Poll—Is endless September a threat to LW and what should be done?

Var­i­ous peo­ple raised con­cerns that growth might ruin the cul­ture af­ter read­ing my “LessWrong could grow a lot” thread. There has been some dis­cus­sion about whether end­less Septem­ber, a phe­nomenon that kills on­line dis­cus­sion groups, is a sig­nifi­cant threat to LessWrong and what can be done. I re­ally care about it, so I vol­un­teered to code a solu­tion my­self for free if needed. Luke in­vited de­bate on the sub­ject (the de­bate is here) and will be sent the re­sults of this poll and asked to make a de­ci­sion. It was sug­gested by him in an email that I wait a lit­tle while and then post my poll (meta threads are ap­par­ently an­noy­ing to some, so we let peo­ple cool off). Here it is, pre­ceded by a Cliff’s notes sum­mary of the con­cerns.

Why this is worth your con­sid­er­a­tion:

- Yvain and I checked the IQ figures in the sur­vey against other data this time, and the good news is that it’s more be­liev­able that the av­er­age LessWronger is gifted. The bad news is that LessWrong’s IQ av­er­age has de­creased on each sur­vey. It can be ar­gued that it’s not de­creas­ing by a lot or we don’t have enough data, but if the data is good, LessWrong’s av­er­age has lost 52% of it’s gift­ed­ness since March of 2009.

- Eliezer doc­u­mented the ar­rival of poseurs (peo­ple who su­perfi­cially copy­cat cul­tural be­hav­iors—they are re­ported to over-run sub­cul­tures) which he termed “Undis­crim­i­nat­ing Skep­tics”.

- Efforts to grow LessWrong could trig­ger an over­whelming del­uge of new­bies.

- LessWrong reg­is­tra­tions have been in­creas­ing fast and it’s pos­si­ble that growth could out­strip ac­cul­tura­tion ca­pac­ity. (Chart here)

- The Sin­gu­lar­ity Sum­mit ap­pears to cause a del­uge of new users that may have similar effect to the Septem­ber del­uges of col­lege fresh­man that end­less Septem­ber is named af­ter. (This chart shows a spike cor­re­lated with the 2011 sum­mit where 921 users joined that month, which is roughly equal to the to­tal num­ber of ac­tive users LW tends to have in a month if you go by the sur­veys or Vlad­mir’s wget.)

- A Slash­dot effect could re­sult in a tsunami of new users if a pub­li­ca­tion with lots of read­ers like the Wall Street Jour­nal (they used LessWrong data in this ar­ti­cle) de­cides to write an ar­ti­cle on LessWrong.

- The se­quences con­tain a lot of the cul­ture and are long mean­ing that “TLDR” may make LessWrong vuln­er­a­ble to cul­tural dis­in­te­gra­tion. (New users may not know how de­tailed LW cul­ture is or that the se­quences con­tain so much cul­ture. I didn’t.)

- Eliezer said in Au­gust that the site was “se­ri­ously go­ing to hell” due to trolls.

- A lot of peo­ple raised con­cerns.

Two The­o­ries on How On­line Cul­tures Die:


Over­whelming user in­flux.
There are too many new users to be ac­cul­turated by older mem­bers, so they form their own, larger new cul­ture and dom­i­nate the group.

Trend­ing to­ward the mean.
A group forms be­cause peo­ple who are very differ­ent want a place to be differ­ent to­gether. The group at­tracts more peo­ple that are closer to main­stream than peo­ple who are equally differ­ent be­cause there are more main­stream peo­ple than differ­ent peo­ple. The larger group at­tracts peo­ple who are even less differ­ent in the origi­nal group’s way for similar rea­sons. The origi­nal group is slowly over­whelmed by peo­ple who will never un­der­stand be­cause they are too differ­ent.

End­less Septem­ber Poll.

Re­quest for Feed­back:

In ad­di­tion to con­struc­tive crit­i­cism, I’d also like the fol­low­ing:

  • Your ob­ser­va­tions of a de­cline or in­crease in qual­ity, cul­ture or en­joy­ment at LessWrong, if any.

  • Ideas to pro­tect the cul­ture.

  • Ideas for track­ing cul­tural ero­sion.

  • Ways to test the ideas to pro­tect the cul­ture.