"There has been some dilution as a result. The stories that get voted up are not quite as good, and the average tone in comment threads is slightly less polite."
The problem is voting. As you might guess, I think a more meritocratic approach to social news (e.g., intelligence testing through puzzles) would give you better results when you have more users.
I'm not so sure. That's not so much meritocratic as a test of someone's determination or the amount of time they have to waste.
True meritocracy in online voting situations would be something like giving more importance to votes from people with higher karma. So, the people who get voted up more, then get more power in voting situations themselves. That way, the site becomes whatever the "top users" want, a true meritocracy.
I'm also a fan of giving precedence depending on the "age" of the user. That is, how long they've been a member of the site. This means the site may not evolve as quickly, but it appears that's something people want to avoid anyway. It tends to be the new "influx" of users that comes after popularity that drags a site down, so letting older users have more say could make sense.
"I'm not so sure. That's not so much meritocratic as a test of someone's determination or the amount of time they have to waste."
That depends on the nature of the puzzle. Not many people can solve Rubik's Cube no matter how determined they may be.
"True meritocracy in online voting situations would be something like giving more importance to votes from people with higher karma. So, the people who get voted up more, then get more power in voting situations themselves. That way, the site becomes whatever the "top users" want, a true meritocracy."
That may not lead to a meritocracy. I think intelligence testing is more reliable.
A good argument, but do you think a user's raw intelligence could correlate to the quality of their links and link preferences?
I don't. The people who post all the political links to Reddit seem quite intelligent, as do many of the people starting pointless arguments. The ability to troll and incite flame wars might even go UP with intelligence.
Perhaps if the puzzles you proposed measured "emotional intelligence" or domain specific knowledge we'd stand a chance, but raw intelligence doesn't prove much. Many of the nicest, most "savvy" people I know are not of high intelligence in an IQ sense, but are still very successful in their fields, and vice versa.
Yes, you can have different sorts of puzzles to measure different sorts of intelligence. It would make a fascinating experiment to see the impact of the puzzle type on link and discussion quality.
In fact, my next puzzle-based social news site will involve a word-based puzzle, which may reward users more who have high "verbal intelligence".
Just a guess, but I think domain specific knowledge testing would work best. I know enough grouchy "intelligent" people to go for the raw intelligence option, but if people were tested on their general knowledge of the sector that the site is based on, that's going to result in, hopefully, more informed discussion and links.
Indeed, perhaps this is why sites like Reddit and Hacker News were / are such high quality sites at the start of their lives, because only the most informed people knew about them. Once the less informed get involved, it all goes to pot. So testing how "informed" someone is might, effectively, resolve it all.. but I can't help but feel a lot of people would be turned off by such "testing"!
So testing how "informed" someone is might, effectively, resolve it all.. but I can't help but feel a lot of people would be turned off by such "testing"!
Puzzle based social news only appeals to people that like solving puzzles. That's fine and I'm glad your site exists, just don't expect your model to supplant vote based story ranking.
The problem is voting. As you might guess, I think a more meritocratic approach to social news (e.g., intelligence testing through puzzles) would give you better results when you have more users.