Ruby

• Ruby 21 Oct 2020 4:49 UTC

  9 points

  in reply to: Kenny’s comment on: LessWrong FAQ

  Sorry for the de­layed re­ply yet again, my bad.

  Thanks for dig­ging into the T&C, your anal­y­sis seems cor­rect to me. I think we do want users to re­tain their rights – more than once peo­ple have been re­luc­tant to post if might cost them their rights and other pub­lish­ing op­por­tu­ni­ties. I think it’d be very costly to make writ­ing seem costlier to au­thors.

  Re­gard­ing an API end­point – we have one! See here for a tu­to­rial. The main thing is that if you wanted to archive ev­ery­thing, I’m not sure it’s so great if you wanted to get com­ments too. Pul­ling all the posts and all the com­ments and as­sem­bling cor­rectly would be a has­sle. I think you’re bet­ter try­ing to get archive via In­ter­net Archive or similar. Deltas would in­deed be tricky. But I as­sure you, we are not go­ing to let the site data get lost!

• Ruby 20 Oct 2020 19:39 UTC

  14 points

  on: Launch­ing Fore­cast, a com­mu­nity for crowd­sourced pre­dic­tions from Facebook

  In­ter­est­ingly, Pre­dic­tion Mar­kets for In­ter­net Points was all about set­ting up pre­dic­tion mar­kets on Face­book with in­ter­net points. I’m cu­ri­ous whether Paul would think the cur­rent app to fulfill what he was imag­in­ing there.

• Ruby 20 Oct 2020 19:34 UTC

  5 points

  on: Launch­ing Fore­cast, a com­mu­nity for crowd­sourced pre­dic­tions from Facebook

  Very neat! I had a look at the desk­top web­site. On the whole, pretty good. I could imag­ine it fa­cil­i­tat­ing some good dis­cus­sion around pre­dic­tions, though seems like it’s just get­ting started.

  Cur­rently, pre­dic­tions are made with a slider, com­ing from a LessWrong back­ground, I’d pre­fer if the slider had in­di­ca­tions of prob­a­bil­ity (50%, 75%, 82%) the cur­rent thing has me guess­ing and doesn’t show me un­til I con­firm. A chance to re­view be­fore con­firm­ing would be good.
  
  On the “Rea­son” cards, it says “10 peo­ple”, “8 peo­ple”, but doesn’t in­di­cate what that means. I’m guess­ing it’s num­ber who en­dorsed the rea­son.

  --

  A big­ger topic I’m cu­ri­ous in is how this re­lates to Face­book over­all. I can’t imag­ine very many FB users be­ing in­ter­ested in pre­dic­tion mar­kets. (NPE’s other listed pro­jects of mu­sic col­lab­o­ra­tion and sports ex­pe­riences would have much broader ap­peal, I’d think.) So I’m cu­ri­ous what the vi­sion is and how Face­book ends up mo­ti­vated to host this kind of pro­ject.

• Ruby 15 Oct 2020 5:10 UTC

  5 points

  in reply to: Kenny’s comment on: LessWrong FAQ

  Ah, I re­gret that the par­ent com­ment here slipped my no­tice.
  
  There’s no prob­lem with you archiv­ing the site for your per­sonal/​pri­vate use. Peo­ple have made var­i­ous epubs and mobi’s of the Se­quences and other posts at times. Ul­ti­mately we want peo­ple to read the con­tent re­gard­less of form.
  
  Copy­right stuff is in­cluded in the Terms of Use doc­u­ment linked at the end of the FAQ. I’ll check with oth­ers be­fore try­ing to pub­li­cly in­ter­pret the le­gal mean­ing.

• Ruby 15 Oct 2020 5:06 UTC

  5 points

  in reply to: Kenny’s comment on: The LessWrong Team

  I’m glad you like the site! You’re wel­come.
  
  I’m cu­ri­ous though, what do you mean by “for­mal­ized”?

• Ruby 6 Oct 2020 21:11 UTC

  2 points

  on: Incredulity

  Talk:Incredulity

  I don’t re­ally see a need for this page. Delete? Redi­rect to Real­ity is nor­mal? Redi­rect to Ab­sur­dity heuris­tic? --Zack M. Davis 20:06, 17 Novem­ber 2009 (UTC)

• Ruby 6 Oct 2020 20:50 UTC

  2 points

  on: Guess­ing The Teacher’s Password

  From the old dis­cus­sion page:

  Talk:Guess­ing the teacher’s password

  On du­pli­cat­ing blog con­tent verbatim

  If some text on the wiki is hor­rible, then of course that’s im­por­tant and needs to be fixed, but it seems like a bad policy to have pages that are strictly ver­ba­tim quotes from Eliezer-posts. The pur­pose of the wiki is to sum­ma­rize key con­cepts from the OB/​LW canon, so that we have some­thing to show other than, “Go read ev­ery­thing Eliezer’s ever writ­ten.” If we have to du­pli­cate too much blog com­ment ver­ba­tim, there’s not much point in hav­ing a wiki. Hence my re­cent rol­lback of Peer­in­finity’s rol­lback. Zack M. Davis 02:34, 23 Oc­to­ber 2009 (UTC)

  But what if Eliezer’s words already are the best sum­mary? Why change them just for the sake of not quot­ing him di­rectly? --PeerIn­finity 03:21, 23 Oc­to­ber 2009 (UTC)

  For the same rea­son that “guess­ing the teacher’s pass­word” is a prob­lem in con­ven­tional schools. The ideas are what’s im­por­tant, and the ideas shouldn’t be so de­pen­dent on any one canon­i­cal phras­ing. I would guess that hit­ting the same ma­te­rial from slightly differ­ent an­gles as­sists in learn­ing it deeply. Fur­ther­more, wiki ar­ti­cles should be con­cise, leav­ing the ex­tended ex­po­si­tion to the full blog­posts. --Zack M. Davis 04:11, 23 Oc­to­ber 2009 (UTC)

• Ruby 6 Oct 2020 20:42 UTC

  2 points

  on: In­finite Set Atheism

  From the old dis­cus­sion page:

  Talk:In­finite set atheism

  I’m cu­ri­ous as to why this is an in­ter­est­ing sub­ject. Afaik, in­finity only ever ex­ists as an ab­strac­tion, and the ex­is­tence is an ax­iom of ZFC; which is to say you’re just say­ing “in­finity ex­ists!” and cross­ing your arms and hop­ing it doesn’t ex­plode. So my point here is, who in the world is both math­e­mat­i­cally liter­ate, and not an “in­finite set athe­ist” to the ex­tent that they re­fer to in­finity as some­thing other than an ab­strac­tion? Magfrump 20:53, 6 Jan­uary 2010 (UTC)

  I’m given to un­der­stand that many take the pos­si­bil­ity of an in­finite uni­verse se­ri­ously. Cf. Bostrom’s “In­finite Ethics” (PDF) Zack M. Davis 14:25, 7 Jan­uary 2010 (UTC)

  Yuck. This whole main para­graph is based on the mis­con­cep­tion that car­di­nal­ity is the only mea­sure of size. I want to fix this but I don’t see how with­out kil­ling that whole para­graph. Any­way, do we re­ally need to call this “in­finite set athe­ism” just be­cause Eliezer Yud­kowsky calls him­self an “in­finite set athe­ist”? Can’t we just say “fini­tism”? Look­ing it up, I guess this is more of what’s called “strict fini­tism”—but this isn’t re­ally a sub­ject I know… Sniffnoy 13:35, 7 Jan­uary 2010 (UTC)

  If you think the para­graph is bet­ter off dead, then by all means kill it: wiki is the medium in which noth­ing is sa­cred. Although this does seem patch­able—some­thing like, one gen­er­al­iza­tion of the “size” of the set is called the car­di­nal­ity. Un­tu­tored in­tu­ition (&c., &c.).” On re­flec­tion I am in­clined to agree with you that this should be moved to fini­tism or such­like. (We already have enough in­sider jar­gon, best not to cre­ate more un­nec­es­sar­ily.) I might do this a lit­tle bit later when I’ve looked up com­mon us­age of fini­tism. Zack M. Davis 14:25, 7 Jan­uary 2010 (UTC)

• Ruby 6 Oct 2020 20:39 UTC

  2 points

  on: Meth­ods of Ra­tion­al­ity (Fan­fic­tion)

  From the old dis­cus­sion page:

  Talk: Meth­ods Of Ra­tion­al­ity (fan­fic­tion)

  I’m not sure ‘pos­tu­lates’ is the right word in ”… which pos­tu­lates that the or­phaned Harry Pot­ter has been raised by a sci­en­tist step­father …”. Thomblake 20:36, 11 May 2010 (UTC)

  Any rea­son to use non-stan­dard nam­ing scheme for this wiki? --Vladimir Nesov 22:32, 11 May 2010 (UTC)

  Edited to use the stan­dard nam­ing scheme. --Vladimir Nesov 16:03, 13 May 2010 (UTC)

  Be­cause I want to dump a lot of au­thor’s notes and so on here, linked to from the pro­file. Long awk­ward links would be un­helpful. Non­stan­dard name is fine, it’s a non­stan­dard use of the wiki and shouldn’t re­ally be linked from any­where else. Please don’t move it again? --Eliezer Yud­kowsky 21:37, 13 May 2010 (UTC)

  (Copied my re­ply from a pri­vate dis­cus­sion.)

  The only down­side seems to be a few ex­tra spaces in links, but I don’t see how it’s par­tic­u­larly im­por­tant, given that the length of the origi­nal links is already big enough, so any com­mu­ni­ca­tion of the links is go­ing to be via copy-paste, iden­ti­cally in both cases. So the dis­agree­ment is purely aes­thet­i­cal, in which case con­sis­tency should win ;-)

  If you want to move the pages out of the do­main of reg­u­lar LessWrong wiki, you could cre­ate them un­der your user page. Hav­ing a page for the fic on Less Wrong wiki seems ap­pro­pri­ate though.--Vladimir Nesov 15:00, 27 May 2010 (UTC)

  This wiki would also be ap­pro­pri­ate for archiv­ing au­thor’s notes, so that peo­ple who can’t/​don’t check of­ten enough can still read them. Also, would you con­sider archiv­ing the differ­ent ver­sions of chap­ters that you post? Sev­eral times you have men­tioned al­ter­nate ver­sions in the au­thor’s notes, and this would be ap­pro­pri­ate to show the differ­ences to any in­ter­ested read­ers. --Dy­namic 15:26, 12 May 2010 (UTC)

  Com­ment thread list

  I just made some re­vi­sions to the Blog Posts sec­tion that I think en­hance us­abil­ity. For­ever now I’ve been frus­trated with this page, be­cause I saw the long list of blog posts and thought it was al­most always one ver­sion be­hind. Ac­tu­ally I was just not look­ing care­fully enough to see that the most re­cent was in its own para­graph above the list. So I re­versed the list or­der and put the cur­rent one as a mem­ber of the same list, so fu­ture peo­ple don’t make the same mis­take I did. Hope this is okay. --75th (talk) 11:33, 31 Jan­uary 2013 (EST)

  Any­one know who’s main­tain­ing the hp­mor site? there’s a stale link for one of the celebrity en­dorse­ments (http://​​www.wired.com/​​geek­mom/​​2011/​​07/​​harry-pot­ter-and-the-meth­ods-of-ra­tio­nal­ity/​​ is a valid al­ter­na­tive)

• Ruby 6 Oct 2020 20:33 UTC

  2 points

  on: Less­wrong:Aboutpage

  From the old dis­cus­sion page:

  Talk:Less­wrong:Aboutpage

  I’ve been try­ing to im­prove this about page in line with the sug­ges­tions I made in my re­cent dis­cus­sion post, but User:Grog­nor has been re­vert­ing my ed­its. I’d like to work with Grog­nor to find a ver­sion of the page we both agree on. It would be great to hear other users weigh in on this is­sue. I don’t see the down­side of mak­ing my sug­gested changes; even if Grog­nor does not think cultish ap­pear­ances are a prob­lem, oth­ers like me do, and it seems like we ought to be able to find a way to write the about page that avoids them and is just as effec­tive in other re­spects as the old about page.--John Maxwell IV 23:43, 15 March 2012 (UTC)

  Please stop talk­ing about me like I’m some kind of malev­olent force that only wants to crush your dreams. Grog­nor 05:47, 16 March 2012 (UTC)

  I re­moved most of my refer­ences to you.--John Maxwell IV 06:50, 16 March 2012 (UTC)

  Man, let’s just chill out and re­lax. The name of the game is Bold, Rev­ert, Dis­cuss, brother. You and Groggy need to talk about your feel­ings for a while. Then we can come to a con­sen­sus about the about page. Paper-ma­chine 06:08, 16 March 2012 (UTC)

  pa­per ma­chine, what do you like about the virtues so much? If you look at this thread I linked to, there are sev­eral peo­ple ex­press­ing am­bivalence about them. If you just want us to have a long about page, do you think you could re­place the virtues with some other con­tent? Also, wedrifed said he never liked the virtues page, with a cough… What else is that sup­posed to mean?

  I have a lot of ideas for this page and I’d re­ally like to re­think it from the ground up, but I’m not go­ing to in­vest the time in that if even sim­ple, ap­par­ently widely sup­ported changes like this are re­verted. Less wrong has a ton of great stuff in its archives, and I don’t see any rea­son to im­me­di­ately show­case a con­tro­ver­sial piece. I don’t ob­ject to link­ing to the virtues in a non-cultish way, but re­pro­duc­ing them en­tirely doesn’t seem like the best use of space. Although re­ally, ideally we would not men­tion them at all since they are not hosted on Less Wrong. This is the about page for Less Wrong, so we should show­case things that are writ­ten here. There are plenty of in­tro­duc­tory pieces on this do­main.--John Maxwell IV 06:08, 17 March 2012 (UTC)

  I strongly dis­agree that we know at this point that the change is “ap­par­ently widely sup­ported”. On the other hand, I’m merely am­biva­lent about ac­tu­ally re­mov­ing the sec­tion. --Vladimir Nesov 10:33, 18 March 2012 (UTC)

  I agree with JM4 on non-in­clu­sion of the 12 Virtues. They are: 1. awe­some; 2. un­for­tu­nately weird in this con­text (their weird­ness works in other con­texts); 3. in­con­ve­niently long in this con­text. --Matt 05:58, 18 March 2012 (UTC)

  I agree with this ar­gu­ment. Paper-ma­chine 19:40, 18 March 2012 (UTC)

  JM4-- I have no per­sonal stake in whether or not the virtues ap­pear on the about page. I have no per­sonal stake in hav­ing “a long about page.” My prob­lem is with your mishan­dling of con­sen­sus. Cur­rently I count two peo­ple in that thread that are “ex­press­ing am­bivalence” about the virtues; that is not “sev­eral peo­ple.” Say­ing “sev­eral peo­ple”, and then say­ing “also, wedrifed (sic) said,” dou­ble-counts ev­i­dence. My goals are 1) to diffuse an edit war and 2) build con­sen­sus. Given Matt’s points above, I cur­rently agree that it’s rea­son­able to re­move the twelve virtues from the about page. Paper-ma­chine 19:53, 18 March 2012 (UTC)

  You’re right that changes to this page should be taken se­ri­ously, and that I pre­ma­turely de­clared the con­flict de­cided. Let me know if you think suffi­cient con­sen­sus has been built to re­move the 12 virtues—I’m a wiki am­a­teur.--John Maxwell IV 00:17, 20 March 2012 (UTC)

  How would you guys feel about a “best less wrong new­comer ex­pe­rience” con­test where en­trants would rewrite this and Less­wrong:Homepage? I re­ally think they could use ad­di­tional op­ti­miza­tion pres­sure, es­pe­cially given that there is con­tent du­pli­cated be­tween the two pages. They are some of the most viewed pages on the site.--John Maxwell IV 06:42, 20 March 2012 (UTC)

  Some data on land­ing page traf­fic: http://​​less­wrong.com/​​lw/​​at9/​​what_if_the_front_page/​​611q—Matt 05:40, 21 March 2012 (UTC)

• Ruby 6 Oct 2020 20:32 UTC

  2 points

  on: In­ter­view Series On Risks From AI

  From the old dis­cus­sion page:

  Talk:In­ter­view se­ries on risks from AI

  Is there a more gen­eral com­pendium of is­sue/​po­si­tion/​ar­gu­ment on this is­sue?

  Some­thing that all users could con­tribute to?

  Seems very much has been said on this is­sue and who cares about a limited sum­mary of the re­sults of only one se­ries of in­ter­views /​ one sur­vey.

  Care has been taken with re­spect to the ques­tions asked.

  At two oc­ca­sions I asked the LessWrong com­mu­nity to re­fine the ques­tions:

  1. http://​​less­wrong.com/​​lw/​​65a/​​help_writ­ing_mar­vin_min­sky/​​

  2. http://​​less­wrong.com/​​lw/​​9as/​​tem­plate_ques­tions_re­gard­ing_pos­si­ble_risks_from/​​

  I changed the ques­tions on both oc­ca­sions. Even though I did not agree with all of the sug­ges­tions, there is no in­di­ca­tion that I did not take care with re­spect to what ques­tions to ask.

  The ques­tions were both neu­tral and to the point. —Pre­ced­ing un­signed com­ment added by XiXiDu (talk • con­tribs)

• Ruby 6 Oct 2020 20:29 UTC

  2 points

  on: Computronium

  Talk:Computronium

  Since this con­cept is on wikipe­dia, and doesn’t add a new per­spec­tive or lo­cal cita­tions, it should prob­a­bly stay out of the wiki. --Vladimir Nesov 19:12, 8 June 2010 (UTC)

  Also, if peo­ple are go­ing to copy from Wikipe­dia, we need to be care­ful to pre­serve the CC li­cense. JoshuaZ 13:43, 9 June 2010 (UTC)

• Ruby 5 Oct 2020 19:47 UTC

  2 points

  on: Pro­gram­ming Resources

  For­mat­ting could prob­a­bly be made nicer, but there’s a lot and I don’t know this is su­per used cur­rently.

• Ruby 3 Oct 2020 2:39 UTC

  2 points

  on: Friendly Ar­tifi­cial Intelligence

  aaa

• Ruby 3 Oct 2020 2:39 UTC

  2 points

  on: Friendly Ar­tifi­cial Intelligence

  Talk:Friendly ar­tifi­cial intelligence

  Contents

  [hide]

  • 1Ac­cessibil­ity of the article

  • 2Merge FAI and uFAI ar­ti­cles?

  • 3unFriendly vs. unfriendly

  • 4FAI as “hav­ing a pos­i­tive rather than nega­tive effect”?

  Ac­cessibil­ity of the article

  I think this ar­ti­cle should also work as a first in­tro­duc­tion to the con­cept (refer­ring to the nec­es­sary ex­ter­nal doc­u­ments per­haps), so that one can put a link to it on the web, clar­ify­ing the use of the con­cept. Presently it’s quite opaque, and dis­cusses only some ar­cane stuff while tak­ing too much un­der­stand­ing for granted. --Vladimir Nesov 23:20, 4 Novem­ber 2009 (UTC)

  Merge FAI and uFAI ar­ti­cles?

  This is to sug­gest that the Friendly ar­tifi­cial in­tel­li­gence and Un­friendly ar­tifi­cial in­tel­li­gence pages be merged into one ar­ti­cle (per­haps Friendly and unFriendly ar­tifi­cial in­tel­li­gence), the ra­tio­nale here be­ing that the two con­cepts are just too closely linked (by a nega­tion, in fact) to de­serve sep­a­rate pages. It’s hard to ex­plain FAI with­out at the same time ex­plain­ing why uFAI would be bad. Any thoughts, or should I just do it? --Zack M. Davis 02:17, 30 Oc­to­ber 2009 (UTC)

  I recom­mend merg­ing the ar­ti­cles but leav­ing a redi­rec­tion from Un­friendly AI. --Wedrifid 04:51, 30 Oc­to­ber 2009 (UTC)

  The con­cepts an­swer differ­ent ques­tions: FAI ar­ti­cle should talk about what it means and what it takes to make AI Friendly, while UFAI ar­ti­cle about why there is the dan­ger in ar­bi­trary AIs to be­gin with. There seems to be lit­tle over­lap in these con­cerns. Cur­rently, some dis­cus­sion that should go into UFAI is in Paper­clip max­i­mizer. I’m restor­ing the UFAI ar­ti­cle, but let’s see it its con­cept can be made an ex­plicit topic in FAI ar­ti­cle. --Vladimir Nesov 12:55, 31 Oc­to­ber 2009 (UTC)

  unFriendly vs. unfriendly

  I also sug­gest that “unFriendly AI” be con­sis­tently ren­dered with the cap­i­tal F, as Friendli­ness is be­ing used as tech­ni­cal term dis­tinct from or­di­nary hu­man friendli­ness. Com­pare the prece­dent. --Zack M. Davis 02:17, 30 Oc­to­ber 2009 (UTC)

  Con­sis­tent cap­i­tal­i­sa­tion is ap­pro­pri­ate. I Suggest that the de­ci­sion is be­tween ’Un­friendly and unFriendly. --wedrifid

  The dis­tinc­tion be­tween cap­i­tal-F Friendli­ness and the dic­tio­nary word is already un­usual enough, funny cap­i­tal­iza­tion seems a lit­tle too much, even though his­tor­i­cally it’s pop­u­lar. --Vladimir Nesov 13:26, 31 Oc­to­ber 2009 (UTC)

  Eliezer seems to use unFriendly con­sis­tently; maybe we should con­sult him? --Zack M. Davis 18:40, 31 Oc­to­ber 2009 (UTC)

  But it’s not an ac­tual prob­lem worth car­ing about, is it? There just can’t be a strong rea­son one way or the other. --Vladimir Nesov 20:24, 31 Oc­to­ber 2009 (UTC)

  It seems non­triv­ial to me, al­though I could just be un­usu­ally (over)sen­si­tive to this sort of thing. --Zack M. Davis 03:46, 1 Novem­ber 2009 (UTC)

  Us­ing ‘un­friendly’ in the place of a cap­i­tal­ised var­i­ant con­veys an er­ror in un­der­stand­ing of the kind that could well leave hu­man­ity ex­tinct. Con­fus­ing uFAI with the literal de­scrip­tion ‘un­friendly’ is an er­ror along the lines that en­courage molec­u­lar smiley face dis­asters. ‘unFriendly’ at least con­veys that we are refer­ring to any su­per-in­tel­li­gence not meet­ing the strict stan­dard we la­bel Friendly. It’s definitely non­triv­ial. --Wedrifid 07:05, 2 Novem­ber 2009 (UTC)

  Right, so I men­tioned this to Eliezer at the re­cent meetup. If I re­call cor­rectly, he said he fa­vored the cap­i­tal-F (unFriendly or UnFriendly) but that Michael Vas­sar had the fi­nal call if there was still dis­agree­ment. For now, I’m go­ing to be us­ing unFriendly (UnFriendly at the be­gin­ning of a sen­tence). --Zack M. Davis 22:20, 10 Novem­ber 2009 (UTC)

  FAI as “hav­ing a pos­i­tive rather than nega­tive effect”?

  A Friendly Ar­tifi­cial In­tel­li­gence (FAI) is an ar­tifi­cial gen­eral in­tel­li­gence that has a pos­i­tive rather than nega­tive effect on hu­man­ity.

  This first sen­tence seems in­ad­e­quate, since it im­plies that hav­ing a pos­i­tive effect on hu­man­ity is a suffi­cient con­di­tion for FAI. Surely we don’t only re­quire that it has pos­i­tive effects on hu­man­ity, but that the pos­i­tive effects were ex­pected with very high prob­a­bil­ity by virtue of its de­sign. The sen­tence above makes it sound as if FAI is noth­ing more than see­ing whether the effects were pos­i­tive or nega­tive, as if Friendli­ness is syn­ony­mous with “benefit­ing hu­man­ity”, re­gard­less of why hu­man­ity was benefited and how eas­ily it might have turned out differ­ently.--Anonym 01:54, 1 Novem­ber 2009 (UTC)

  I agree; I’ll edit the page. --Zack M. Davis 03:38, 1 Novem­ber 2009 (UTC)

  Vladimir_Nesov re­verted this back to the origi­nal with the com­ment, “that you can’t get there by luck is a con­clu­sion down the road, not defi­ni­tion.” Con­sider an AI whose first ac­tion is to flip a fair coin and then de­stroy hu­man­ity if the coin lands heads, or solve an out­stand­ing math­e­mat­i­cal con­jec­ture if tails. After ei­ther de­stroy­ing hu­man­ity or solv­ing a con­jec­ture, it de­stroys it­self. Un­der Vladimir’s fa­vored defi­ni­tion, if the coin lands tails, it is an FAI, and if it lands heads, well, it’s not an FAI and we no longer ex­ist. Am I the only one who thinks this is silly, and that if such a thing is con­sid­ered a FAI even though it was just as likely to de­stroy us, there is no point in even us­ing the term? --Anonym 03:18, 8 May 2010 (UTC)

  AI-af­ter-it-de­cided-to-do-good is good, AI-af­ter-it-de­cided-to-do-evil is bad, and AI-be­fore-it-flipped-the-coin, as in your de­scrip­tion, is on the net bad. You judge a sys­tem as whole, based on what you ex­pect from it, not in ret­ro­spect, based on what ac­tu­ally did hap­pen. What will ac­tu­ally hap­pen is not a prop­erty of our sys­tem, and so can’t be used to cat­e­go­rize it as “Friendly” or not. (Of course, “solv­ing a math­e­mat­i­cal con­jec­ture” is nowhere near an op­ti­mal thing ac­cord­ing to hu­man val­ues to do with the world, so shouldn’t be seen as Friendly, and “pos­i­tive effect on hu­man­ity” is a sim­plifi­ca­tion.) --Vladimir Nesov 14:20, 8 May 2010 (UTC)

  You make my point for me. I said that it’s not just the effect that de­ter­mines whether it’s an FAI or an uFAI. I said that we have to take into ac­count the prob­a­bil­ity of fu­ture harm or benefit based on what know of the sys­tem, and you state ba­si­cally the same thing when you say “You judge a sys­tem as whole, based on what you ex­pect from it, not in ret­ro­spect, based on what ac­tu­ally did hap­pen”. If we judge based on what we ex­pect, then we cer­tainly do not just con­sider “hav­ing a pos­i­tive rather than nega­tive effect”, which is my point en­tirely. --Anonym 03:29, 14 May 2010 (UTC)

  As I said, sim­plifi­ca­tion. It’s the same is­sue as with “ra­tio­nal­ists win”. If you know of a way to im­prove the word­ing, go ahead, but if it makes the point harder to get across, the cur­rent ver­sion would be prefer­able. --Vladimir Nesov 11:12, 14 May 2010 (UTC)

• Ruby 3 Oct 2020 2:25 UTC

  2 points

  on: Sequences

  From the old dis­cus­sion page:

  Talk:Sequences

  The Se­quences page is prob­a­bly the most im­por­tant page on the Wiki. As such, speed of user ex­pe­rience is more im­por­tant than a vague urge for ab­stract sym­me­try or con­sis­tency. Where the origi­nal se­quence guides are blog posts, we want users—es­pe­cially new users—to visit those origi­nal se­quence guides im­me­di­ately. We don’t want to send them to a Wiki page that for­wards to the se­quence guide and force them to click twice or read the same con­tent over again. User in­ter­face stud­ies show that re­quiring one more click re­sults in a sig­nifi­cant drop-off in par­ti­ci­pa­tion, and this is very de­bil­i­tat­ing when it comes to the Se­quences page. We want the user to click and see some­thing in­ter­est­ing and at­trac­tive as fast as pos­si­ble.

  Hence I’m rol­ling back var­i­ous ed­its that in­crease the num­ber of user clicks. I de­signed the page the way it is for a rea­son. --Eliezer Yud­kowsky 18:43, 10 De­cem­ber 2009 (UTC)

  
  Eliezer, that’s caus­ing prob­lems. See the is­sue I’ve just moved from #215:

  I’ve no­ticed that not all se­quences have a page on the wiki. E.g. http://​​wiki.less­wrong.com/​​wiki/​​Se­quences#The_Quan­tum_Physics_Se­quence links to http://​​less­wrong.com/​​lw/​​r5/​​the_quan­tum_physics_se­quence/​​ in­stead of a wiki page. I as­sume this is be­cause the LW post already lists the ar­ti­cles in the se­quence. This is a prob­lem though as the ar­ti­cle nav­i­ga­tion code as­sumes there is a wiki page for each se­quence and uses that to build the list of ar­ti­cles in the se­quence. The re­sult is ar­ti­cles in se­quences with­out a wiki page don’t have the se­quence listed in the ar­ti­cle nav­i­ga­tion links. E.g. http://​​less­wrong.com/​​lw/​​r9/​​quan­tum_me­chan­ics_and_per­sonal_iden­tity/​​

  --Matt 04:04, 13 Au­gust 2010 (UTC)

  As I’ve said on the tracker, it’s not re­ally so:

  There is a wiki page for quan­tum physics se­quence, it’s just not linked from the main se­quences page, be­cause the post that is linked has ab­stracts, while the wiki page doesn’t (and prob­a­bly shouldn’t, since it’ll be just a copy-paste of what’s already available). Here it is: http://​​wiki.less­wrong.com/​​wiki/​​The_Quan­tum_Physics_Sequence

  The same goes for fun the­ory se­quence: http://​​wiki.less­wrong.com/​​wiki/​​The_Fun_The­ory_Sequence

  --Vladimir Nesov 08:26, 13 Au­gust 2010 (UTC)

  Wording

  “Map and Ter­ri­tory con­tains some of the most im­por­tant in­tro­duc­tory posts and es­says.

  “If you don’t read the se­quences on Mys­te­ri­ous An­swers to Mys­te­ri­ous Ques­tions and Re­duc­tion­ism, noth­ing else on Less Wrong will make much sense.”

  This has bugged me some­what since I first came here. “Map and Ter­ri­tory is ‘in­tro­duc­tory’, sug­gest­ing that I’m best off read­ing it first, but at the same time it won’t make much sense with­out MAMQ, but I should read it first for some rea­son any­way?”. --Doc­u­ment 04:22, 6 Jan­uary 2011 (UTC)

  Alter­na­tive formats

  It wasn’t til I’d already added links un­der each sec­tion head­ing that I no­ticed the “Alter­na­tive For­mats” sec­tion at the bot­tom and dis­cov­ered OneWhoFrogs’s ap­par­ently more com­plete col­lec­tion; sorry about that. I’d try to clean things up, but I’m up late and I’m hop­ing it’s still closer to ideal ac­cessibil­ity than be­fore than be­fore. (Edit: an­other thing that should be done is mak­ing links to sum­mary posts more promi­nent, and pos­si­bly link­ing this.) --Doc­u­ment 08:18, 6 Jan­uary 2011 (UTC)

  I’m re­vert­ing all your ed­its, as they seem to be re­dun­dant (cor­rect me if there’s some ma­te­rial not in­cluded in the last sec­tion, I haven’t ac­tu­ally checked; in that case, it should be added to the last sec­tion), and dis­tort struc­ture of the page (by in­ter­rupt­ing the text that de­scribes the se­quences). --Vladimir Nesov 13:58, 16 Jan­uary 2011 (UTC)

  Chronolog­i­cal Format

  http://​​yud­kowsky.net/​​ra­tio­nal/​​over­com­ing-bias notes that the posts may make more sense in chronolog­i­cal for­mat, and for me, it seems like http://​​www.cs.auck­land.ac.nz/​​~and­whay/​​postlist.html offers a bet­ter start­ing point than the “map and ter­ri­tory” se­quence (ie, just a bit more ex­cit­ing to start read­ing). So, I sug­gest that this be men­tioned as an al­ter­na­tive in the wiki page, pos­si­bly in the in­tro­duc­tion af­ter the bit about “the most sys­tem­atic way...”

  Com­mu­ni­ca­tion sequence

  Work in progress: User:Chriswa­ter­guy #A com­mu­ni­ca­tion se­quence.

  I don’t know whether it should even­tu­ally be linked here, but it’s pos­si­bly of in­ter­est to peo­ple be­sides my­self. --Chriswa­ter­guy (talk) 11:44, 8 De­cem­ber 2014 (AEDT)

• Ruby 3 Oct 2020 2:22 UTC

  2 points

  on: Less Wrong/​Ar­ti­cle Summaries

  Talk:Less Wrong/​Ar­ti­cle summaries

  < Talk:Less Wrong

  I don’t find the yearly cat­e­go­riza­tion of posts use­ful. I think it’d be use­ful to or­ga­nize them, eg. by Ra­tion­al­ity: From AI to Zom­bies’ table of con­tents.

  I was think­ing of cre­at­ing a sep­a­rate ar­ti­cle that does so, but then I re­al­ize it’d in­tro­duce a du­pli­ca­tion—the sum­mary text would be lo­cated in the new ar­ti­cle, and in the eg. 2006_Ar­ti­cles/​Sum­maries ar­ti­cle. The eas­iest way of deal­ing with that is to just have one ar­ti­cle sum­maries page.

  What do you think?

  - Adam Zerner, 4/​27/​17

• Ruby 3 Oct 2020 2:21 UTC

  2 points

  on: Joy In Discovery

  Talk:Joy in discovery

  Dis­cov­ery isn’t all about know­ing stuff about the world out there. A lot is in ob­tain­ing con­cep­tual tools to look at the world. The de­scrip­tions of these tools need to get into one’s head to get one started, but not nec­es­sar­ily the info about the world that is en­abled by these tools. Hence “ev­i­den­tial en­tan­gle­ment” is a wrong as­so­ci­a­tion (btw, a bet­ter link for that phrase, if it was to put in the ar­ti­cle in the first place, is Beliefs re­quire ob­ser­va­tions). --Vladimir Nesov 12:27, 31 Oc­to­ber 2009 (UTC)

• Ruby 3 Oct 2020 2:17 UTC

  2 points

  on: Etiquette

  Talk:Etiquette

  I don’t know if “Eti­quette” is the proper ti­tle for this page. Thoughts? MrHen 20:50, 8 Jan­uary 2010 (UTC)

  Please start pages already with at least proper stubs, so that the in­tent is clear and there is some value and qual­ity in the page as it is. It’s very im­por­tant not to al­low bit rot in. For now, there are pages FAQ and Less Wrong, which are more about the blog than the wiki, as pos­si­ble places to start or in­ter­link. --Vladimir Nesov 21:54, 8 Jan­uary 2010 (UTC)

  I couldn’t find a tem­plate for Stubs. What are you talk­ing about? MrHen 22:22, 8 Jan­uary 2010 (UTC)

  I’m talk­ing about even a half-writ­ten page be­ing pre­sentable, stub tem­plate is only an ic­ing. I added info about page sta­tus tem­plates in­clud­ing stub at Help:Style Guide#For­mat­ting (al­though this one you should’ve been able to find in the code of the many stub pages on the wiki). --Vladimir Nesov 22:57, 8 Jan­uary 2010 (UTC)

  Okay. So… how should I do that for this page? In­stead of guess­ing in­cor­rectly again I’m just go­ing to start ask­ing. I do not know of a high-level post talk­ing about what high-level post eti­quette at LessWrong is. I cer­tainly have no au­thor­ity to write the first notes on the sub­ject. I’ll add what I can. MrHen 23:22, 8 Jan­uary 2010 (UTC)

  You could write up a com­plete (with­out “dan­gling ends”), if short and sim­ple, po­si­tion state­ment on what you’ve ob­served the post eti­quette to be/​what it should be. I’ve up­dated Help:Eti­quette to in­clude a note about bit rot. The rea­son to be more con­cerned about it here is that the num­ber of con­tri­bu­tions is low, and so bad-qual­ity bits left and for­got­ten by users over time threaten to over­whelm the wiki if al­lowed. --Vladimir Nesov 23:57, 8 Jan­uary 2010 (UTC)

• Ruby 3 Oct 2020 2:16 UTC

  2 points

  on: De­bate Tools

  Talk:De­bate tools

  Sugges­tions for im­prove­ment: “pros/​cons/​first men­tioned” for­mat is ar­tifi­cial, and notes there are of lit­tle use (ei­ther re­move them all, or in­te­grate in the sum­mary, ex­cept for “first men­tioned”). The word “Sum­mary” doesn’t need to be used. --Vladimir Nesov 07:57, 30 Au­gust 2010 (UTC)

  Sugges­tion: Would pol.is count as a de­bate tool? It’s more like a con­sul­ta­tion /​ con­sen­sus-build­ing tool… https://​​blog.pol.is/​​pol-is-in-taiwan-da7570d372b5?gi=d4362adab60d