‘Eternal September’ situations are caused by new users as a proportion of total users. 10,000 new users in a month could cause an eternal September here, but 10,000 new users would barely be a blip on, say, Pinterest’s culture. Growth at a steady percentage of users seems like the safest way to head off an eternal September situation; as the site gains more users it becomes naturally more resilient to culture shift by newbies.
Is LessWrong increasing in growth%, decreasing in growth%, or are growth% numbers roughly staying the same?
I agree that it is safer to keep growth at a manageable pace than it is to try and grow it faster while also trying to prevent endless September. However, I would disagree with the idea that managing the pace of growth will prevent endless September. So relying on pacing to solve the problem seems like a bad idea to me.
My perspective is that groups trend toward normal over time as they grow larger—regardless of the pace. I think this is most noticeable during a deluge, that a large deluge will definitely speed up the process, and that growth curves are such that this can seem to happen over night. It may be “correlation implies causation” to assume that the reason a culture is watered down is due to too many people, I think it’s because it eventually attracts a snowball of people who are ever less like the original group. People who start a new group have some kind of difference—why would you break away from the herd if this is not so? Once you’ve started the group, it is less likely to find people who are very much like you than people who are just plain similar. So as the group grows, the people attracted to it are similar to the ones already there, but not really similar. Of course, as the circle of people enlarges, what is defined as similar becomes more relaxed over time. Different people are necessarily overwhelmed by mainstream people eventually due to the fact that there are more mainstream people than different people. That’s the way I’ve observed it happening. If you know an example of a culture that actually stayed the same with a very large numbers of users, I’d be interested in hearing about that. You mentioned Pinterest, but do we know whether their beginning culture was retained? It sounds like you are just saying the current culture would be retained. For all I know, they’ve already been through an endless September and the culture they have now is very different from the one they began with.
I’m glad you brought this up, worded in that way, because it revealed these different perspectives.
Because I think endless September can happen regardless of whether we pace growth, I still think it’s important to do something to prevent that.
As for whether it’s increasing or decreasing or staying the same… over the long-term, they’re definitely increasing. In the recent short-term, they’re decreasing. That looks to be due to the falling back you might see after a large spike (the spikes last a little and taper off), perhaps combined with a common dip in numbers (It looks pretty normal for this site for them to be low for a few months in a row).
‘Eternal September’ situations are caused by new users as a proportion of total users. 10,000 new users in a month could cause an eternal September here, but 10,000 new users would barely be a blip on, say, Pinterest’s culture. Growth at a steady percentage of users seems like the safest way to head off an eternal September situation; as the site gains more users it becomes naturally more resilient to culture shift by newbies.
Is LessWrong increasing in growth%, decreasing in growth%, or are growth% numbers roughly staying the same?
I agree that it is safer to keep growth at a manageable pace than it is to try and grow it faster while also trying to prevent endless September. However, I would disagree with the idea that managing the pace of growth will prevent endless September. So relying on pacing to solve the problem seems like a bad idea to me.
My perspective is that groups trend toward normal over time as they grow larger—regardless of the pace. I think this is most noticeable during a deluge, that a large deluge will definitely speed up the process, and that growth curves are such that this can seem to happen over night. It may be “correlation implies causation” to assume that the reason a culture is watered down is due to too many people, I think it’s because it eventually attracts a snowball of people who are ever less like the original group. People who start a new group have some kind of difference—why would you break away from the herd if this is not so? Once you’ve started the group, it is less likely to find people who are very much like you than people who are just plain similar. So as the group grows, the people attracted to it are similar to the ones already there, but not really similar. Of course, as the circle of people enlarges, what is defined as similar becomes more relaxed over time. Different people are necessarily overwhelmed by mainstream people eventually due to the fact that there are more mainstream people than different people. That’s the way I’ve observed it happening. If you know an example of a culture that actually stayed the same with a very large numbers of users, I’d be interested in hearing about that. You mentioned Pinterest, but do we know whether their beginning culture was retained? It sounds like you are just saying the current culture would be retained. For all I know, they’ve already been through an endless September and the culture they have now is very different from the one they began with.
I’m glad you brought this up, worded in that way, because it revealed these different perspectives.
Because I think endless September can happen regardless of whether we pace growth, I still think it’s important to do something to prevent that.
As for whether it’s increasing or decreasing or staying the same… over the long-term, they’re definitely increasing. In the recent short-term, they’re decreasing. That looks to be due to the falling back you might see after a large spike (the spikes last a little and taper off), perhaps combined with a common dip in numbers (It looks pretty normal for this site for them to be low for a few months in a row).