Chris_Leong

• Chris_Leong 15 May 2018 10:14 UTC

  47 points

  in reply to: gjm's comment on: De­cou­pling vs Con­tex­tu­al­is­ing Norms

  Ac­tu­ally, I like De­cou­pling vs. Con­tex­tu­al­is­ing more too, es­pe­cially as they be­come sin­gle words.

• Chris_Leong 1 Oct 2017 13:33 UTC

  44 points

  on: Slack

  I clicked +1, but in way of feed­back I would sug­gest try­ing to be more pre­cise with your defi­ni­tions. “The ab­sence of bind­ing con­straints on be­hav­ior” sounds just like a syn­onym for free­dom. If that was the con­cept that this ar­ti­cle iden­ti­fied, it would be kind of pointless, but you’ve ac­tu­ally iden­ti­fied a new and use­ful con­cept.

  This has a few ad­van­tages:

  • Firstly, it makes the ar­ti­cle eas­ier to un­der­stand. Some­one peo­ple learn bet­ter by ex­am­ple, oth­ers bet­ter by ex­plicit defi­ni­tion.

  • Se­condly, it helps you make sure that what you have iden­ti­fied is in­deed a sin­gle con­cept and not a few closely re­lated ideas rol­led into one.

  • Third, it al­lows you clar­ify the con­cept in your own head and pick more cen­tral ex­am­ples to illus­trate it.

  • Fourth, it helps set so­cial norms by en­courag­ing other peo­ple to care­fully define their terms.

  I would make an al­ter­na­tive defi­ni­tion as fol­lows:

  • Firstly, we start off by as­sum­ing some kind of re­source (ie. time, en­ergy, money, so­cial cap­i­tal)

  • Now we can define Slack as keep­ing some of a re­source spare so that you can spend it when op­por­tu­ni­ties come up (ie. to do things eth­i­cally/​prop­erly/​just for fun/​for per­sonal growth, ect).

  This is a more spe­cific con­cept than free­dom, it is re­lated to the free­dom pro­vided by hav­ing spare re­sources.

  There is also a typ­i­cal paradigm of how peo­ple end up in situ­a­tions with­out enough slack. Ba­si­cally, they un­der­es­ti­mate the fu­ture op­por­tu­nity cost of ei­ther spend­ing or com­mit­ting to spend a par­tic­u­lar re­source now. For ex­am­ple, they spend a lot money at a cas­ino, with­out re­al­is­ing that there is a chance that they may lose their job, in which case their ex­cess money would sud­denly be­come much more use­ful. Or they com­mit to too many pro­jects be­cause they want to be agree­able in the pre­sent, with­out re­al­is­ing how a lack of time for re­cu­per­a­tion will wear them out when the work ends up more tiring than they ex­pect. More broadly, it is de­cid­ing what you can spend or com­mit now, whilst ig­nor­ing the un­knowns that might pop up in the fu­ture.

No op­tion to re­port spam

• Chris_Leong 29 Oct 2017 4:57 UTC

  38 points

  on: Inad­e­quacy and Modesty

  I’m re­ally happy that you are writ­ing a book on this topic. I mean, the Se­quences and the other dis­cus­sion on Less Wrong has given us a lot of tools with which to form our own opinion, but then we need to figure out how to bal­ance this against the opinions of ex­perts with more do­main-spe­cific knowl­edge. There’s a sense in which all the other knowl­edge isn’t of any use un­less we know when to ac­tu­ally use it.

  “Now, on the mod­est view, this was the un­fairest test imag­in­able. Out of all the times that I’ve ever sug­gested that a gov­ern­ment’s policy is sub­op­ti­mal, the rare time a gov­ern­ment tries my preferred al­ter­na­tive will se­lect the most main­stream, high­est-con­ven­tional-pres­tige poli­cies I hap­pen to ad­vo­cate, and those are the very policy pro­pos­als that mod­esty is least likely to dis­ap­prove of.”

  This is a pretty big deal, so I wanted to em­pha­sise it. Let’s sup­pose you come up with 50 poli­cies you think the gov­ern­ment should im­ple­ment. 10 get im­ple­mented and 8 work out well. Pretty good right? But what if 30 of your poli­cies would have been ut­terly stupid and this is ob­vi­ous to any of the ex­perts? This effect could com­pletely de­stroy your at­tempts at cal­libra­tion.

• Chris_Leong 3 Mar 2018 0:22 UTC

  37 points

  on: Monthly Meta: Com­mon Knowledge

  Slack: Slack is one of the con­cepts that seems to have gained the most trac­tion. Un­for­tu­nately, I don’t think we have a clear defi­ni­tion yet. Zvi defined Slack as the ab­sence of bind­ing con­straints on be­havi­our. I sug­gested that this would make it merely a syn­onym for free­dom and that his ar­ti­cle seems to de­scribe, “free­dom pro­vided by hav­ing spare re­sources”, but Zvi wasn’t happy with it and I don’t know if he’s set­tled on a pre­cise defi­ni­tion yet.

  I sus­pect that this idea is so pop­u­lar be­cause it gives peo­ple a word to de­scribe the dis­ad­van­tages of run­ning a sys­tem near max­i­mum ca­pac­ity. Zvi ex­plains how slack re­duces stress, al­lows you to pur­su­ing op­por­tu­ni­ties that arise, al­lows you to avoid bad trade-offs and al­lows long-term think­ing. In an­other key ar­ti­cle, Rae­mon ex­plains how it can lead to frag­ile de­sign de­ci­sions and mak­ing poor de­ci­sions that go un­no­ticed.

• Chris_Leong 8 Oct 2017 23:15 UTC

  37 points

  on: Con­tra dou­ble crux

  I’ve at­tended a CFAR work­shop. I agree with you that Dou­ble Crux has all of these the­o­ret­i­cal flaws, but it ac­tu­ally seems to work rea­son­ably well in prac­tise, even if these flaws make it kind of con­fus­ing. In prac­tise you just kind of stum­ble through. I strongly agree that if the tech­nique was rewrit­ten so that it didn’t have these flaws, it would be much eas­ier to learn as the stum­bling bit isn’t the most con­fi­dence in­spiring (this is when the in per­son as­sis­tance be­comes im­por­tant).
  
  One of the key el­e­ments that haven’t seen men­tioned here is this sep­a­ra­tion be­tween try­ing to per­suade the other per­son and try­ing to find out where your point of view differs. When you are try­ing to con­vince the other per­son it is much eas­ier to miss, for ex­am­ple, when there’s a differ­ence in a core as­sump­tion. Dou­ble Crux lets you un­der­stand the broad struc­ture of their be­liefs so that you can at least figure out the right kinds of things to say later to per­suade them that won’t be im­me­di­ately dis­missed.

On Disingenuity

• Chris_Leong 29 Oct 2017 23:04 UTC

  33 points

  on: In defence of epistemic modesty

  I feel like the in­side view and all-things-con­sid­ered view needs to be its own post. The idea of main­tain­ing two sep­a­rate prob­a­bil­ities for differ­ent situ­a­tions is pretty im­por­tant, but many peo­ple won’t get this in­sight as this post is rather long. (Ad­mit­tedly it’s slightly more com­plex in prac­tise be­cause the in­side view might de­pend on judge­ments about other things which in turn have their own in­side view and all-things-con­sid­ered view).

• Chris_Leong 11 May 2018 8:54 UTC

  31 points

  in reply to: query's comment on: Affor­dance Widths

  I sus­pect there’s a prac­tise effect here as well. Figur­ing out how to be as­sertive with­out be­ing dom­i­neer­ing or bossy is hard. Peo­ple who have grown up be­ing as­sertive will have had the op­por­tu­nity to learn, but those who try to be­come as­sertive be­cause they know its im­por­tant for the work­place won’t have de­vel­oped the judge­ment yet.

• Chris_Leong 26 Feb 2018 5:25 UTC

  31 points

  on: Meta-tations on Moder­a­tion: Towards Public Archipelago

  In­stead of purely fo­cus­ing on whether peo­ple will use these pow­ers well, per­haps we should also talk about ways to nudge them to­wards re­spon­si­bil­ity?

  • What if au­thors had to give rea­sons for the first 10 or 20 com­ments that they deleted? This would avoid cre­at­ing a long term bur­den, but it would also nudge them to­wards think­ing care­fully about which com­ments they would or would not delete at the start

  • Per­haps reign of ter­ror mod­er­a­tion should re­quire more karma than norm en­forc­ing? This would en­courage peo­ple to lean to­wards norm en­forc­ing and to only switch to reign of ter­ror if norm en­forc­ing wasn’t work­ing for them

  • I already posted this as a re­ply to a com­ment fur­ther up, but per­haps au­thors should only be able to col­lapse com­ments at first and then later be given delete pow­ers. Again, it would nudge them to­wards the former, rather than the later

  • For the 2nd and 3rd idea to work, the de­lay sys­tem couldn’t be based purely on karma as many au­thors already have enough karma to gain these pow­ers in­stantly. There should ideally be to some de­lay in gain­ing ac­cess to the higher level fea­tures even in this case

• Chris_Leong 10 Sep 2018 3:23 UTC

  30 points

  AF

  on: Com­ment on de­ci­sion theory

  I don’t sup­pose you could clar­ify what the un­re­solved is­sues in de­ci­sion the­ory are. What are the biggest is­sues that haven’t been solved for UDT or FDT? What is a co-or­di­na­tion prob­lem that hasn’t been solved? And what still isn’t known about coun­ter­fac­tu­als?

• Chris_Leong 22 May 2018 7:49 UTC

  30 points

  in reply to: Jan_Kulveit's comment on: The Berkeley Com­mu­nity & The Rest Of Us: A Re­sponse to Zvi & Benquo

  I don’t sup­pose I could per­suade you to write up a post with what you con­sider to be some of the most im­por­tant in­sights from net­work the­ory? I’ve started to think my­self that some of our mod­els that we tend to use within the ra­tio­nal­ity com­mu­nity are overly sim­plis­tic.

• Chris_Leong 26 Feb 2018 2:27 UTC

  30 points

  on: Meta-tations on Moder­a­tion: Towards Public Archipelago

  I’m a huge fan of the Archipelago model, but I’m un­sure how well our cur­rent struc­ture is suited to it. On Red­dit, you only have to learn one set of mod­er­a­tion norms per sub. On Less Wrong, learn­ing one set of mod­er­a­tion norms per au­thor seems like a much higher bur­den.

  In fact, the Archipelago model it­self, is one set of norms per group. If you don’t like the norms of a par­tic­u­lar group, you just don’t go there. This is harder when there isn’t a clear sep­a­ra­tion and ev­ery­one’s posts are all mixed to­gether.

Sum­mary: Sur­real Decisions

[Question] What are some con­crete prob­lems about log­i­cal coun­ter­fac­tu­als?

The In­spec­tion Para­dox is Everywhere

• Chris_Leong 29 Jan 2018 23:37 UTC

  26 points

  in reply to: Raemon's comment on: Critch on “Tak­ing AI Risk Se­ri­ously”

  I’d love to see more con­ver­sa­tions be­ing writ­ten up, es­pe­cially as some­one who is out­side of the Bay Area and so who is prob­a­bly miss­ing out on many ideas that are in wide cir­cu­la­tion there. I won­der if there are any Less Wrongers with jour­nal­ism train­ing who would take up this role.

De­con­fus­ing Log­i­cal Counterfactuals

Sugges­tion: New ma­te­rial shouldn’t be re­leased too fast

• Chris_Leong 11 Apr 2019 0:57 UTC

  24 points

  on: Ex­cerpts from a larger dis­cus­sion about simulacra

  Bau­drillard’s lan­guage seems quite re­li­gious, so I al­most feel that a re­li­gious ex­am­ple might re­late di­rectly to his claims bet­ter. I haven’t re­ally read Bau­drillard, but here’s how I’d ex­plain my cur­rent un­der­stand­ing:

  Stage 1: Peo­ple pray faith­fully in pub­lic be­cause they be­lieve in God and fol­low a re­li­gion. Those who wit­ness this prayer ex­pe­rience a win­dow into the tran­scen­dent.

  Stage 2: Peo­ple re­al­ise that they can gain so­cial sta­tus by pray­ing in pub­lic, so they pre­tend to be­lieve. Many peo­ple are aware of this, so wit­ness­ing an ap­par­ently sincere prayer ceases to be the same ex­pe­rience as you don’t know whether it is gen­uine or not. It still rep­re­sent the tran­scen­dent to some de­gree, but the ex­pe­rience of wit­ness­ing it just isn’t the same.

  Stage 3: Enough peo­ple have started pray­ing in­sincerely that al­most ev­ery­one starts jump­ing on the band­wagon. Publi­cly prayer has ceased to be an in­di­ca­tor of re­li­gios­ity or faith any more, but some par­tic­u­larly naive peo­ple still haven’t re­al­ised the pre­tence. Peo­ple still gain sta­tus from this for speak­ing suffi­ciently el­e­gantly. Peo­ple can’t be too ob­vi­ously fake though or they’ll be pun­ished ei­ther by the few still naive enough to buy into it or by those who want to keep up the pre­tence.

  Stage 4: Pray­ing is now seen purely as a so­cial move which op­er­ates ac­cord­ing to cer­tain rules. It’s no longer nec­es­sary in and of it­self to con­vince peo­ple that you are real, but part of the game may in­clude pun­ish­ments for mak­ing cer­tain moves. For ex­am­ple, if you swear dur­ing your prayer, that might be pun­ished for be­ing in­ap­pro­pri­ate, even though no-one cares about re­li­gion any more, be­cause that’s seen as cheat­ing or break­ing the rules of the game. How­ever, you can be ob­vi­ously fake in ways that don’t vi­o­late these rules, as the spirit of the rules has been for­got­ten. Maybe peo­ple pray for vain things like be­com­ing wealthy. Or they go to church one day, then post pic­tures of them get­ting smashed the next day on Face­book, which all their church friends see, but none of them care. The naive are too few to mat­ter and if they say any­thing, peo­ple will make fun of them.

  I’ll ad­mit that I’ve added some­thing of my own in­ter­pre­ta­tion here, es­pe­cially in terms of how strongly you have to pre­tend to be real at the var­i­ous stage