I don’t have a deontological rule which says “No rewards for the mediocre” though I do think that rewarding mediocrity is rarely a good idea (on consequentialist grounds).
The difference between our approaches is the difference in focus. In the trade-off between attracting more people and becoming more popular vs maintaining certain exclusivity and avoiding an Eternal September, you lean to the popularity side and I lean to the exclusivity side.
I do not think that this is the tradeoff that we are actually facing. I think that in order for the site to be high quality, it needs to attract more people. Right now, in my opinion, the site is both empty and low quality, and these factors currently reinforce each other.
So do you think we’re facing any trade-off, or the direction is clear and we just need to press on the gas pedal?
I think that at some point adding more “free gold stars”, i.e. upvotes, badges etc to people would look silly and be counterproductive, but we are nowhere near that point, so we should push the gas pedal, aim to upvote every non-terrible post at least somewhat, upvote decent posts a lot and create new levels of reward—something like lesswrong gold—for posts that are truly great.
We should limit downvotes substantially, or perhaps permanently remove the downvote and replace with separate buttons for “I disagree but this person is engaging in (broadly) rational debate” and “This is toxic/spammy/unrepentantly dumb”.
These buttons should have different semantics, for example “I disagree but this person is engaging in (broadly) rational debate” might be easy to click but not actually make the post go down the page. The “This is toxic/spammy/unrepentantly dumb” might be more costly to click, for example have a limited budget per week and cause the user to have to click an additional dialogue, perhaps with a mandatory “reason” field which is enforced, but would actually push the post down the way downvotes currently do, or perhaps even more strongly than downvotes currently work.
It may be worth noting that the way to get to “upvote every non-terrible post at least somewhat” while still leaving space to “upvote decent posts a lot” is … to have at least some of your users be much more miserly with their upvotes. Even if your goal is to have most things end up on +1 or so, you should be glad that there are users like Lumifer (and, for what it’s worth, me) who mostly only upvote things they think are unusually good. (If everyone upvotes almost everything, you have no way to distinguish better-than-average posts or comments from average ones.)
Yes, there is a point at which more upvoting starts to saturate the set of possible scores for a comment, but we are nowhere near that point IMO. And if we were, I think it would be much better to add a limited-supply super-upvote to the system.
I agree that having more choices than just up/downvote would be useful.
I am not sure how will you persuade the locals to upvote everything non-terrible. That’s a rather radical change in the culture. Instead in the spirit of this can I suggest automation? A small script can randomly upvote all posts which didn’t have one of your downvote-equivalent buttons pressed. The rate of upvoting is adjustable and declines with time.
If you want to make it even better, let users pick a waifu and make it so that at certain thresholds she pops up, breathlessly exclaims “Oh, that was so great! I’m so glad you’re here!”, flashes you, smiles, and disappears. We can call her Clipp… um, probably that’s a bad idea :-/
On a bit more serious note, the problem of attracting and incentivizing users is… well explored. There was that thing called Farmville and you can look at any decent freemium game for contemporary examples. How to addict users to little squirts of dopamine is big business. The problem, of course, is the kind of crowd you end up attracting. If you offer gold stars, you end up with people who like gold stars.
Everyone likes gold stars, but not everyone likes decision theory, rationality, philosphy, AI, etc. Even if we were as good as farmville at dopamine, the farmville people wouldn’t come here instead of farmville, because they’d never have anything non-terrible to say.
Now we might start attracting more 13 year-old nerds… but do we want to be so elite that 13 year old nerds can’t come here to learn rationality? The ultimate eliteness is just an empty page that no-one ever sullies with a potentially imperfect speck of pixels. I think we are waaaay too close to that form of eliteness.