Revitalizing Less Wrong seems like a lost purpose, but here are some other ideas

This is a re­sponse to in­gres’ re­cent post shar­ing Less Wrong sur­vey re­sults. If you haven’t read & up­voted it, I strongly en­courage you to—they’ve done a fab­u­lous job of col­lect­ing and pre­sent­ing data about the state of the com­mu­nity.

So, there’s a bit of a con­tra­dic­tion in the sur­vey re­sults. On the one hand, peo­ple say the com­mu­nity needs to do more schol­ar­ship, be more rigor­ous, be more prac­ti­cal, be more hum­ble. On the other hand, not much is get­ting posted, and it seems like rais­ing the bar will only ex­ac­er­bate that prob­lem.

I did a query against the sur­vey database to find the com­plaints of top Less Wrong con­trib­u­tors and figure out how best to serve their needs. (Note: it’s a bit hard to read the com­ments be­cause some of them should start with “the com­mu­nity needs more” or “the com­mu­nity needs less”, but adding that info would have meant con­struct­ing a much more com­pli­cated query.) One user wrote:

[it’s not so much that there are] overly high stan­dards, just not a very civil or wel­com­ing cli­mate . why write con­tent for free and get trashed when I can go write a grant ap­pli­ca­tion or a manuscript in­stead?

in­gres em­pha­sizes that in or­der to re­vi­tal­ize the com­mu­nity, we would need more con­tent. Con­tent is im­por­tant, but in­cen­tives for pro­duc­ing con­tent might be even more im­por­tant. So­cial sta­tus may be the in­cen­tive hu­mans re­spond most strongly to. Right now, from a so­cial sta­tus per­spec­tive, the ex­pected value of cre­at­ing a new Less Wrong post doesn’t feel very high. Par­tially be­cause many LW posts are get­ting down­votes and crit­i­cal com­ments, so my Sys­tem 1 says my posts might as well. And par­tially be­cause the Less Wrong brand is weak enough that I don’t ex­pect as­so­ci­at­ing my­self with it will boost my so­cial sta­tus.

When Less Wrong was founded, the pri­mary failure mode guarded against was Eter­nal Septem­ber. If Eter­nal Septem­ber rep­re­sents a sort of digi­tal pop­ulism, Less Wrong was at­tempt­ing a sort of digi­tal elitism. My per­cep­tion is that elitism isn’t work­ing be­cause the benefits of join­ing the elite are too small and the costs are too large. Teddy Roo­sevelt talked about the man in the arena—I think Less Wrong ex­pe­rienced the re­verse of the evap­o­ra­tive cool­ing EY feared, where peo­ple grad­u­ally left the arena as the pro­por­tional num­ber of crit­ics in the stands grew ever larger.

Given where Less Wrong is at, how­ever, I sus­pect the goal of re­vi­tal­iz­ing Less Wrong rep­re­sents a lost pur­pose.

in­gres’ sur­vey re­ceived a to­tal of 3083 re­sponses. Not only is that about twice the num­ber we got in the last sur­vey in 2014, it’s about twice the num­ber we got in 2013, 2012, and 2011 (though much big­ger than the first sur­vey in 2009). It’s hard to know for sure, since pre­vi­ous sur­veys were only ad­ver­tised on the do­main, but it doesn’t seem like the di­as­pora thing has slowed the growth of the com­mu­nity a ton and it may have dra­mat­i­cally ac­cel­er­ated it.

Why has the com­mu­nity con­tinued grow­ing? Here’s one pos­si­bil­ity. Maybe Less Wrong has been re­placed by su­pe­rior al­ter­na­tives.

  • CFAR—in­gres writes: “If LessWrong is se­ri­ous about it’s goal of ‘ad­vanc­ing the art of hu­man ra­tio­nal­ity’ then it needs to figure out a way to do real in­ves­ti­ga­tion into the sub­ject.” That’s ex­actly what CFAR does. CFAR is a su­pe­rior al­ter­na­tive for peo­ple who want some­thing like Less Wrong, but more prac­ti­cal. (They have an alumni mailing list that’s higher qual­ity and more ac­tive than Less Wrong.) Yes, CFAR costs money, be­cause do­ing re­search costs money!

  • Effec­tive Altru­ism—A su­pe­rior al­ter­na­tive for peo­ple who want some­thing that’s more fo­cused on re­sults.

  • Face­book, Tum­blr, Twit­ter—Peo­ple are go­ing to be wast­ing time on these sites any­way. They might as well talk about ra­tio­nal­ity while they do it. Like all those ph­pBB boards in the 00s, Less Wrong has been out­com­peted by the hot new thing, and I think it’s prob­a­bly bet­ter to roll with it than fight it. I also wouldn’t be sur­prised if in­ter­act­ing with oth­ers through so­cial me­dia has been a cause of com­mu­nity growth.

  • SlateS­tarCodex—SSC already checks most of the boxes un­der in­gres’ “Fu­ture Im­prove­ment Wish­list Based On Sur­vey Re­sults”. In my opinion, the av­er­age SSC post has bet­ter schol­ar­ship, rigor, and hu­mil­ity than the av­er­age LW post, and the com­mu­nity seems less in­timi­dat­ing, less ar­gu­men­ta­tive, more ac­cessible, and more ac­cept­ing of out­side view­points.

  • The meatspace com­mu­nity—Meet­ing in per­son has lots of ad­van­tages. Real-time dis­cus­sion us­ing Slack/​IRC also has ad­van­tages.

Less Wrong had a great run, and the su­pe­rior al­ter­na­tives wouldn’t ex­ist in their cur­rent form with­out it. (LW was eas­ily the most com­mon way peo­ple heard about EA in 2014, for in­stance, al­though sam­pling effects may have dis­torted that es­ti­mate.) But that doesn’t mean it’s the best op­tion go­ing for­ward.

There­fore, here are some things I don’t think we should do:

  • Try to be a sec­ond-rate ver­sion of any of the su­pe­rior al­ter­na­tives I men­tioned above. If some­one’s go­ing to put some­thing to­gether, it should fulfill a real com­mu­nity need or be the best al­ter­na­tive available for what­ever pur­pose it serves.

  • Try to get old con­trib­u­tors to re­turn to Less Wrong for the sake of get­ting them to re­turn. If they’ve judged that other ac­tivi­ties are a bet­ter use of time, we should prob­a­bly trust their judge­ment. It might be sen­si­ble to make an ex­cep­tion for old posters that never trans­ferred to the in-per­son com­mu­nity, but they’d be harder to track down.

  • Try to solve the same sort of prob­lems Ar­bital or Me­tac­u­lus is op­ti­miz­ing for. No rea­son to step on the toes of other pro­jects in the com­mu­nity.

But that doesn’t mean there’s noth­ing to be done. Here are some pos­si­ble weak­nesses I see with our cur­rent setup:

  • If you’ve got a great idea for a blog post, and you don’t already have an on­line pres­ence, it’s a bit hard to reach lots of peo­ple, if that’s what you want to do.

  • If we had a good sys­tem for in­cen­tiviz­ing peo­ple to write great stuff (as op­posed to merely tol­er­at­ing great stuff the way LW cul­ture his­tor­i­cally has), we’d get more great stuff writ­ten.

  • It can be hard to find good con­tent in the di­as­pora. Pos­si­ble solu­tion: Weekly “di­as­pora roundup” posts to Less Wrong. I’m too busy to do this, but any­one else is more than wel­come to (as­sum­ing both peo­ple read­ing LW and peo­ple in the di­as­pora want it).

  • EDIT 11/​27/​16 - Re­cently peo­ple have been ar­gu­ing that so­cial me­dia gen­er­ates rel­a­tively su­perfi­cial dis­cus­sions. This plau­si­bly un­der­mines my “lost pur­pose” the­sis.

in­gres men­tions the pos­si­bil­ity of Scott Alexan­der some­how open­ing up SlateS­tarCodex to other con­trib­u­tors. This seems like a clearly su­pe­rior al­ter­na­tive to re­vi­tal­iz­ing Less Wrong, if Scott is down for it:

  • As I men­tioned, SSC already seems to have solved most of the cul­ture & philos­o­phy prob­lems that peo­ple com­plained about with Less Wrong.

  • SSC has no short­age of con­tent—Scott has in­creased the rate at which he cre­ates open threads to deal with an ex­cess of com­ments.

  • SSC has a stronger brand than Less Wrong. It’s been linked to by Ezra Klein, Ross Douthat, Bryan Ca­plan, etc.

But the most im­por­tant rea­sons may be be­hav­ioral rea­sons. SSC has more traf­fic—peo­ple are in the habit of vis­it­ing there, not here. And the post­ing habits peo­ple have ac­quired there seem more con­ducive to com­mu­nity. Chang­ing habits is hard.

As in­gres writes, re­vi­tal­iz­ing Less Wrong is prob­a­bly about as difficult as cre­at­ing a new site from scratch, and I think cre­at­ing a new site from scratch for Scott is a su­pe­rior al­ter­na­tive for the rea­sons I gave.

So if there’s any­one who’s in­ter­ested in im­prov­ing Less Wrong, here’s my hum­ble recom­men­da­tion: Go tell Scott Alexan­der you’ll build an on­line fo­rum to his speci­fi­ca­tion, with SSC com­mu­nity feed­back, to provide a bet­ter solu­tion for his overflow­ing open threads. Once you’ve solved that prob­lem, keep mak­ing im­prove­ments and sub­fora so your fo­rum be­comes the best available al­ter­na­tive for more and more use cases.

And here’s my hum­ble sug­ges­tion for what an SSC fo­rum could look like:

As I men­tioned above, Eter­nal Septem­ber is analo­gous to a sort of digi­tal pop­ulism. The ma­jor so­cial me­dia sites of­ten have a “mob rule” cul­ture to them, and peo­ple are in­creas­ingly see­ing the dis­ad­van­tages of this model. Less Wrong tried to achieve digi­tal elitism and it didn’t work well in the long run, but that doesn’t mean it’s im­pos­si­ble. has found a model for digi­tal elitism that works. There may be other work­able mod­els out there. A work­able model could even turn in to a suc­cess­ful com­pany. Fight the hot new thing by be­com­ing the hot new thing.

My pro­posal is based on the idea of eigen­democ­racy. (Recom­mended that you read the link be­fore con­tin­u­ing—eigen­democ­racy is cool.) In eigen­democ­racy, your trust score is a com­pos­ite rat­ing of what trusted peo­ple think of you. (It sounds like in­finite re­cur­sion, but it can be re­solved us­ing lin­ear alge­bra.)

Ei­gen­democ­racy is a com­pli­cated idea, but a sim­ple way to get most of the way there would be to have a fo­rum where hav­ing lots of karma gives you the abil­ity to up­vote mul­ti­ple times. How would this work? Let’s say Scott starts with 5 karma and ev­ery­one else starts with 0 karma. Each point of karma gives you the abil­ity to up­vote once a day. Let’s say it takes 5 up­votes for a post to get fea­tured on the side­bar of Scott’s blog. If Scott wants to fea­ture a post on the side­bar of his blog, he up­votes it 5 times, net­ting the per­son who wrote it 1 karma. As Scott fea­tures more and more posts, he gains a mod­er­a­tion team full of peo­ple who wrote posts that were good enough to fea­ture. As they fea­ture posts in turn, they gen­er­ate more co-mod­er­a­tors.

Why do I like this solu­tion?

  • It acts as a cul­tural preser­va­tion mechanism. On red­dit and Twit­ter, sheer num­bers rule when de­ter­min­ing what gets visi­bil­ity. The red­dit-like vot­ing mechanisms of Less Wrong meant that the site de­liber­ately kept a some­what low pro­file in or­der to avoid get­ting over­run. Even if SSC ex­pe­rienced a large in­flux of new users, those users would only gain power to af­fect the visi­bil­ity of con­tent if they proved them­selves by mak­ing qual­ity con­tri­bu­tions first.

  • It takes the mod­er­a­tion bur­den off of Scott and dis­tributes it across trusted com­mu­nity mem­bers. As the com­mu­nity grows, the mod team grows with it.

  • The in­cen­tives seem well-al­igned. Writ­ing stuff Scott likes or meta-likes gets you recog­ni­tion, mod pow­ers, and the abil­ity to con­trol the dis­cus­sion—forms of so­cial sta­tus. Con­trast with so­cial me­dia sites where hy­per­bole is a short­cut to at­ten­tion, fol­low­ers, up­votes. Also, un­like Less Wrong, there’d be no pun­ish­ment for writ­ing a low qual­ity post—it sim­ply doesn’t get fea­tured and is one more click away from the SSC home­page.

TL;DR—De­spite ap­pear­ances, the Less Wrong com­mu­nity is ac­tu­ally do­ing great. Any suc­ces­sor to Less Wrong should try to offer com­pel­ling ad­van­tages over op­tions that are already available.