Frontpage Posting and Commenting Guidelines

Wel­come to the new LessWrong! Our goal with the LessWrong front­page is to host high-qual­ity dis­cus­sion on a wide range of top­ics, in a way that al­lows users to make bet­ter col­lec­tive progress to­ward the truth.

New posts au­to­mat­i­cally get posted to a user’s per­sonal blog, where peo­ple are free to talk about what­ever they like. By de­fault, mod­er­a­tors will con­sider whether the post is a good fit for the front­page. If you don’t want the post to ap­pear on front­page, you can uncheck “mod­er­a­tors can pro­mote”.

1. Things to shoot for on frontpage

1.1. Use­ful­ness, nov­elty, and fun. The front­page of this site is for se­ri­ous in­tel­lec­tual en­gage­ment with in­ter­est­ing ideas, with a fo­cus on ideas that are im­por­tant but challeng­ing to eval­u­ate. Topics that lack in­her­ent im­por­tance are OK if the dis­cus­sion qual­ity is high enough, and par­tic­u­larly if the dis­cus­sion is use­ful for other pur­poses, like build­ing skills; but the best top­ics will usu­ally be con­se­quen­tial and ne­glected ones.

1.2. Ac­cu­racy, kind­ness, and rele­vance to the dis­cus­sion at hand, in the spirit of the Vic­to­rian Sufi Bud­dha ideal.

1.3. Clar­ity and open­ness about what you be­lieve, your rea­sons for be­liev­ing it, and what would cause you to change your mind. Try to make con­crete pre­dic­tions and bets, and to note the cruxes for your be­liefs, where pos­si­ble. It’s not always easy to clearly ar­tic­u­late a be­lief, and it’s great to note places where you’re un­cer­tain about what you be­lieve, about your rea­sons, and about your cruxes. We don’t want peo­ple to feel like they have to con­ceal or im­me­di­ately aban­don their be­liefs when­ever those be­liefs turn out to be non­triv­ial to ar­tic­u­late or jus­tify. But in­cre­men­tal progress to­ward more clar­ity and open­ness, even if it’s in­com­plete, is highly val­ued here.

A corol­lary of 1.3 is that we of­ten pre­fer de­scrip­tive lan­guage (in­clud­ing lan­guage de­scribing your cur­rent be­liefs, emo­tional state, etc.) over pre­scrip­tive lan­guage, all else be­ing equal. Pre­scrip­tions are ob­vi­ously an es­sen­tial part of com­mu­ni­ca­tion, but de­scrip­tions are gen­er­ally eas­ier to re­late to ev­i­dence, pre­dic­tions, and cruxes. We en­courage putting a fo­cus on them for that rea­son.

2. Things to keep to a minimum

2.1. Com­mu­nity-fo­cused dis­cus­sion — i.e., dis­cus­sion about the LessWrong/​ra­tio­nal­ity com­mu­nity, as op­posed to dis­cus­sion about par­tic­u­lar ob­ject-level top­ics. We want to avoid dy­nam­ics like (from Feyn­man):

When I was in high school, one of the first hon­ors I got was to be made a mem­ber of the “Arista,” which was a group of kids who got good grades, hmm? Every­body wanted to be a mem­ber of the Arista, and when I got into this Arista I dis­cov­ered that what they did in their meet­ings was to sit around to dis­cuss who else was “wor­thy” to join “this won­der­ful group that we are,” okay?

If you want to dis­cuss the com­mu­nity more gen­er­ally, and you don’t ex­pect the dis­cus­sion to be of much in­ter­est to peo­ple who just want to talk about ob­ject-level is­sues (in psy­chol­ogy, or physics, or zo­ol­ogy, or cryp­tog­ra­phy, or…), it’s best to leave it in your per­sonal blog sec­tion.

Ques­tions about the site it­self are wel­come in the Meta sec­tion.

2.2. Crowd­ed­ness i.e., top­ics that are already re­ally widely dis­cussed in the pub­lic sphere, and where it will there­fore be harder to say some­thing new.

2.3 Things of fleet­ing im­por­tance — i.e., top­ics that will only be of in­ter­est for a cou­ple of weeks, like dis­cus­sions of what a poli­ti­cian has been do­ing. We want the front­page of LessWrong to serve both as a train­ing ground for as­piring ra­tio­nal­ists and as an archive of ac­cu­mu­lated col­lec­tive knowl­edge. The ideal dis­cus­sion will there­fore both help build skills and help build knowl­edge that are valuable down the line. Not ev­ery dis­cus­sion needs to achieve that ideal, but it’s a use­ful one to keep in mind.

We may build fea­tures in the fu­ture that are for more short-form and clearly ephemeral con­tent on LessWrong. If so, this will be in a new sec­tion of the site built to be less like a repos­i­tory of time­less in­for­ma­tion and dis­cus­sion, and more like (e.g.) a Face­book feed.

2.4 Hot-but­ton poli­ti­cal is­sues Highly poli­ti­cized is­sues tend to be very viral, which can of­ten lead to them dom­i­nat­ing dis­cus­sion. Th­ese is­sues of­ten (though not always) score poorly on tractabil­ity and ne­glect­ed­ness; they’re of­ten emo­tion­ally charged in ways that make con­ver­gence and skill-build­ing more challeng­ing; and dis­cus­sion is of­ten trig­gered by tran­sient news items, as op­posed to deep new in­sights that will be equally rele­vant years down the line. “Poli­tics is an im­por­tant do­main to which we should in­di­vi­d­u­ally ap­ply our ra­tio­nal­ity—but it’s a ter­rible do­main in which to learn ra­tio­nal­ity”. This means that highly poli­ti­cized is­sues will of­ten score poorly on 1.1, 2.2, and 2.3.

Of course, what counts as a “hot-but­ton poli­ti­cal is­sue” isn’t always clear, and we don’t want to en­courage ag­o­niz­ing or ar­gu­ing about what counts. (See 2.1.) We just want to en­courage users to use their judg­ment and do their best to keep it to a min­i­mum, so that other top­ics aren’t crowded out.

3. Off-limits things

3.1. Se­ri­ous vi­o­la­tions of dis­course norms — Threat­en­ing be­hav­ior, need­lessly harsh per­sonal at­tacks, ha­rass­ment, doxxing, and so on.

3.2. Con­sis­tently dis­rup­tive or low-qual­ity con­tent — Spam, dis­cus­sion de­railing, and so on.

A list of users with bans or pub­lic warn­ings can be found here.

4. How mod­er­a­tion works

Com­pared to mod­er­a­tors on other on­line fo­rums, mod­er­a­tors on LessWrong are granted greater abil­ity to change and im­prove the web­site, and are trusted with more in­for­ma­tion. Th­ese roles of re­spon­si­bil­ity are only given to trusted mem­bers of the com­mu­nity, and they are known as the Sun­sh­ine Reg­i­ment.

The new, weighted karma sys­tem is de­signed to bring good con­tent to the top. How­ever, this karma sys­tem is based on the vot­ing pat­terns of many in­di­vi­d­u­als, most of whom do not have the time to re­flect on big-pic­ture trends, nor the re­sources to sub­stan­tially change those trends. In a clas­sic tragedy of the com­mons, when there are thou­sands of peo­ple vot­ing, no in­di­vi­d­ual is in­cen­tivised to spend a lot of time con­sid­er­ing their vote.

The in­cen­tives set up by the karma sys­tem can be con­sid­ered the com­mu­nity’s Sys­tem 1, and the Sun­sh­ine Reg­i­ment can be thought of as the com­mu­nity’s Sys­tem 2. Sun­sh­ines think about what in­cen­tive gra­di­ents are be­ing pro­duced, and are given the re­sources to in­fluence the in­cen­tive gra­di­ents in a more sub­stan­tial way (e.g. karma re­wards on com­ments), al­low­ing the com­mu­nity to plan around ob­sta­cles and achieve more com­plex goals.

There are no hard rules about what com­ments each mem­ber of the Sun­sh­ine Reg­i­ment will give karma re­wards to. If your sub­mis­sion has re­ceived a karma re­ward, it will be sig­nified by a small star icon on that com­ment or post. If your sub­mis­sion has been re­moved by a Sun­sh­ine, they will leave a note ex­plain­ing why the com­ment was in­ap­pro­pri­ate or un­suited to the LessWrong front­page.

Mem­bers of the Sun­sh­ine Reg­i­ment will have ac­cess to more in­for­ma­tion than other users, al­low­ing them to no­tice nega­tive pat­terns of be­havi­ours, such as sock­pup­pet ac­counts and mass down­vot­ing. The ex­tra in­for­ma­tion is:

  • Ac­cess to the iden­tities of vot­ers on any com­ment/​post, and to the vot­ing his­tory of all users.

  • The IP ad­dress a user wrote a post or com­ment from.

Sun­sh­ines of the 1st LessWrong Reg­i­ment are: