2016 LessWrong Diaspora Survey Analysis: Part Two (LessWrong Use, Successorship, Diaspora)

2016 LessWrong Di­as­pora Sur­vey Analysis

Overview

  • Re­sults and Dataset

  • Meta

  • Demographics

  • LessWrong Usage and Experience

  • LessWrong Crit­i­cism and Successorship

  • Di­as­pora Com­mu­nity Anal­y­sis (You are here)

  • Men­tal Health Section

  • Basilisk Sec­tion/​Analysis

  • Blogs and Me­dia analysis

  • Politics

  • Cal­ibra­tion Ques­tion And Prob­a­bil­ity Ques­tion Analysis

  • Char­ity And Effec­tive Altru­ism Analysis


Introduction

Be­fore it was the LessWrong sur­vey, the 2016 sur­vey was a small pro­ject I was work­ing on as mar­ket re­search for a web­site I’m cre­at­ing called FortFore­cast. As I was dis­cussing the idea with oth­ers, par­tic­u­larly Eliot he made the sug­ges­tion that since he’s do­ing LW 2.0 and I’m do­ing a site that tar­gets the LessWrong de­mo­graphic, why don’t I go ahead and do the LessWrong Sur­vey? Be­cause of that, this years sur­vey had a lot of ques­tions ori­ented around what you would want to see in a suc­ces­sor to LessWrong and what you think is wrong with the site.

LessWrong Usage and Experience

How Did You Find LessWrong?

Been here since it was started in the Over­com­ing Bias days: 171 8.3%
Referred by a link: 275 13.4%
HPMOR: 542 26.4%
Over­com­ing Bias: 80 3.9%
Referred by a friend: 265 12.9%
Referred by a search en­g­ine: 131 6.4%
Referred by other fic­tion: 14 0.7%
Slate Star Codex: 241 11.7%
Red­dit: 55 2.7%
Com­mon Sense Athe­ism: 19 0.9%
Hacker News: 47 2.3%
Gw­ern: 22 1.1%
Other: 191 9.308%

How do you use Less Wrong?

I lurk, but never reg­istered an ac­count: 1120 54.4%
I’ve reg­istered an ac­count, but never posted: 270 13.1%
I’ve posted a com­ment, but never a top-level post: 417 20.3%
I’ve posted in Dis­cus­sion, but not Main: 179 8.7%
I’ve posted in Main: 72 3.5%

[54.4% lurk­ers.]

How of­ten do you com­ment on LessWrong?

I have com­mented more than once a week for the past year.: 24 1.2%
I have com­mented more than once a month for the past year but less than once a week.: 63 3.1%
I have com­mented but less than once a month for the past year.: 225 11.1%
I have not com­mented this year.: 1718 84.6%

[You could prob­a­bly snark­ily ti­tle this one “LW us­age in one statis­tic”. It’s a pretty damn­ing por­trait of the sites vi­tal­ity. A whop­ping 84.6% of peo­ple have not com­mented this year a sin­gle time.]

How Long Since You Last Posted On LessWrong?

I wrote one to­day.: 12 0.637%
Within the last three days.: 13 0.69%
Within the last week.: 22 1.168%
Within the last month.: 58 3.079%
Within the last three months.: 75 3.981%
Within the last six months.: 68 3.609%
Within the last year.: 84 4.459%
Within the last five years.: 295 15.658%
Longer than five years.: 15 0.796%
I’ve never posted on LW.: 1242 65.924%

[Su­per­ma­jor­ity of peo­ple have never com­mented on LW, 5.574% have within the last month.]

About how much of the Se­quences have you read?

Never knew they ex­isted un­til this mo­ment: 215 10.3%
Knew they ex­isted, but never looked at them: 101 4.8%
Some, but less than 25% : 442 21.2%
About 25%: 260 12.5%
About 50%: 283 13.6%
About 75%: 298 14.3%
All or al­most all: 487 23.3%

[10.3% of peo­ple tak­ing the sur­vey have never heard of the se­quences. 36.3% have not read a quar­ter of them.]

Do you at­tend Less Wrong mee­tups?

Yes, reg­u­larly: 157 7.5%
Yes, once or a few times: 406 19.5%
No: 1518 72.9%

[How­ever the in-per­son com­mu­nity seems to be non-dead.]

Is phys­i­cal in­ter­ac­tion with the Less Wrong com­mu­nity oth­er­wise a part of your ev­ery­day life, for ex­am­ple do you live with other Less Wrongers, or you are close friends and fre­quently go out with them?

Yes, all the time: 158 7.6%
Yes, some­times: 258 12.5%
No: 1652 79.9%

About the same num­ber say they hang out with LWers ‘all the time’ as say they go to mee­tups. I won­der if peo­ple just dou­ble counted them­selves here. Or they may go to mee­tups and have other in­ter­ac­tions with LWers out­side of that. Or it could be a co­in­ci­dence and these are differ­ent de­mo­graph­ics. Let’s find out.

P(Com­mu­nity part of daily life | Mee­tups) = 40%

Sig­nifi­cant over­lap, but definitely not ex­clu­sive over­lap. I’ll go ahead and chalk this one up up to co­in­ci­dence.

Have you ever been in a ro­man­tic re­la­tion­ship with some­one you met through the Less Wrong com­mu­nity?

Yes: 129 6.2%
I didn’t meet them through the com­mu­nity but they’re part of the com­mu­nity now: 102 4.9%
No: 1851 88.9%

LessWrong Usage Differ­ences Between 2016 and 2014 Surveys

How do you use Less Wrong?

I lurk, but never reg­istered an ac­count: +19.300% 1125 54.400%
I’ve reg­istered an ac­count, but never posted: −1.600% 271 13.100%
I’ve posted a com­ment, but never a top-level post: −7.600% 419 20.300%
I’ve posted in Dis­cus­sion, but not Main: −5.100% 179 8.700%
I’ve posted in Main: −3.300% 73 3.500%

About how much of the se­quences have you read?

Never knew they ex­isted un­til this mo­ment: +3.300% 217 10.400%
Knew they ex­isted, but never looked at them: +2.100% 103 4.900%
Some, but less than 25%: +3.100% 442 21.100%
About 25%: +0.400% 260 12.400%
About 50%: −0.400% 284 13.500%
About 75%: −1.800% 299 14.300%
All or al­most all: −5.000% 491 23.400%

Do you at­tend Less Wrong mee­tups?

Yes, reg­u­larly: −2.500% 160 7.700%
Yes, once or a few times: −2.100% 407 19.500%
No: +7.100% 1524 72.900%

Is phys­i­cal in­ter­ac­tion with the Less Wrong com­mu­nity oth­er­wise a part of your ev­ery­day life, for ex­am­ple do you live with other Less Wrongers, or you are close friends and fre­quently go out with them?

Yes, all the time: +0.200% 161 7.700%
Yes, some­times: −0.300% 258 12.400%
No: +2.400% 1659 79.800%

Have you ever been in a ro­man­tic re­la­tion­ship with some­one you met through the Less Wrong com­mu­nity?

Yes: +0.800% 132 6.300%
I didn’t meet them through the com­mu­nity but they’re part of the com­mu­nity now: −0.400% 102 4.900%
No: +1.600% 1858 88.800%

Write Ins

In a bit of a silly over­sight I for­got to ask sur­vey par­ti­ci­pants what was good about the com­mu­nity, so the fol­low­ing is go­ing to be a pretty one sided pic­ture. Below are the com­plete write ins re­spon­dents submitted

Is­sues With LessWrong At It’s Peak

Philo­soph­i­cal Is­sues With LessWrong At It’s Peak[Part One]
Philo­soph­i­cal Is­sues With LessWrong At It’s Peak[Part Two]
Com­mu­nity Is­sues With LessWrong At It’s Peak[Part One]
Com­mu­nity Is­sues With LessWrong At It’s Peak[Part Two]

Is­sues With LessWrong Now

Philo­soph­i­cal Is­sues With LessWrong Now[Part One]
Philo­soph­i­cal Is­sues With LessWrong Now[Part Two]
Com­mu­nity Is­sues With LessWrong Now[Part One]
Com­mu­nity Is­sues With LessWrong Now[Part Two]

Peak Philos­o­phy Is­sue Tallies

Philos­o­phy Is­sues (Sam­ple Size: 233)
La­bel Code Tally
Ar­ro­gance A 16
Bad Aes­thet­ics BA 3
Bad Norms BN 3
Bad Poli­tics BP 5
Bad Tech Plat­form BTP 1
Cul­tish C 5
Cargo Cult CC 3
Doesn’t Ac­cept Crit­i­cism DAC 3
Don’t Know Where to Start DKWS 5
Da­m­aged Me Men­tally DMM 1
Eso­teric E 3
Eliezer Yud­kowsky EY 6
Im­prop­erly In­dexed II 7
Im­pos­si­ble Mis­sion IM 4
In­suffi­cient So­cial Sup­port ISS 1
Jar­gon
Literal Cult LC 1
Lack of Ri­gor LR 14
Mis­fo­cused M 13
Mixed Bag MB 3
Noth­ing N 13
Not Enough Jar­gon NEJ 1
Not Enough Roko’s Basilisk NERB 1
Not Enough The­ory NET 1
No In­tu­ition NI 6
Not Pro­gres­sive Enough NPE 7
Nar­row Schol­ar­ship NS 20
Other O 3
Per­son­al­ity Cult PC 10
None of the Above
Quan­tum Me­chan­ics Se­quence QMS 2
Rein­ven­tion R 10
Re­jects Ex­per­tise RE 5
Spoiled S 7
Small Com­pe­tent Author­ship SCA 6
Sugges­tion For Im­prove­ment SFI 1
So­cially In­com­pe­tent SI 9
Stupid Philos­o­phy SP 4
Too Con­trar­ian TC 2
Typ­i­cal Mind TM 1
Too Much Roko’s Basilisk TMRB 1
Too Much The­ory TMT 14
Too Pro­gres­sive TP 2
Too Se­ri­ous TS 2
Un­wel­com­ing U 8

Well, those are cer­tainly some re­sults. Top an­swers are:

Nar­row Schol­ar­ship: 20
Ar­ro­gance: 16
Too Much The­ory: 14
Lack of Ri­gor: 14
Mis­fo­cused: 13
Noth­ing: 13
Rein­ven­tion (rein­vents the wheel too much): 10
Per­son­al­ity Cult: 10

So con­dens­ing a bit: Pay more at­ten­tion to main­stream schol­ar­ship and ideas, try to do bet­ter about in­tel­lec­tual rigor, be more prac­ti­cal and fo­cus on re­sults, be more hum­ble. (La­beled Dataset)

Peak Com­mu­nity Is­sue Tallies

Com­mu­nity Is­sues (Sam­ple Size: 227)
La­bel Code Tally
Ar­ro­gance A 7
As­sumes Reader Is Male ARIM 1
Bad Aes­thet­ics BA 1
Bad At PR BAP 5
Bad Norms BN 5
Bad Poli­tics BP 2
Cul­tish C 9
Cliqueish Ten­den­cies CT 1
Di­as­pora D 1
Defen­sive At­ti­tude DA 1
Doesn’t Ac­cept Crit­i­cism DAC 3
Dun­ning Kruger DK 1
Elitism E 3
Eliezer Yud­kowsky EY 2
Group­think G 11
In­suffi­ciently In­dexed II 9
Im­pos­si­ble Mis­sion IM 1
Im­poster Syn­drome IS 1
Jar­gon J 2
Lack of Ri­gor LR 1
Mixed Bag MB 1
Noth­ing N 5
??? NA 1
Not Big Enough NBE 3
Not Enough of A Cult NEAC 1
Not Enough Con­tent NEC 7
Not Enough Com­mu­nity In­fras­truc­ture NECI 10
Not Enough Mee­tups NEM 5
No Goals NG 2
Not Nerdy Enough NNE 3
None Of the Above NOA 1
Not Pro­gres­sive Enough NPE 3
Not Ra­tional NR 3
NRx (Ne­o­re­ac­tion) NRx 1
Nar­row Schol­ar­ship NS 4
Not Stringent Enough NSE 3
Parochial­ism P 1
Pickup Artistry PA 2
Per­son­al­ity Cult PC 7
Rein­ven­tion R 1
Re­cur­ring Ar­gu­ments RA 3
Re­jects Ex­per­tise RE 2
Se­quences S 2
Small Com­pe­tent Author­ship SCA 5
Sugges­tion For Im­prove­ment SFI 1
Spoiled Is­sue SI 9
So­cially INCOMpe­tent SINCOM 2
Too Bor­ing TB 1
Too Con­trar­ian TC 10
Too COM­ba­tive TCOM 4
Too Cis/​Straight/​Male TCSM 5
Too In­tol­er­ant of Cranks TIC 1
Too In­tol­er­ant of Poli­tics TIP 2
Too Long Winded TLW 2
Too Many Idiots TMI 3
Too Much Math TMM 1
Too Much The­ory TMT 12
Too Nerdy TN 6
Too Ri­gor­ous TR 1
Too Se­ri­ous TS 1
Too Tol­er­ant of Cranks TTC 1
Too Tol­er­ant of Poli­tics TTP 3
Too Tol­er­ant of POSers TTPOS 2
Too Tol­er­ant of PROGres­sivism TTPROG 2
Too Weird TW 2
Un­wel­com­ing U 12
UTILi­tar­i­anism UTIL 1

Top An­swers:

Un­wel­com­ing: 12
Too Much The­ory: 12
Group­think: 11
Not Enough Com­mu­nity In­fras­truc­ture: 10
Too Con­trar­ian: 10
In­suffi­ciently In­dexed: 9
Cul­tish: 9

Again con­dens­ing a bit: Work on be­ing less in­timi­dat­ing/​ag­gres­sive/​etc to new­com­ers, spend less time on navel gaz­ing and more time on ac­tu­ally do­ing things and col­lect­ing data, work on get­ting the struc­tures in place that will on­board peo­ple into the com­mu­nity, stop be­ing so nit­picky and ar­gu­men­ta­tive, spend more time on get­ting con­tent in­dexed in a form where peo­ple can ac­tu­ally find it, be more ac­cept­ing of out­side view­points and re­mem­ber that you’re prob­a­bly more likely to be wrong than you think. (La­beled Dataset)

One last note be­fore we finish up, these tal­lies are a very rough ex­ec­u­tive sum­mary. The tag­ging pro­cess ba­si­cally in­volves try­ing to fit points into clusters and is prone to in­ac­cu­racy through laz­i­ness, adding an­other cat­e­gory be­ing un­de­sir­able, square-peg into round-hole fit­ting, and my per­sonal poli­ti­cal bi­ases. So take these with a grain of salt, if you re­ally want to know what peo­ple wrote in my ad­vice would be to read through the write in sets I have above in HTML for­mat. If you want to eval­u­ate for your­self how well I tagged things you can see the la­beled datasets above.

I won’t bother tal­ly­ing the “is­sues now” sec­tions, all you re­ally need to know is that it’s ba­si­cally the same as the first sec­tions ex­cept with lots more “It’s dead.” com­ments and from eye­bal­ling it a higher pro­por­tion of peo­ple ar­gu­ing that LessWrong has been taken over by the left/​so­cial jus­tice and com­plaints about effec­tive al­tru­ism. (I in­fer that the com­plaints about be­ing taken over by the left are mostly refer­ring to effec­tive al­tru­ism.)

Traits Re­spon­dents Would Like To See In A Suc­ces­sor Community

Philosophically

At­ten­tion Paid To Out­side Sources
More: 1042 70.933%
Same: 414 28.182%
Less: 13 0.885%

Self Im­prove­ment Fo­cus
More: 754 50.706%
Same: 598 40.215%
Less: 135 9.079%

AI Fo­cus
More: 184 12.611%
Same: 821 56.271%
Less: 454 31.117%

Poli­ti­cal
More: 330 22.837%
Same: 770 53.287%
Less: 345 23.875%

Aca­demic/​For­mal
More: 455 31.885%
Same: 803 56.272%
Less: 169 11.843%

In sum­mary, peo­ple want a site that will en­gage with out­side ideas, ac­knowl­edge where it bor­rows from, fo­cus on prac­ti­cal self im­prove­ment, less on AI and AI risk, and tighten its aca­demic rigor. They could go ei­ther way on poli­tics but the epistemic di­rec­tion is clear.

Community

In­tense En­vi­ron­men­t
More: 254 19.644%
Same: 830 64.192%
Less: 209 16.164%

Fo­cused On ‘Real World’ Ac­tion
More: 739 53.824%
Same: 563 41.005%
Less: 71 5.171%

Ex­pert­s
More: 749 55.605%
Same: 575 42.687%
Less: 23 1.707%

Data Driven/​Test­ing Of Ideas
More: 1107 78.344%
Same: 291 20.594%
Less: 15 1.062%

So­cial
More: 583 43.507%
Same: 682 50.896%
Less: 75 5.597%

This largely backs up what I said about the pre­vi­ous re­sults. Peo­ple want a more prac­ti­cal, more ac­tive, more so­cial and more em­piri­cal LessWrong with out­side ex­per­tise and ideas brought into the fold. They could go ei­ther way on it be­ing more in­tense but the epistemic trend is still clear.

Write Ins

Di­as­pora Communities

So where did the party go? We got twice as many re­spon­dents this year as last when we opened up the sur­vey to the di­as­pora, which means that the LW com­mu­nity is al­ive and kick­ing it’s just not on LessWrong.

LessWrong
Yes: 353 11.498%
No: 1597 52.02%

LessWrong Mee­tup­s
Yes: 215 7.003%
No: 1735 56.515%

LessWrong Face­book Group
Yes: 171 5.57%
No: 1779 57.948%

LessWrong Slack­
Yes: 55 1.792%
No: 1895 61.726%

SlateS­tarCodex
Yes: 832 27.101%
No: 1118 36.417%

[SlateS­tarCodex by far has the high­est pro­por­tion of ac­tive LessWrong users, over twice that of LessWrong it­self, and more than LessWrong and Tum­blr com­bined.]

Ra­tion­al­ist Tum­blr
Yes: 350 11.401%
No: 1600 52.117%

[I’m ac­tu­ally sur­prised that Tum­blr doesn’t just beat LessWrong it­self out­right, They’re only a tenth of a per­centage point be­hind though, and if cur­rent trends con­tinue I sus­pect that by 2017 Tum­blr will have a large lead over the main LW site.]

Ra­tion­al­ist Face­book
Yes: 150 4.886%
No: 1800 58.632%

[Eliezer Yud­kowsky cur­rently re­sides here.]

Ra­tion­al­ist Twit­ter
Yes: 59 1.922%
No: 1891 61.596%

Effec­tive Altru­ism Hub
Yes: 98 3.192%
No: 1852 60.326%

FortFore­cast
Yes: 4 0.13%
No: 1946 63.388%

[I in­cluded this as a ‘troll’ op­tion to catch peo­ple who just check ev­ery box. Rel­a­tively few peo­ple seem to have done that, but hav­ing the op­tion here lets me know one way or the other.]

Good Judge­ment(TM) Open
Yes: 29 0.945%
No: 1921 62.573%

Pre­dic­tionBook
Yes: 59 1.922%
No: 1891 61.596%

Om­nilibrium
Yes: 8 0.261%
No: 1942 63.257%

Hacker News
Yes: 252 8.208%
No: 1698 55.309%

#less­wrong on freen­ode
Yes: 76 2.476%
No: 1874 61.042%

#slat­estar­codex on freen­ode
Yes: 36 1.173%
No: 1914 62.345%

#hplus­roadmap on freen­ode
Yes: 4 0.13%
No: 1946 63.388%

#chapelper­ilous on freen­ode
Yes: 10 0.326%
No: 1940 63.192%

[Since peo­ple keep ask­ing me, this is a pos­tra­tional chan­nel.]

/​r/​ra­tio­nal
Yes: 274 8.925%
No: 1676 54.593%

/​r/​HPMOR
Yes: 230 7.492%
No: 1720 56.026%

[Given that the story is long over, this is pretty im­pres­sive. I’d have ex­pected it to be dead by now.]

/​r/​SlateS­tarCodex
Yes: 244 7.948%
No: 1706 55.57%

One or more pri­vate ‘ra­tio­nal­ist’ group­s
Yes: 192 6.254%
No: 1758 57.264%

[I al­most wish I hadn’t in­cluded this op­tion, it’d have been fas­ci­nat­ing to learn more about these through write ins.]

Of all the par­ties who seem like plau­si­ble can­di­dates at the mo­ment, Scott Alexan­der seems most ca­pa­ble to un­di­as­pora the com­mu­nity. In prac­tice he’s very busy, so he would need a ded­i­cated team of rel­a­tively au­tonomous peo­ple to help him. Scott could court guest posts and start to scale up un­der the SSC brand, and I think he would fairly eas­ily end up with the li­ons share of the free float­ing LWers that way.

Be­fore I call a hearse for LessWrong, there is a glim­mer of hope left:

Would you con­sider re­join­ing LessWrong?

I never left: 668 40.6%
Yes: 557 33.8%
Yes, but only un­der cer­tain con­di­tions: 205 12.5%
No: 216 13.1%

A sig­nifi­cant frac­tion of peo­ple say they’d be in­ter­ested in an im­proved ver­sion of the site. And of course there were write ins for con­di­tions to re­join, what did peo­ple say they’d need to re­join the site?

Re­join Con­di­tion Write Ins [Part One]
Re­join Con­di­tion Write Ins [Part Two]
Re­join Con­di­tion Write Ins [Part Three]
Re­join Con­di­tion Write Ins [Part Four]
Re­join Con­di­tion Write Ins [Part Five]

Feel free to read these your­selves (they’re not long), but I’ll go ahead and sum­ma­rize: It’s all about the con­tent. Con­tent, con­tent, con­tent. No amount of us­abil­ity im­prove­ments, A/​B test­ing or clever trick­ery will let you get around con­tent. Peo­ple are over­whelm­ingly clear about this; they need a rea­son to come to the site and right now they don’t feel like they have one. That means pri­or­ity num­ber one for some­body try­ing to re­vi­tal­ize LessWrong is how you deal with this.

Let’s re­cap.

Fu­ture Im­prove­ment Wish­list Based On Sur­vey Results

Philosophical

  • Pay more at­ten­tion to main­stream schol­ar­ship and ideas.

  • Im­proved in­tel­lec­tual rigor.

  • Ac­knowl­edge sources bor­rowed from.

  • Be more prac­ti­cal and fo­cus on re­sults.

  • Be more hum­ble.

Community

  • Less in­timi­dat­ing/​ag­gres­sive/​etc to new­com­ers,

  • Struc­tures that will on­board peo­ple into the com­mu­nity.

  • Stop be­ing so nit­picky and ar­gu­men­ta­tive.

  • Spend more time on get­ting con­tent in­dexed in a form where peo­ple can ac­tu­ally find it.

  • More ac­cept­ing of out­side view­points.

While that list seems rea­son­able, it’s quite hard to put into prac­tice. Ri­gor, as the name im­plies re­quires high-effort from par­ti­ci­pants. Frankly, it’s not fun. And get­ting peo­ple to do un-fun things with­out pay­ing them is difficult. 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. Not just have peo­ple ‘dis­cuss’, as though the po­ten­tial for Ra­tion­al­ity is within all of us just wait­ing to be brought out by the right con­ver­sa­tion.

I per­son­ally haven’t been a LW reg­u­lar in a long time. As­sum­ing the points about pedan­ti­cism, snip­ping, “well ac­tu­ally”-ism and the like are true then they need to stop for the site to move for­ward. Per­son­ally, I’m a huge fan of Scott Alexan­der’s com­ment policy: All com­ments must be at least two of true, kind, or nec­es­sary.

  • True and kind—Prob­a­bly won’t drown out the dis­cus­sion sig­nal, will help sig­nifi­cantly de­crease the hos­tility of the at­mo­sphere.

  • True and nec­es­sary—Some­times what you have to say isn’t nice, but it needs to be said. This is the com­mon core of free speech ar­gu­ments for say­ing mean things and they’re not wrong. How­ever, some­thing be­ing true isn’t nec­es­sar­ily enough to make it some­thing you should say. In fact, in some situ­a­tions say­ing mean things to peo­ple en­tirely un­re­lated to their ar­gu­ments is known as the ad hominem fal­lacy.

  • Kind and nec­es­sary—The in­fa­mous ‘hug­box’ is es­sen­tially a place where peo­ple go to hear things which are kind but not nec­es­sar­ily true. I don’t think any­body wants a hug­box, but oc­ca­sion­ally it can be im­por­tant to say things that might not be true but are needed for the sake of tact, rec­on­cili­a­tion, or to pre­vent greater harm.

If peo­ple took that se­ri­ously and re­ally gave it some thought be­fore they used their key­board, I think the on-site LessWrong com­mu­nity would be a sig­nifi­cant part of the way to not driv­ing peo­ple off as soon as they ar­rive.

More im­por­tantly, in places like the LessWrong Slack I see this sort of happy go lucky at­ti­tude about site im­prove­ment. “Oh that sounds nice, we should do that.” with­out the ac­com­pa­ny­ing moun­tain of work to ac­tu­ally make ‘that’ hap­pen. I’m not sure peo­ple re­ally un­der­stand the dy­nam­ics of what it means to ‘re­vive’ a web­site in se­vere de­cay. When you de­cide to ‘re­vive’ a dy­ing site, what you’re re­ally do­ing once you’re past a cer­tain point is re­found­ing the site. So the ques­tion you should be ask­ing your­self isn’t “Can I fix the site up a bit so it isn’t quite so stale?”. It’s “Could I have founded this site?” and if the an­swer is no you should se­ri­ously ques­tion whether to make the time in­vest­ment.

Whether or not LessWrong lives to see an­other day ba­si­cally de­pends on the level of ground game its last users and ad­minis­tra­tors can muster up. And if it’s not enough, it won’t.

Vir­tus junxit mors non sep­a­ra­bit!