Viliam

• Viliam 29 Sep 2020 20:36 UTC

  4 points

  in reply to: AllAmericanBreakfast’s comment on: AllAmer­i­canBreak­fast’s Shortform

  In a perfect situ­a­tion, it would be pos­si­ble to achieve mean­ingful ex­pe­riences with­out pain, but usu­ally it is not pos­si­ble. A per­son who op­ti­mizes for short-term pain avoidance, will not reach the mean­ingful ex­pe­rience. Be­cause op­ti­miz­ing for short-term pain avoidance is nat­u­ral, we have to re­mind our­selves to over­come this in­stinct.

• Viliam 29 Sep 2020 19:28 UTC

  2 points

  on: Numer­acy ne­glect—A per­sonal postmortem

  So, what now?

  There are many free text­books on­line (1, 2, 3...), so maybe choose math, down­load some­thing that starts sim­ple, read it, if it is too com­pli­cated put it away and down­load some­thing else. Down­load the text­books to a book reader, so you can read them while you travel, etc. If you get stuck some­where, search the an­swer on­line, if that fails ask in Less Wrong short­form.

  You lost some time, but it is not too late to learn. If you start now, af­ter ten years you will be happy that you did.

  I used to be good at math at high school, but then I chose com­puter sci­ence at uni­ver­sity, and ended up mak­ing stupid web­sites for 20 years. A few months ago, I de­cided to give it an­other try, and down­loaded a few books. (I think I still have solid high-school knowl­edge, so I de­cided to go ahead and chose set the­ory.) First time I read a book, I didn’t un­der­stand most of it. Then I read it again and did some of the ex­er­cises, and sud­denly it made much more sense. Now I un­der­stand even some Wikipe­dia ar­ti­cles which are definitely not writ­ten for be­gin­ners. (The in­ter­sec­tion be­tween “un­der­stands an es­o­teric topic”, “can ex­plain it clearly”, and “will­ing to edit Wikipe­dia ar­ti­cles” is small, some­times nonex­is­tent.) I don’t have much free time with job and kids, but I try to reg­u­larly find an hour or two. But I am also picky; if a book doesn’t work for me, I throw it away and take an­other. That’s the ad­van­tage of free down­load­ing. (Ahem.)

• Viliam 28 Sep 2020 22:26 UTC

  0 points

  in reply to: habryka’s comment on: What hard sci­ence fic­tion sto­ries also got the so­cial sci­ences right?

  Peo­ple usu­ally sub­stan­tially in­crease the amount of work they do, and gen­er­ally re­port higher lev­els of en­gage­ment and very rarely just give up.

  In short term, sure. In long term? I look around me and see peo­ple so tired of tak­ing pre­cau­tions against COVID-19 that they would rather die than spend an­other day wear­ing a face mask.

  In the book, the time in­ter­vals were much longer, given the dis­tances in uni­verse, and the speed of light. Peo­ple were ca­pa­ble of dra­matic de­ci­sions when the threat was de­tected. A few years later, with threat still on the way, they were already burned out. Sounds re­al­is­tic to me.

  And the “cos­mic so­ciol­ogy” is Med­i­ta­tions on Moloch turned up to eleven.

• Viliam 28 Sep 2020 19:47 UTC

  6 points

  in reply to: AllAmericanBreakfast’s comment on: AllAmer­i­canBreak­fast’s Shortform

  I started out feel­ing a lot smarter. I think it was com­mu­nity val­i­da­tion + the promise of mys­ti­cal knowl­edge.

  Too smart for your own good. You were sup­posed to be­lieve it was about ra­tio­nal­ity. Now we have to ban you and erase your com­ment be­fore other peo­ple can see it. :D

  Now I’ve started to feel dumber. Prob­a­bly be­cause the les­sons have sunk in enough that I catch my own bad ideas and no­tice just how many of them there are. [...] you have to ac­cept feel­ing stupid all the time. But I still look down that old road and I’m glad I’m not walk­ing down it any­more.

  Yeah, same here.

• Viliam 27 Sep 2020 9:34 UTC

  10 points

  on: Blog posts as epistemic trust builders

  Build­ing rep­u­ta­tion by re­peated in­ter­ac­tion.

  But it needs to be the type of in­ter­ac­tion where you no­tice and re­mem­ber the au­thor. For ex­am­ple, if you go to LessWrong, you are more likely to as­so­ci­ate “I read this on LessWrong” with the in­for­ma­tion, than if you just vis­ited LessWrong ar­ti­cles from links shared on so­cial net­works. (And it is prob­a­bly eas­ier to re­mem­ber Zvi than an av­er­age au­thor at LessWrong, be­cause Zvi re­cently posted a se­quence of ar­ti­cles, which is eas­ier to re­mem­ber than an equal num­ber of ar­ti­cles on un­re­lated top­ics.) You need to no­tice “ar­ti­cles by Zvi” as a sep­a­rate cat­e­gory first, and only then your brain can de­cide to as­so­ci­ate trust with this cat­e­gory.

  (Slate Star Codex takes this a bit fur­ther, be­cause for my brain it is eas­ier to re­mem­ber “I read this on SSC” than to re­mem­ber the set of ar­ti­cles writ­ten by Scott on LessWrong. This is brand­ing. If your qual­ity is con­sis­tently high, mak­ing the fact “this was writ­ten by me” more no­tice­able in­creases your rep­u­ta­tion.)

  The flip side of the coin is that the cul­ture of shar­ing hy­per­links on so­cial net­works de­stroys trust. If you read hun­dred ar­ti­cles from hun­dred differ­ent sources ev­ery day, your brain has a prob­lem to keep tabs. Be­fore in­ter­net, when you reg­u­larly read maybe 10 differ­ent jour­nals, you grad­u­ally no­ticed that X is re­li­able and Y is un­re­li­able. Be­cause some­times you read ten re­li­able sto­ries on one day, and ten un­re­li­able sto­ries on a differ­ent day, and it felt differ­ently. But on in­ter­net, there are hun­dred web­sites, and you switch be­tween them, so even if a few of the are no­to­ri­ously bad, it is hard to no­tice. Even harder, be­cause the same web­site can have mul­ti­ple au­thors with wildly differ­ent qual­ity. A sci­en­tist and a crack­pot can have a blog on the same do­main. With pa­per sources, the au­thors within one source were more bal­anced. (LessWrong is also kinda bal­anced, es­pe­cially if you only con­sider the up­voted ar­ti­cles.)

• Viliam 27 Sep 2020 9:04 UTC

  2 points

  in reply to: niplav’s comment on: ni­plav’s Shortform

  If the ad­block­ers be­come too pop­u­lar, web­sites will up­date to cir­cum­vent them. It will be a lot of work at the be­gin­ning, but prob­a­bly pos­si­ble.

  Cur­rently, most ads are in­jected by JavaScript that down­loads them from a differ­ent do­main. That al­lows ad­block­ers to block any­thing com­ing from a differ­ent do­main, and the ads are blocked rel­a­tively sim­ply.

  The straight­for­ward solu­tion would be to move ad in­jec­tion to the server side. The PHP (or what­ever lan­guage) code gen­er­at­ing the page would con­tact the ad server, down­load the ad, and in­ject it into the gen­er­ated HTML file. From the client per­spec­tive, it is now all com­ing from the same do­main; it is even part of the same page. The client can­not see the in­ter­ac­tion be­tween server and third party.

  Prob­lem with this solu­tion is that it is too easy for the server to cheat; to down­load thou­sand ex­tra ads with­out dis­play­ing them to any­one. The ad­ver­tis­ing com­pa­nies must find a way to pro­tect them­selves from fraud.

  But if smart peo­ple start think­ing about it, they will prob­a­bly find a solu­tion. The solu­tion doesn’t have to work perfectly, only statis­ti­cally. For ex­am­ple, the server dis­play­ing the ad could also take the client’s finger­print and send it to the ad­ver­tis­ing com­pany. Now this finger­print can of course ei­ther be real, or fic­tional if the server is cheat­ing. But the ad­ver­tis­ing com­pany could cross-com­pare finger­prints com­ing from thou­sand servers. If many differ­ent servers re­port hav­ing no­ticed the same iden­tity, the iden­tity is prob­a­bly real. If a server re­ports too many iden­tities that no one else have ever seen, the iden­tities are prob­a­bly made up. The ad­ver­tis­ing com­pany would sus­pect fraud if the frac­tion of unique iden­tities re­ported by one server ex­ceeds 20%. Some­thing like this.

• Viliam 26 Sep 2020 21:56 UTC

  5 points

  on: Most Peo­ple Aren’t Fishermen

  My opinion on mar­riage is con­ser­va­tive—peo­ple should get mar­ried when they want to have kids. They don’t sac­ri­fice to each other; they to­gether pay the costs of cre­at­ing a good en­vi­ron­ment for their kids to grow up in.

  If you don’t want to have kids, you can have sex or live to­gether also with­out mar­riage, and di­vorce made mar­riage kinda use­less as a sig­nal of com­mit­ment. (Okay, there are other rea­sons, too, such as tax benefits.)

  From this per­spec­tive, I am quite sur­prised that you see mar­riage as an op­po­site of growth mind­set. Mak­ing a com­mit­ment to rad­i­cally change your ev­ery­day life for the next 20 years, and tak­ing re­spon­si­bil­ity for challenges you never ex­pe­rienced be­fore, know­ing that there is no way to stop this train with­out some­one get­ting hurt...

  Similarly, strate­gi­cally mak­ing a sac­ri­fice counts as “growth” in my books. (Jor­dan Peter­son agrees.)

  Not know­ing your friend’s buddy of course makes it im­pos­si­ble for me to guess whether his de­ci­sion was a re­sult of ma­tu­rity or… some­thing com­pletely differ­ent.

  They don’t want to take the risky leap in be­com­ing fish­er­men. As long as they keep re­ceiv­ing enough fish, they’ll tol­er­ate the mis­ery.

  Who knows what would hap­pen if the risk be­came smaller, e.g. thanks to the UBI. You seem to as­sume that peo­ple who don’t ac­cept risk now, they sim­ply are the type of per­son who would never take a risk. But maybe many peo­ple con­sider some smaller lev­els of risk ac­cept­able (e.g. “there is a chance I will spend three years work­ing on some­thing that ul­ti­mately fails, and if I switch to a reg­u­lar ca­reer later, I will be three years be­hind my peers”), and some higher lev­els of risk un­ac­cept­able (e.g. “there is a chance I will lose my lifelong sav­ings and live in poverty, or get sick with­out hav­ing good health­care”). And maybe too many peo­ple live in a situ­a­tion where try­ing some­thing rev­olu­tion­ary would re­quire the un­ac­cept­able lev­els of risk.

  By the way, some peo­ple work in cor­po­ra­tions be­cause they need to ac­cu­mu­late the cap­i­tal nec­es­sary for start­ing their own com­pany. And some peo­ple work in cor­po­ra­tions be­cause their com­pany failed and now they have to pay their debts. Both of these can take many years.

• Viliam 26 Sep 2020 15:43 UTC

  3 points

  in reply to: crabman’s comment on: What is com­plex­ity sci­ence? (Not com­pu­ta­tional com­plex­ity the­ory) How use­ful is it? What ar­eas is it re­lated to?

  Yes, this is a motte of “emer­gence”.

  The prob­le­matic part is when you turn the con­cept of “de­spite un­der­stand­ing the rules of all lit­tle pieces, it is still difficult for a hu­man to pre­dict some pat­terns of their in­ter­ac­tion” into a noun, and then kinda sug­gest that it refers to a mys­te­ri­ous thing that many difficult-to-pre­dict pat­terns have in com­mon, and that there is a way to study this mys­te­ri­ous thing it­self, and by do­ing so gain in­sight (go­ing be­yond “yep, com­plex things with many parts are of­ten difficult to pre­dict”) into all these difficult-to-pre­dict pat­terns.

  In other words, if you make it seem as if un­der­stand­ing of e.g. gliders and biolog­i­cal evolu­tion (two ex­am­ples of “emer­gence”) al­lows you to bet­ter pre­dict stock mar­kets (an­other ex­am­ple of “emer­gence”… there­fore, they all should have some­thing in com­mon, and you can study that).

  Quot­ing Eliezer: (source)

  Taken liter­ally, that de­scrip­tion fits ev­ery phe­nomenon in our uni­verse above the level of in­di­vi­d­ual quarks [...] There’s noth­ing wrong with say­ing “X emerges from Y,” where Y is some spe­cific, de­tailed model with in­ter­nal mov­ing parts. [...] Grav­ity arises from the cur­va­ture of space­time, ac­cord­ing to the spe­cific math­e­mat­i­cal model of Gen­eral Rel­a­tivity. Chem­istry arises from in­ter­ac­tions be­tween atoms, ac­cord­ing to the spe­cific model of quan­tum elec­tro­dy­nam­ics.

  The phrase “emerges from” is ac­cept­able, just like “arises from” or “is caused by” are ac­cept­able, if the phrase pre­cedes some spe­cific model to be judged on its own mer­its. How­ever, this is not the way “emer­gence” is com­monly used. “Emer­gence” is com­monly used as an ex­pla­na­tion in its own right.

• Viliam 26 Sep 2020 13:29 UTC

  3 points

  in reply to: Richard_Kennaway’s comment on: What is com­plex­ity sci­ence? (Not com­pu­ta­tional com­plex­ity the­ory) How use­ful is it? What ar­eas is it re­lated to?

  Similar here. Read­ing the ti­tle, think­ing “ex­plain­ing how ex­po­nen­tial com­plex­ity is worse than lin­ear will be a piece of cake”. Read­ing the text, think­ing “okay, how is this differ­ent from cy­ber­net­ics?”

  Even Wikipe­dia just says “study of com­plex­ity and com­plex sys­tems”, and then points to­wards com­pu­ta­tional com­plex­ity and sys­tems the­ory. Wikipe­dia has its flaws, but...

  Even among the re­sources linked as “some courses/​primers/​in­tro­duc­tions”, half of them do not con­tain words “com­plex­ity the­ory” or “com­plex­ity sci­ence”. Which makes me doubt:

  It is at least not 100% crack­pot­tery, since some books are pub­lished by Prince­ton uni­ver­sity press and Oxford uni­ver­sity press.

  Just be­cause those books con­tain the word “com­plex” or “com­plex­ity”, doesn’t mean they sup­port the idea of “com­plex­ity sci­ence”.

• Viliam 26 Sep 2020 12:34 UTC

  3 points

  in reply to: MikkW’s comment on: MikkW’s Shortform

  ei­ther A) most LW’ers aren’t in­vest­ing in stocks

  Does LW 2.0 still have the func­tion­al­ity to make polls in com­ments? (I don’t re­mem­ber see­ing any re­cently.) This seems like the ques­tion that could be eas­ily an­swered by a poll.

• Viliam 25 Sep 2020 20:55 UTC

  2 points

  in reply to: snog toddgrass’s comment on: Shittests are ac­tu­ally good

  Seems that we mostly agree here, the ma­jor dis­agree­ment is about ter­minol­ogy.

  I dis­agree about too wide use of “shit-test­ing” to in­clude… maybe not test­ing in gen­eral, but still more than the nar­row mean­ing in the PUA liter­a­ture… which is ap­prox­i­mately “pur­pose­fully an­noy­ing your part­ner, in or­der to find out whether the part­ner is good at keep­ing their bound­aries”.

  I agree that if there are in­com­pat­i­bil­ities be­tween peo­ple, it’s bet­ter to find them sooner rather than later. And that some­times you need to search for the pos­si­ble in­com­pat­i­bil­ities ac­tively.

• Viliam 24 Sep 2020 22:38 UTC

  6 points

  on: Los­ing the for­est for the trees with grid drawings

  Iron­i­cally, Draw­ing on the Right Side of the Brain recom­mends as an ex­er­cise to draw the pic­ture up­side-down, so that the “for­est” does not dis­tract you from get­ting the de­tails right.

  (But it is not as­sumed that the re­sult­ing pic­ture will be beau­tiful, and there is also no grid that would in­tro­duce ar­tifi­cial line bends.)

  Per­haps a metaphor could be made for that, too, that some­times fo­cus­ing on the big pic­ture pre­vents you from notic­ing that you got the de­tails wrong, which can also ruin the out­come.

• Viliam 24 Sep 2020 22:28 UTC

  10 points

  on: Shittests are ac­tu­ally good

  The phrase “If you can’t han­dle me at my worst, you don’t de­serve me at my best” is some­times the idea.

  I would as­sume that your cur­rent “worst” is the best pre­dic­tor of your fu­ture be­hav­ior. And frankly, shouldn’t I? I think it is a con­sen­sus among the peo­ple who use the word that shit-tests never end.

  I am am­biva­lent about the whole idea of shit-test­ing. On one hand, it makes sense to test your part­ner’s re­ac­tion to your bad be­hav­ior. Be­cause, if you stay to­gether for a long time, sooner or later some bad be­hav­ior will hap­pen; life will throw a lot of stress on you, and you will snap. You need the kind of part­ner who can sur­vive it grace­fully. If it is some­one who would col­lapse, or go nu­clear, that is a time-bomb; bet­ter avoid that.

  On the other hand, if some­one oc­cas­sion­ally be­haves badly even when ev­ery­thing goes fine, it doesn’t ex­actly give me con­fi­dence that the per­son will try their best when things get hard. When a life-or-death situ­a­tion hap­pens (and by the same logic, sooner or later it will), would you want your part­ner to choose ex­actly that mo­ment for their next shit-test? And what makes you so sure they wouldn’t, if they already do it ha­bit­u­ally?

  So… shit-test­ing al­lows you to se­lect a bet­ter part­ner… but at the same time, “be­ing the kind of per­son who shit-tests their part­ner” makes you a worse part­ner. (Which is kinda your part­ner’s prob­lem, not yours, but still...)

  It’s like those “if you re­ally love me, you will do X for me” situ­a­tions, when some­one de­mands an ar­bi­trary sac­ri­fice X as a proof of love. If you are too fo­cused on sig­nal­ing your love, you may miss the larger pic­ture, which is that a per­son who loves you would not ask you to make ar­bi­trary sac­ri­fices. So you are set­ting your­self up for a one-sided re­la­tion­ship; and the right an­swer would be to walk away, and find some­one else who is will­ing to re­cip­ro­cate your love. (Even if you be­lieve that suffi­ciently strong one-sided love may even­tu­ally elicit the same feel­ings in the other party, it still makes more sense to choose some­one who will not abuse you be­fore that hap­pens, as­sum­ing it hap­pens at all.)

  How to get out of this dilemma? Ar­bi­trar­ily test­ing your part­ner is bad, leav­ing them untested is dan­ger­ous...

  Per­haps, if you could ob­serve your part­ner in tests that life throws at them nat­u­rally. That would re­quire to spend a lot of time to­gether. If you want to speed it up, you could choose a situ­a­tion that in­creases stress lev­els nat­u­rally, for some good rea­son. For ex­am­ple, spend a va­ca­tion in moun­tains to­gether. Or some­thing else that gets you tired and un­com­fortable, but for rea­sons bet­ter than one per­son choos­ing to an­noy the other.

  (I won­der if shit-test­ing was also so fre­quent in the past, or whether it is an adap­ta­tion to the mod­ern dat­ing mar­ket where you have to test your part­ners quickly.)

• Viliam 24 Sep 2020 17:33 UTC

  4 points

  on: The ra­tio­nal­ist com­mu­nity’s lo­ca­tion problem

  Ra­tion­al­ists may be less likely than av­er­age to want kids, but that doesn’t mean none of us are hav­ing them.

  Many peo­ple don’t want to have kids in their 20s, and change their mind later. Ten years later, I could imag­ine that many ra­tio­nal­ists will feel am­bigu­ous, and then some­thing can start a chain re­ac­tion of hav­ing kids.

  Ac­tu­ally, I think it would be su­per cool to have a gen­er­a­tion of kids of ap­prox­i­mately the same age, whose par­ents are ra­tio­nal­ists liv­ing next to each other and can co­or­di­nate on school choice /​ home­school­ing /​ pro­vid­ing ex­tra les­sons in free time.

• Viliam 24 Sep 2020 17:12 UTC

  2 points

  on: I have dis­cov­ered a new kind of un­em­ploy­ment.

  Firms im­pose higher effort de­mands on work­ers; work­ers have to com­plete more tasks (for a higher wage) or be fired.

  This sounds cor­rect, but I thought it was spe­cific for IT. I mean the pop­u­lar trends of be­ing “full-stack de­vel­oper” and “dev-ops”, which in my opinion both mean: -- Why should I hire two or three spe­cial­ists, when one per­son could do ev­ery­thing alone? And if the pro­ject size re­quires hiring two or three peo­ple any­way, at least this will make them more re­place­able, and I can im­me­di­ately move one of them to an­other pro­ject when the worst crisis is over. And if be­ing un­able to main­tain top-level ex­per­tise at too many things at the same time makes them feel like im­pos­tors, at least it will keep them hum­ble.

  Do you sug­gest it also hap­pens in other in­dus­tries? Your ar­ti­cles has “tech­nol­ogy” in its ti­tle, but seems to talk about econ­omy in gen­eral. Could you per­haps provide more spe­cific in­for­ma­tion about other in­dus­tries? Un­for­tu­nately, I am not qual­ified to com­ment on the sec­ond part of your ar­ti­cle.

  I will ad­mit that the claim Amer­i­can work­ers have be­come Stakhanovites is a bold one. It’s the sort of claim that im­me­di­ately raises all sorts of ob­jec­tions and ques­tions, like: how is that even pos­si­ble in a cap­i­tal­ist econ­omy, and why hasn’t it also hap­pened out­side of the U.S?

  I don’t have enough data, but is it pos­si­ble that this is more about IT (and also in other coun­tries) than about Amer­i­cans? Be­cause the hy­poth­e­sis “nerds suck at ne­go­ti­a­tion, and are eas­ily brain­washed” would ex­plain a few things we see. I mean, even com­rade Stakhanov didn’t spend his free time im­prov­ing his Github port­fo­lio.

• Viliam 24 Sep 2020 15:31 UTC

  6 points

  in reply to: AllAmericanBreakfast’s comment on: AllAmer­i­canBreak­fast’s Shortform

  There are things like “ly­ing for a good cause”, which is a text­book ex­am­ple of what will go hor­ribly wrong be­cause you al­most cer­tainly un­der­es­ti­mate the sec­ond-or­der effects. Like the “do not wear face masks, they are use­less” ex­pert ad­vice for COVID-19, which was a “clever” dark-arts move aimed to pre­vent peo­ple from buy­ing up nec­es­sary med­i­cal sup­plies. A few months later, hun­dreds of thou­sands have died (also) thanks to this ad­vice.

  (It would prob­a­bly be use­ful to com­pile a list of ly­ing for a good cause gone wrong, just to drive home this point.)

  Think­ing about his­tor­i­cal record of peo­ple pro­mot­ing the use of dark arts within ra­tio­nal­ist com­mu­nity, con­sider In­ten­tional In­sights. Turned out, the or­ga­ni­za­tion was also us­ing the dark arts against the ra­tio­nal­ist com­mu­nity it­self. (There is a more gen­eral les­son here: when­ever a fan of dark arts tries to make you see the wis­dom of their ways, you should as­sume that at this very mo­ment they are prob­a­bly already us­ing the same tech­niques on you. Why wouldn’t they, given their ex­pressed be­lief that this is the right thing to do?)

  The gen­eral prob­lem with ly­ing is that peo­ple are bad at keep­ing mul­ti­ple in­de­pen­dent mod­els of the world in their brains. The eas­iest, in­stinc­tive way to con­vince oth­ers about some­thing is to start be­liev­ing it your­self. To­day you de­cide that X is a strate­gic lie nec­es­sary for achiev­ing goal Y, and to­mor­row you re­al­ize that ac­tu­ally X is more cor­rect than you origi­nally as­sumed (this is how self-de­cep­tion feels from in­side). This is in con­flict with our goal to un­der­stand the world bet­ter. Also, how would you strate­gi­cally lie as a group? Post it openly on­line: “Hey, we are go­ing to spread the lie X for in­stru­men­tal rea­sons, don’t tell any­one!” :)

  Then there are things like “us­ing tech­niques-or­thog­o­nal-to-truth to pro­mote true things”. Here I am quite guilty my­self, be­cause I have long ago ad­vo­cated turn­ing the Se­quences into a book, rea­son­ing, among other things, that for many peo­ple, a book is in­her­ently higher-sta­tus than a web­site. Ob­vi­ously, con­vert­ing a web­site to a book doesn’t in­crease its truth value. This comes with smaller risks, such as get­ting high on your own sup­ply (con­vinc­ing our­selves that ar­ti­cles in the book are in­her­ently more valuable than those that didn’t make it for what­ever rea­son, e.g. be­ing writ­ten af­ter the book was pub­lished), or wast­ing too many re­sources on things that are not our goal.

  But at least, in this cat­e­gory, one can openly and cor­rectly de­scribe their be­liefs and goals.

  Me­taphor­i­cally, rea­son is tra­di­tion­ally as­so­ci­ated with vi­sion/​light (e.g. “en­light­en­ment”), ig­no­rance and de­cep­tion with blind­ness/​dark­ness. The “dark side” also refer­ences Star Wars, which this nerdy au­di­ence is fa­mil­iar with. So, if the use of the term it­self is an ex­am­ple of dark arts (which I sup­pose it is), at least it is the type where I can openly ex­plain how it works and why we do it, with­out ru­in­ing its effect.

  But does it make us up­date too far against the use of de­cep­tion? Uhm, I don’t know what is the op­ti­mal amount of de­cep­tion. Un­like Kant, I don’t be­lieve it’s liter­ally zero. I also be­lieve that peo­ple err on the side of ly­ing more than is op­ti­mal, so a nudge in the op­po­site di­rec­tion is on av­er­age an im­prove­ment, but I don’t have a proof for this.

• Viliam 23 Sep 2020 20:58 UTC

  2 points

  in reply to: fraidykluofficer.com’s comment on: Com­par­a­tive ad­van­tage and when to blow up your island

  In this ex­am­ple it is as­sumed that the en­tire is­land is liter­ally owned by one per­son. So, if you wish, this per­son may be a metaphor for a strong cen­tral­ized gov­ern­ment.

  De­stroy­ing your pro­duc­tion ca­pac­ity is a strate­gic mis­take, and ex­poses you to black­mail in the fu­ture. A smart owner (or a smart cen­tral­ized gov­ern­ment) would not let that hap­pen. If you want to give me free ba­nanas, okay, I will take them; but I will still keep my ba­nana plan­ta­tion ready. That way, I get free ba­nanas to­day and keep my abil­ity to pro­duce ba­nanas to­mor­row.

  (And the other side of the same coin is that a smart owner—or cen­tral­ized gov­ern­ment—will try to ex­pand their fu­ture pro­duc­tion ca­pac­i­ties. For ex­am­ple, if to­day it is for me more prof­itable to grow ba­nanas than to write com­puter soft­ware, I might strate­gi­cally de­cide to write soft­ware any­way, at least part-time, be­cause two or three years later my soft­ware-writ­ing skills are likely to in­crease dra­mat­i­cally, while my ba­nana-grow­ing skills would prob­a­bly re­main the same. So the com­par­a­tive ad­van­tage of to­mor­row may re­ward me for writ­ing soft­ware, but in or­der to get there, I need to ac­cept some dis­ad­van­tage to­day.)

  That said, an­other ques­tion is whether sub­sidies are the best way to keep your pro­duc­tion ca­pac­ity, and what amount of sub­sidies is op­ti­mal. (Of course, the farm­ers will always say “more is bet­ter” for ob­vi­ous rea­sons.) If we dis­cuss real-life agri­cul­ture, I would even challenge which types of prod­ucts should we sub­si­dize: if the goal is to pre­vent starv­ing, we prob­a­bly do not need to pro­tect our meat pro­duc­tion—if the other coun­tries keep giv­ing us cheap meat, let them; and if they sud­denly stop do­ing that (in the un­likely case that all meat-sub­si­diz­ing coun­tries would co­or­di­nate to do this in the same year), we may have a year or two of mostly veg­e­tar­ian diet, but no one is go­ing to die.

  In other words, al­though some pro­tec­tion of pro­duc­tion ca­pac­ity is strate­gi­cally im­por­tant, it doesn’t nec­es­sar­ily fol­low that the farm­ing sub­sidies, as we know them now, are any­where near the op­ti­mal solu­tion. (Speci­fi­cally, I think that sub­sidies of meat pro­duc­tion are com­pletely un­nec­es­sary—it is un­likely that all other coun­tries would stop sub­si­diz­ing meat at the same year, and in the un­likely case that would hap­pen, we would sur­vive any­way.)

• Viliam 23 Sep 2020 13:25 UTC

  5 points

  in reply to: johnswentworth’s comment on: Com­par­a­tive ad­van­tage and when to blow up your island

  In gen­eral, yes, but there can be other fac­tors that re­duce the pos­si­bil­ity to in­ter­act with many pos­si­ble part­ners.

  Geo­graph­i­cal lo­cal mo­nop­o­lies—there are thou­sands of is­lands in the ocean, but most of them are too far from your home. You could re­place your near­est trade part­ner with some­one fur­ther away, at an ex­tra cost; and if your near­est trade part­ner pushes you too far, you will do it. But within that in­ter­val, the ne­go­ti­a­tion is im­por­tant.

  Upfront trans­ac­tion costs—even if the trade part­ners are equiv­a­lent, but it is costly to start in­ter­act­ing with an­other one (you have to do a com­pli­cated back­ground check, you need to adapt to their speci­fics), this again cre­ates an ex­tra cost of switch­ing, and an in­ter­val within which it is about ne­go­ti­a­tion.

  Both can ap­ply at the same time.

  There is also a gray line be­tween “car­tel” and “peo­ple do­ing the same thing, act­ing self­ishly, but up­dat­ing on their com­peti­tors’ past ac­tions”. To make it sim­ple, imag­ine that that a fair price for a ton of ba­nanas is $100. (Fair price = what would be the mar­ket bal­ance if any­one could trade with any­one, in a world with zero trans­ac­tion costs.) But there is a $8 cost for trad­ing with some­one who is not your ge­o­graph­i­cally near­est trade part­ner. In this situ­a­tion, the ba­nana buy­ers can in­di­vi­d­u­ally pre­com­mit to buy at e.g. $95, be­cause they know that you will pre­fer to sell them for $95 rather than sell some­one else for $100, pay $8 for tran­sit, and only keep $92.

  Now imag­ine the ba­nana buy­ers have a web­site, where they pub­li­cly share their ex­pe­rience. (This is perfectly le­gal, right?) And there is this highly up­voted ar­ti­cle called: “Don’t buy ba­nanas for $100, you can get them for $95 us­ing game the­ory”. It be­comes com­mon knowl­edge that the ba­nana sel­l­ers suck at ne­go­ti­a­tion (they don’t have an analog­i­cal web­site), and that most ba­nana buy­ers only pay $95. -- Armed with this knowl­edge, you can now pre­com­mit to only pay $90 for a ton of ba­nanas next year, be­cause now it is known that the best price your neigh­bor can get from any­one else is $95.

  How many iter­a­tions can hap­pen, de­pends on the ex­act shape of diminish­ing re­turns. For ex­am­ple, even if I was will­ing to pay $100 for my first ton of ba­nanas, but us­ing my power of pre­com­mit­ment I already got them from my neigh­bor for $85, I am prob­a­bly not will­ing to pay $100 for the sec­ond ton of ba­nanas. Sup­pose the sec­ond ton of ba­nanas is only worth $90 to me. But to ob­tain it, from some­one who is not my neigh­bor, I would have to pay $85 + $8, which is more. So I will not defect against the new equil­ibrium. -- Here I act al­most like a car­tel mem­ber (my first ton of ba­nanas is worth $100 to me, and at the end I only buy one ton, and yet I pre­com­mit­ted to not pay more than $85), but I am still only fol­low­ing my self­ish in­cen­tives, and at no point I am sac­ri­fic­ing a po­ten­tial ex­tra profit in fa­vor of keep­ing the bal­ance.

  I feel like I am rein­vent­ing here the Marx­ist class con­flict, in a more gen­eral form, with em­pha­sis on shar­ing ne­go­ti­a­tion tac­tics. The essence is that one side shares their ne­go­ti­a­tion tricks, which work in­di­vi­d­u­ally even if no one else is us­ing them (this is what makes it not a car­tel), but quickly be­come a new stan­dard if shared; and the new stan­dard—and the com­mon knowl­edge thereof—be­comes a more pow­er­ful lev­er­age (this is what makes it car­tel-like in effect) in the fol­low­ing iter­a­tion of ne­go­ti­a­tion. The power to say: “Yes, you no­ticed that I am us­ing this dirty trick against you, but we both know that all my com­peti­tors use ex­actly the same trick, so you can­not pun­ish me by switch­ing to an­other. And it is perfectly le­gal, be­cause we co­or­di­nated this pub­li­cly. Your side as a whole sucks at ne­go­ti­a­tion, my side suc­cess­fully turned it into a global lev­er­age, and you as an in­di­vi­d­ual face an up­hill bat­tle here.”

• Viliam 20 Sep 2020 20:05 UTC

  8 points

  on: Viliam’s Shortform

  1) There was this fa­mous marsh­mal­low ex­per­i­ment, where the kids had an op­tion to eat one marsh­mal­low (phys­i­cally pre­sent on the table) right now, or two of them later, if they waited for 15 min­utes. The sci­en­tists found out that the kids who waited for the two marsh­mal­lows were later more suc­cess­ful in life. The stan­dard con­clu­sion was that if you want to live well, you should learn some strat­egy to de­lay grat­ifi­ca­tion.

  (A less known re­sult is that the op­ti­mal strat­egy to get two marsh­mal­lows was to stop think­ing about marsh­mal­lows at all. Kids who fo­cused on how awe­some it would be to get two marsh­mal­lows af­ter re­sist­ing the temp­ta­tion, were less suc­cess­ful at ac­tu­ally re­sist­ing the temp­ta­tion com­pared to the kids who dis­tracted them­selves in or­der to for­get about the marsh­mal­lows—the one that was there and the hy­po­thet­i­cal two in the fu­ture—com­pletely, e.g. they just closed their eyes and took a nap. Iron­i­cally, when some­one gives you a lec­ture about the marsh­mal­low ex­per­i­ment, clos­ing your eyes and tak­ing a nap is al­most cer­tainly not what they want you to do.)

  After the origi­nal ex­per­i­ment, some peo­ple challenged the naive in­ter­pre­ta­tion. They pointed out that whether de­lay­ing grat­ifi­ca­tion ac­tu­ally im­proves your life, de­pends on your en­vi­ron­ment. Speci­fi­cally, if some­one tells you that giv­ing up a marsh­mal­low now will let you have two in the fu­ture… how much should you trust their word? Maybe your ex­pe­rience is that af­ter trust­ing some­one and giv­ing up the marsh­mal­low in front of you, you later get… a rep­u­ta­tion of be­ing an easy mark. In such case, grab­bing the marsh­mal­low and ig­nor­ing the talk is the right move. -- And the cor­re­la­tion the sci­en­tists found? Yeah, sure, peo­ple who can de­lay grat­ifi­ca­tion and hap­pen to live in an en­vi­ron­ment that re­wards such be­hav­ior, will suceed in life more than peo­ple who live in an en­vi­ron­ment that pun­ishes trust and long-term think­ing, duh.

  Later ex­per­i­ments showed that when the ex­per­i­menter es­tab­lishes them­selves as an un­trust­wor­thy per­son be­fore the ex­per­i­ment, fewer kids re­sist tak­ing the marsh­mal­low. (Duh. But the point is that their pre­vi­ous lives out­side the ex­per­i­ment have also shaped their ex­pec­ta­tions about trust.) The les­son is that our adap­ta­tion is more com­plex than was origi­nally thought: the abil­ity to de­lay grat­ifi­ca­tion de­pends on the na­ture of the en­vi­ron­ment we find our­selves in. For rea­sons that make sense, from the evolu­tion­ary per­spec­tive.

  2) Read­ers of Less Wrong of­ten re­port hav­ing prob­lems with pro­cras­ti­na­tion. Also, many provide an ex­am­ple when they re­al­ized at young age, on a deep level, that adults are un­re­li­able and in­sti­tu­tions are in­com­pe­tent.

  I won­der if there might be a con­nec­tion here. Some­thing like: re­al­iz­ing the profound abyss be­tween how our civ­i­liza­tion is, and how it could be, is a su­per­stim­u­lus that switches your brain per­ma­nently into “we are doomed, eat all your marsh­mal­lows now” mode.

• Viliam 20 Sep 2020 15:29 UTC

  13 points

  in reply to: Ericf’s comment on: Against Victimhood

  A sys­tem­at­i­cally op­pressed group can still be wrong. Be­ing op­pressed gives you an ex­pe­rience other peo­ple don’t have, but doesn’t give you epistemic su­per­pow­ers. You can still de­rive wrong con­clu­sions, de­spite hav­ing ac­cess to spe­cial data.

  Anec­dote time: When I was a kid, I was bul­lied by some­one who did lots of sport. As a re­sult, I de­vel­oped an un­con­scious aver­sion to sport. (Be­cause I didn’t want to be like him, and I didn’t want to par­ti­ci­pate in things that re­minded me of him.) Ob­vi­ously, this only fur­ther re­duced the qual­ity of my life. Years later, I found some great friends, who also did lots of sport. Soon, the aver­sion dis­ap­peared. My un­con­scious­ness de­cided it was ac­tu­ally okay to be like them.

  Maybe I am gen­er­al­iz­ing my ex­pe­rience too much, but look­ing at some groups, it seems like they fol­low the same al­gorithm (some­times ex­cept for the happy end­ing, yet). At some mo­ment in his­tory, your group hap­pens to be at the bot­tom of the so­cial lad­der. Others—the bad guys—have the money, the ed­u­ca­tion, the in­sti­tu­tions, etc. Your group starts as­so­ci­at­ing money, ed­u­ca­tion, and in­sti­tu­tions with the bad things that were done to them. The differ­ence is that when this hap­pens on a group level, the be­lief gets re­in­forced cul­turally, be­cause your friends and fam­ily all had the same ex­pe­rience.

  A few decades or cen­turies later, your group also gets an ac­cess to ed­u­ca­tion, money, and in­sti­tu­tions. (And I am not nec­es­sar­ily talk­ing about equal ac­cess here; just about some ac­cess, as op­posed to your an­ces­tors who had none.) But now ev­ery­one knows that these are things your peo­ple tra­di­tion­ally don’t have, and who­ever as­pires to get them is per­ceived as a traitor, as some­one who wants to join the bad guys. You can­not dis­cuss ra­tio­nally whether get­ting more ed­u­ca­tion, more money, and more of your peo­ple in in­sti­tu­tions is ac­tu­ally a good thing for your group, be­cause it in­creases your in­di­vi­d­ual and col­lec­tive power. The group as a whole is flinch­ing away from the painful ex­pe­rience in the col­lec­tive mem­ory, and the in­di­vi­d­u­als who go against the grain get pun­ished.

  (An ex­am­ple would be black peo­ple polic­ing each other against “act­ing white”, but a similar mechanism ap­plies in situ­a­tions where one group of white peo­ple was his­tor­i­cally op­pressed by an­other group of white peo­ple, be­cause of differ­ent lan­guage or re­li­gion or what­ever.)

  But of course, there may be also le­gi­t­i­mate rea­sons to dis­trust strate­gies that work for other peo­ple. For ex­am­ple, ed­u­ca­tion means ac­quiring debt in re­turn for higher ex­pected in­come in the fu­ture. If you know that the “higher in­come” is not go­ing to hap­pen, e.g. be­cause of racism, then ed­u­ca­tion is not as prof­itable for you as it would be for the ma­jor­ity.