romeostevensit

• romeostevensit 22 Jul 2019 23:27 UTC

  4 points

  on: Where is the Mean­ing?

  Ex­pla­na­tions can be about a few differ­ent things. I’ve been hav­ing some luck split­ting them into var­i­ant (who, what, where) and in­var­i­ant (how, why, what) parts. Aris­to­tle and Sun Tzu had their own type sys­tems for causal­ity, and I’m cu­ri­ous why I haven’t been able to find much in ei­ther philos­o­phy or com­puter sci­ence about it. One guess is that I just haven’t found the right key­words yet, but I have gone hunt­ing around cita­tion chains from knowl­edge rep­re­sen­ta­tion and Pearl’s stuff, as well as the stuff from here and here. The modal logic stuff seems promis­ing but most of what’s been built on it seems like epicy­cles.

• romeostevensit 22 Jul 2019 16:41 UTC

  16 points

  in reply to: mr-hire's comment on: Ap­peal to Con­se­quence, Value Ten­sions, And Ro­bust Organizations

  In the ‘keep the or­ga­ni­za­tion from be­ing over­run’ sense, see also seal­ion­ing. The search space of worth­while things is very large and idiosyn­crat­i­cally ex­plored by well mean­ing, in­tel­li­gent peo­ple. Ag­gres­sive value laden ‘log­i­cal ar­gu­ments’ of­ten point to a tacit value to have ev­ery­one con­verge on the same set of meta­heuris­tics. This is be­cause the per­son do­ing this has a strong need for in­ter­nal con­sis­tency that they are ex­ter­nal­iz­ing onto their so­cial space. And there’s noth­ing wrong with want­ing in­ter­nal con­sis­tency. But if pressed hard, it is anti-truth seek­ing as an ag­gre­gate strat­egy be­cause you lose out on the con­silience of hav­ing differ­ent peo­ple pur­su­ing differ­ent search meth­ods. Episte­mol­ogy is a team sport. The ob­jec­tion would be ‘but if we don’t then ar­gue about what we’ve dis­cov­ered what’s the point?’ The point is that ad­ver­sar­ial pro­cesses as a part of the truth seek­ing pro­cess needs to be con­sen­sual. This ap­plies dou­bly when you aren’t in a 101 space and peo­ple might be sick of a dy­namic where sim­ple seem­ing ques­tions with com­pli­cated an­swers make newer mem­bers feel en­ti­tled to the effort needed to ex­plain said com­pli­cated an­swers. This is one of the rea­sons well writ­ten blog posts that can be refer­enced by name can be so helpful for com­mu­nity dis­course.

  I like this post by the way and my com­ment wasn’t an ob­jec­tion to it.

• romeostevensit 21 Jul 2019 8:34 UTC

  6 points

  on: Do you fear the rock or the hard place?

  In gen­eral: mak­ing so­matic re­ac­tions part of the mu­tual un­der­stand­ing of what is hap­pen­ing. The more you are aware of when you’re ab­stract­ing from a so­matic re­ac­tion the more you’ll rec­og­nize when oth­ers are do­ing it too. E-Prime is aimed at this, as the step of sym­bol­iz­ing one thing as an­other thing (X *is* Y) is a mo­ment in which you can catch this.

• romeostevensit 20 Jul 2019 6:43 UTC

  16 points

  on: Ap­peal to Con­se­quence, Value Ten­sions, And Ro­bust Organizations

  I feel like the elephant in the room is that con­vinc­ing log­i­cal ar­gu­ments are of­ten only weak to mod­er­ate ev­i­dence for some­thing.

• romeostevensit 18 Jul 2019 15:12 UTC

  6 points

  on: Nor­mal­is­ing util­ity as will­ing­ness to pay

  What does it mean to pay util­ity?

• romeostevensit 18 Jul 2019 7:10 UTC

  26 points

  on: Dialogue on Ap­peals to Consequences

  The bar that is set for ap­peals to con­se­quences im­ply the sort of equil­ibrium world you’ll end up in. Er­ring on the side of higher is bet­ter, be­cause it is hard to go the other way be­cause epistemic stan­dards tend to slide in the face of lo­cal in­cen­tives.

  I also want to note an ar­gu­men­ta­tive tac­tic that oc­curs on the tacit level whereby peo­ple will push you into a state where you need to ex­pend more en­ergy on av­er­age per truth bit than they do, so they even­tu­ally win by at­tri­tion. Re­lated to evap­o­ra­tive cool­ing. The sub­jec­tive ex­pe­rience of this feels like talk­ing to the cops. You sense that no big wins are available (be­cause they have their bot­tom line) but big losses are, so you stop talk­ing. If you’ve en­coun­tered this dy­namic, you rec­og­nize things like this

  > “You still haven’t re­futed my ar­gu­ment. If you don’t do so, I win by de­fault.”

  as part of the sup­port­ing frame­work for the dy­namic and it will make you very an­gry...which oth­ers will then use as part of the dy­namic which makes you an­gry which......

• romeostevensit 18 Jul 2019 6:55 UTC

  5 points

  on: Why it feels like ev­ery­thing is a trade-off

  Slip­ping out of pro­duc­tion pos­si­bil­ity fron­tier zero-sum re­la­tions in­volves slip­ping side­ways along some other di­men­sion. Sad­dle points in gra­di­ent de­scent. This is yet an­other rea­son why ‘If a prob­lem seems hard, the rep­re­sen­ta­tion is prob­a­bly wrong’ of­ten holds.

• romeostevensit 18 Jul 2019 6:51 UTC

  2 points

  in reply to: ChristianKl's comment on: 87,000 Hours or: Thoughts on Home Ownership

  Higher in more ex­pen­sive mar­kets, but yes.

• romeostevensit 17 Jul 2019 4:17 UTC

  3 points

  on: What is your Per­sonal Knowl­edge Man­age­ment sys­tem?

  Work­flowy. Dy­nal­ist and oth­ers have more fea­tures, but I don’t want more fea­tures. More fea­tures means more de­ci­sions. I or­ga­nize by month and do a re­view at the end of months, plus a year end re­view when col­laps­ing into my archive tab. Tags for things like book notes, quotes, rou­tines etc.

• romeostevensit 17 Jul 2019 4:11 UTC

  5 points

  on: How Can Ra­tion­al­ists Join Other Com­mu­ni­ties In­ter­ested in Truth-Seek­ing?

  Since The Ra­tion­al­ity Quo­tient mostly showed that Ra­tion­al­ity isn’t much of a thing on top of g, that means that de­spite not car­ing about the qual­ity of ar­gu­ments too much, other peo­ple aren’t suffer­ing worse life out­comes. One can take it as an op­por­tu­nity to be cu­ri­ous about why that might be. What might oth­ers who seem to be less ex­plic­itly/​ver­bally com­mit­ted to truth be get­ting right in other ways? I’ve found that spiritual com­mu­ni­ties are good for this, and more open to re­flec­tion than most, once the right se­man­tic flags are un­der­stood and trans­lated.

• romeostevensit 16 Jul 2019 22:03 UTC

  2 points

  in reply to: ChristianKl's comment on: 87,000 Hours or: Thoughts on Home Ownership

  https://​​www.less­wrong.com/​​posts/​​X7G3HfRuMzCSBzgbM/​​87-000-hours-or-thoughts-on-home-own­er­ship#Sofu2NKFTMsT6y4it

• romeostevensit 15 Jul 2019 9:06 UTC

  2 points

  on: How Should We Cri­tique Re­search? A De­ci­sion Perspective

  I call it de­ci­sion lev­er­age, for hand­i­ness, and look­ing for it changes ca­sual con­ver­sa­tion as well as anal­y­sis.

• romeostevensit 11 Jul 2019 7:07 UTC

  7 points

  on: How can guessti­mates work?

  Hu­man civ­i­liza­tions have sur­vivor­ship bias such that you’re see­ing the iter­ated out­put of those that found pa­ram­e­ter in­sen­si­tive parts of crit­i­cal dis­tri­bu­tions. Ex­oge­nous shocks are com­mon, and civs that weren’t metastable dis­ap­peared. This hap­pened on both the cul­tural and ge­netic level (you are already a civ­i­liza­tion of trillions of cells). As the pace of change ac­cel­er­ates things might get more dan­ger­ous as peo­ple push farther/​harder within the span of a sin­gle life­time be­fore nat­u­ral self cor­rec­tions kick in (four horse­men). The cur­rent trend could be seen as the hu­man race es­sen­tially tin­ker­ing with can­cer (run­away, un­con­trol­lable growth) to see if it will grant us im­mor­tal­ity be­fore it kills us.

• romeostevensit 10 Jul 2019 21:20 UTC

  10 points

  in reply to: Hazard's comment on: Magic is Dead, Give me Attention

  I liked the vuln­er­a­bil­ity of ad­mit­ting want­ing at­ten­tion and the ex­plo­ra­tion of mul­ti­ple con­se­quences of that, the way cop­ing strate­gies evolve and change as a re­sult of pres­sures.

• romeostevensit 10 Jul 2019 20:50 UTC

  5 points

  on: Magic is Dead, Give me Attention

  I re­ally ap­pre­ci­ated this post.

• romeostevensit 10 Jul 2019 17:07 UTC

  8 points

  in reply to: Vladimir_Nesov's comment on: Do bond yield curve in­ver­sions re­ally in­di­cate there is likely to be a re­ces­sion?

  What Liron said. A small num­ber of days ac­counts for most stock gains. Sharpe con­cluded that you’d need to make calls right about 74% of the time to win by timing.

  There are ways to sys­tem­at­i­cally make money. Diver­sifi­ca­tion, not timing, low time prefer­ence, lev­er­age, not be­ing loss averse and other fac­tors al­low you to har­vest money off of the poorly di­ver­sified, the mar­ket timers, the high time prefer­ence, the lev­er­age averse, and the loss averse over time.

• romeostevensit 10 Jul 2019 4:45 UTC

  5 points

  on: Do bond yield curve in­ver­sions re­ally in­di­cate there is likely to be a re­ces­sion?

  The pop­u­lar say­ing is that bond yield in­ver­sions have pre­dicted 6 of the last 3 re­ces­sions. In gen­eral, just run your same searches and ap­pend the word ‘myth’, ‘mis­con­cep­tion’, ‘doesn’t’ etc. If you use pos­i­tive search meth­ods the in­ter­net will tell you what­ever you want.

  More gen­er­ally, mar­ket timers lose hard. Set your al­lo­ca­tions such that you’re com­fortable with max draw­downs and plug money into it re­li­ably. Re­ces­sions last on av­er­age 9 months and have an av­er­age draw down of 20%, not much to worry about in the long run.

• romeostevensit 9 Jul 2019 5:57 UTC

  25 points

  on: “Ra­tion­al­iz­ing” and “Sit­ting Bolt Upright in Alarm.”

  I think peo­ple have alarm fa­tigue due to in­suffi­cient re­s­olu­tion of their pri­or­ity func­tion to make de­ci­sions with­out tons of effort. There are a large num­ber of things in the world that are deeply im­por­tant, but of up­stream im­por­tance is know­ing that there’s an or­der that will have to de facto be fol­lowed whether you are aware of the things that de­ter­mine said or­der or not. And that you won’t per­son­ally have time to get to al­most any of the ur­gent things.

  I’m point­ing this out be­cause lots of things shout­ing for at­ten­tion with high­est pri­or­ity sig­nals is a recipe for burnout, while at the same time we do in fact need peo­ple still ca­pa­ble of sit­ting bolt up­right oc­ca­sion­ally.

• romeostevensit 9 Jul 2019 0:47 UTC

  4 points

  in reply to: paulfchristiano's comment on: 87,000 Hours or: Thoughts on Home Ownership

  So is the bet for the con­struc­tion of a fi­nan­cial in­stru­ment that moves like a lev­er­aged case-shiller hous­ing in­dex for a se­lected city with track­ing er­ror + fees smaller by some fac­tor than the av­er­age fric­tional costs in the buy­ing mar­ket?

  edit: oh in this model we’re mostly con­cerned with the rent hedg­ing value rather than the spec­u­la­tive value (that be­ing seen as mostly a thing that sub­si­dizes the rent hedge here) so we’d re­ally want some­thing that pays you when the rent in­dex de­vi­ates sub­stan­tially from its trend. If we were happy with just hedg­ing the long term trend I think that could be done cheaply with fixed in­come lev­ered ap­pro­pri­ately, but I’d have to show that it won’t sud­denly change cor­re­la­tions un­der var­i­ous con­di­tions.

  edit 2: I think there’s some aether vari­able-ing go­ing on here. Ren­tal mar­ket are much more liquid and dis­play lower volatility than hous­ing mar­kets. So in or­der to pur­chase a hedge against rental price (cash flow) ap­pre­ci­a­tion we ex­pose our­selves to much more volatile net worth swings? That only makes sense for peo­ple with poor sav­ings (al­most all ‘nor­mal’ peo­ple. But nor­mal peo­ple aren’t try­ing to buy in the most ex­pen­sive mar­kets any­way).

• romeostevensit 9 Jul 2019 0:22 UTC

  8 points

  in reply to: paulfchristiano's comment on: 87,000 Hours or: Thoughts on Home Ownership

  The ac­tual num­bers are some­what idiosyn­cratic but my main point is that in the pro­cess of in­ves­ti­gat­ing the num­bers most peo­ple don’t eval­u­ate down­sides be­cause they don’t oc­cur to them. Once these costs are taken into ac­count the marginal buyer will flip on the de­ci­sion. In ex­tremely hot mar­kets you are much more likely to be like the marginal buyer rather than the av­er­age buyer.