hamnox

• hamnox 8 Apr 2019 17:19 UTC

  1 point

  on: How did Volde­mort learn the hor­crux spell?

  The in­ter­dict of Mer­lin stops the in­dis­crim­i­nate spread of high level magic, for a defi­ni­tion of high-level that is rel­a­tive to the learner. It is prob­a­bly pos­si­ble in prin­ci­ple to write down a se­ries of hints that lead the reader to figure out enough of the con­cepts them­selves that they can read the next level of hints. I ex­pect lots of peo­ple *try* writ­ing down their ma­te­rial and leav­ing such a trail of hints, and while it usu­ally fails Rid­dle prob­a­bly in­de­pen­dently figured out enough stuff to bridge the gaps.

• hamnox 6 Mar 2019 22:01 UTC

  11 points

  on: What would a zetetic ex­pla­na­tion be for the ra­tio­nal­ity com­mu­nity?

  thing I started typ­ing out:

  A philos­o­phy pro­fes­sor of Oxford Univer­sity founded the Fu­ture of Hu­man­ity In­sti­tute to study what we can do now to en­sure a long flour­ish­ing fu­ture. (Cu­ri­ousity, what else did they do?) One of the effects this group had was to spin off (how?) a group-writ­ten blog, Over­com­ingBias.com, ded­i­cated to the gen­eral theme of how to move our be­liefs closer to re­al­ity.

  One writer, Eliezer Yud­kowsky, worked on Ar­tifi­cial In­tel­li­gence and wanted to warn other peo­ple about dan­gers he’d re­al­ized could come from it. When he tried to talk about the dan­gers, he found that not only did peo­ple not know the dan­gers, they did not un­der­stand the ideas nec­es­sary to un­der­stand the ex­pla­na­tion of the dan­gers. Be­fore he could ex­plain the ideas he thought were most im­por­tant, he had to ex­plain a lot of smaller ideas that built up to it.

  Yud­kowsky’s writ­ing cov­ered a va­ri­ety of top­ics, yet made them all come to­gether to feel like part of the same deep philos­o­phy in the spirit of be­com­ing “less wrong” in one’s un­der­stand­ing of re­al­ity. As his writ­ing gained pop­u­lar­ity, he moved it to a new blog­ging web­site, Less­wrong.com, which any­body could post to.

  Al­most ev­ery­one who par­ti­ci­pated in this com­mu­nity in the early days had read Eliezer’s posts. Whether or not a per­son agreed with his ideas, his posts on var­i­ous top­ics were iconic and pre­cise; they be­came a com­mon ba­sis to start im­por­tant con­ver­sa­tions on. Other peo­ple’s con­tent filled gaps in and built on this com­mon canon. Peo­ple liked hav­ing this com­mon ba­sis enough to try to share it with peo­ple they knew in ev­ery­day life or prefer­en­tially talk with peo­ple who had already read it.

  [Zet: THEN CFAR. THEN HPMOR… Or was it the other way around?]

  some source here and here, I am so bad at not pla­gia­riz­ing.

[Question] What would a zetetic ex­pla­na­tion be for the ra­tio­nal­ity com­mu­nity?

First steps of a ra­tio­nal­ity skill bootstrap

• hamnox 6 Feb 2019 22:01 UTC

  4 points

  on: If Ra­tion­al­ity can be likened to a ‘Mar­tial Art’, what would be the Forms?

  A thing I fre­quently do is liter­ally just take an ob­ject and start ob­serv­ing things about it. No­tice what I no­tice. No­tice what con­clu­sions I gen­er­ate, what as­sump­tions they’re built off of, and what parts of my brain they come from. No­tice in­ter­ac­tions with other men­tal mod­els, no­tice what hy­pothe­ses are testable.

  Also try some kind of pre­ci­sion phys­i­cal art, be­cause the brain’s pri­mary job is and always will be mov­ing your body. There’s a lot of prior art in how to move your body bet­ter, and I would be sur­prised if fine-tun­ing that didn’t trans­late at least a lit­tle bit into in­creased in­tel­lec­tual fit­ness.

  (I should take my own ad­vice, meh.)

• hamnox 3 Sep 2018 0:50 UTC

  14 points

  on: Open Thread Septem­ber 2018

  Folk val­ues—the qual­ities of the “I love sci­ence” crowd as con­trasted to the qual­ities of ac­tual, ex­cep­tional sci­en­tists—mat­ter too. The com­mon folk out­num­ber the epic heroes.

  This holds true even if you be­lieve that ev­ery­one can be­come an epic hero! Peo­ple need to know, rather than guess and hope, that walk­ing the path to be­com­ing an epic hero might look and feel rather differ­ent than do­ing ac­tive epic hero­ing. In the­ory one ought to be able to de­rive the ap­pro­pri­ate in­stru­men­tal goals from the ter­mi­nal goal, but in prac­tice peo­ple very fre­quently mess this up.

  The gen­eral crowd has a differ­ent job than the in­ner cir­cle, and treat­ing this differ­ence as or­thog­o­nal prop­a­gates fewer er­rors than treat­ing it as a mat­ter of de­gree.

  Folk ra­tio­nal­ity needs to strongly pro­tect against in­fo­haz­ards un­til one gets a chance to de­velop less vuln­er­a­ble in­ter­nal habits. Folk ra­tio­nal­ity needs to cel­e­brate suc­cess­fully satis­fic­ing goals and iden­ti­fy­ing pi­cas rather than go­ing for hard op­ti­miza­tion be­cause am­a­teur min-max­ing just spawns Good­hart demons ev­ery which way. Folk ra­tio­nal­ity needs to prize keep­ing so­cial com­mit­ments and good con­flict me­di­a­tion tools; it needs to honor so­cial work­ers straight­for­wardly ad­dress­ing so­cial or re­source prob­lems. Folk ra­tio­nal­ity needs lu­minos­ity, and ther­apy. Folk ra­tio­nal­ity should also have civic duty of proac­tive per­sonal data col­lec­tion, cheer­ing on repli­ca­tions, par­ti­ci­pat­ing in RCTs, and not ghost­ing or lizard­man­ning sur­veys… be­cause sci­ence needs to get done d’arvit.

  In­ter­ested in cruxing

• hamnox 1 Sep 2018 15:15 UTC

  7 points

  on: Why I Don’t Like Scotty Slate Star and His Codex and So Can You

  • Will there be a record­ing or notes taken & posted some­where?

  • Can there a phone or web pres­ence op­tion?

• hamnox 9 Aug 2018 23:12 UTC

  10 points

  on: Strate­gies of Per­sonal Growth

  I like hav­ing these dis­tinc­tions laid out to think about. While it’s on my mind I’d like to share an ex­ten­sion of Brienne’s quad­rants I’d made in my own notes.

  To “Easy vs. Difficult” and “Fast vs. Slow”, I added a third di­men­sion of “Hype vs. Sig­nal”. A grand epiphany can turn out to be in­sight porn. A long gruel to at­tain wiz­ardry could be an in­vest­ment scam. Bug patches can be sur­face-level fads. Tor­toise skill prac­tice might be lo­tus-eat­ing dis­trac­tion.

  (I may have been a bit dis­illu­sioned with ra­tio­nal­ity lore at the time I named these. Be­cause yes, it *was* de­mor­al­iz­ing to get 2-3% re­turns when I ex­pected bursts of 300%.)

  A use­ful core can have many sub­tly-off in­stan­ti­a­tions. The ex­pected sig­nal-to-noise ra­tio mat­ters, when you’re figur­ing out where it makes sense to fo­cus your efforts.

• hamnox 18 Mar 2018 23:27 UTC

  24 points

  on: Mus­ings on LessWrong Peer Review

  Thoughts on the un­suit­abil­ity of adding more thresh­olds of qual­ity con­trol:

  The idea of pro­mot­ing things up a nar­row­ing hi­er­ar­chy falls into one’s lap as an easy fix. It won’t solve the prob­lem, I think though, of how a good ini­tial pro­posal of an idea just does not thrive on the same met­rics as the write-up of com­mon con­text needs to. A first draft is not just a shoddy ver­sion of a fi­nal draft; it ac­tu­ally does some­thing differ­ent than the fi­nal copy. In the pro­gres­sion of idea gen­er­a­tion to canon, some­one has to do the trans­for­ma­tion work. The same per­son could do it; gen­eral skill /​ abil­ity over the whole pro­gres­sion ex­ists and I gen­uinely hope we find a way to nur­ture it in peo­ple. I doubt the sense of count­ing on peo­ple’s skill (and will­ing­ness) at each step be­ing trans­fer­able.

• hamnox 22 Dec 2017 8:59 UTC

  5 points

  in reply to: Benquo's comment on: More Dakka

  With mela­tonin, it’s not any­where as sim­ple as “too strong” an effect. Me­la­tonin is typ­i­cally sold at high doses that don’t re­ally the proper effect at all, which re­sults in peo­ple de­cid­ing to in­crease the dose even more un­til they get a hard knock­out effect which looks like the de­sired thing if you’re des­per­ate and squint­ing, but… no.

• hamnox 19 Dec 2017 6:03 UTC

  15 points

  in reply to: PDV's comment on: Bay Sols­tice 2017 Flopped

  The point is to stop talk­ing about words, and start talk­ing about re­al­ity.

• hamnox 19 Dec 2017 6:00 UTC

  17 points

  in reply to: Raemon's comment on: Bay Sols­tice 2017 Flopped

  “maybe peo­ple flag when they dis­agree but don’t get into a pro­tracted con­ver­sa­tion un­til af­ter­wards.”

  yup that’s what I meant

  At the be­gin­ning, note speci­fi­cally that we’re do­ing the rit­ual thing, that we are tel­ling sto­ries/​songs that are some­what hack­ing our brains, that this only re­ally works if you lean into it with your sys­tem 1, and that we’re try­ing to do this wisely.

  Yes, make it clear what’s go­ing to hap­pen so peo­ple can opt in or out sanely.

  Maybe en­couage peo­ple to do some kind of “silent but visi­ble dis­agree­ment” thing if they dis­agree (I’m not sure if this would work with­out ru­in­ing things)

  Make a ded­i­cated space for peo­ple af­ter­wards to dis­cuss /​ dis­agree /​ ar­gue.

  Th­ese all sound good

• hamnox 19 Dec 2017 4:47 UTC

  4 points

  in reply to: Benquo's comment on: Seed­ing a pro­duc­tive cul­ture: a work­ing hypothesis

  Your less­wrong link is not a less­wrong link

• hamnox 19 Dec 2017 4:40 UTC

  29 points

  on: Bay Sols­tice 2017 Flopped

  This Sols­tice had me think­ing on what I had imag­ined, when I first read about Sols­tice. When I was young and dreamed that we were go­ing to Do This Ra­tion­ally.

  The or­ga­niz­ers would have ac­tual mod­els about what brain but­tons we were push­ing to what effect—en­tan­gling the wellbe­ing effect of light with a spe­cific nar­ra­tive of hu­man progress, evoca­tive and non-rep­re­sen­ta­tive sto­ries, in­duc­ing ex­is­ten­tial fears and di­rect­ing peo­ple to soothe them through so­cial bond­ing with a par­tic­u­lar crowd, deep rhyth­mic res­o­nances that just hit straight to sys-1′s sense of “re­ally big”, etc.—and share them ahead of time to en­able in­formed con­sent.

  The event would en­shrine peo­ple’s right and duty to con­scien­tiously ob­ject at any point they feel the goals or epistemics have drifted in an un­de­sir­able di­rec­tion:

  • put in anti-Asch-con­for­mity plants

  • in­ten­tion­ally give up/​change the most be­loved part of the rit­ual from year to year to avoid sta­tus quo bias

  • give the au­di­ence 5 min­utes to ac­tu­ally con­sider whether to do this thing or what they need to do instead

  • make a place for ob­jec­tors to stand and be counted in­stead of silently bounc­ing out

    • in­voke cu­ri­ousity about (but do not de­mand on-the-spot jus­tifi­ca­tions for) why.

  This year was lovely perfor­mances, nice speeches, an in­ter­est­ing ac­tivity, a good evening of food and en­ter­tain­ment. But I did not get what I wanted.

  I’m not sure how much of what I wanted is ac­tu­ally doable.

  But when I hear load­bear­ing speeches lifted straight from the pre­vi­ous year’s lineup—un­changed from the se­quences, I won­der—have we learned any­thing new at all? I hear a tidy lit­tle myth like The God­dess of Every­thing Else and worry about false ap­pear­ance of con­sen­sus.

  I know the arc of Bay’s Sols­tice has moved more to­wards em­pha­siz­ing com­mu­nity than x-risk, yet I do not think we have changed our rit­ual tool­ing to match this shift. One or two ex­tra in­ter­ac­tions hap­pened that day, but is the au­di­ence any more em­pow­ered to act as a com­mu­nity than it was be­fore? How could we have re­li­ably solved the Tarot Card prob­lem? What con­ver­sa­tions need to con­tinue hap­pen­ing af­ter Sols­tice and how will they hap­pen?

  It would be lu­dicrous to rein­vent some­thing of Sols­tice’s mag­ni­tude ev­ery year. But where there’s risk of your lo­gis­tics and epistemics clash­ing, I think we should err more on the side of vastly sim­plify­ing events than on the side of slop­pier epistemics.

  I think next year I only want yin med­i­ta­tion, oaths sworn by can­dlelight, and a playlist of per­son­ally mean­ingful songs. Per­haps I’ll do it with five friends and one lonely stranger.

• hamnox 1 Dec 2017 20:21 UTC

  4 points

  in reply to: Hazard's comment on: Creat­ing Space to Cul­ti­vate Skill

  I don’t want to bake the in­tended les­son into the prac­tice; if I have to tell you what the moral of a story is then it’s not do­ing a very good job of mak­ing its own point.

• hamnox 1 Dec 2017 20:18 UTC

  4 points

  in reply to: magfrump's comment on: Creat­ing Space to Cul­ti­vate Skill

  It’s an­tag­o­nis­tic to that end be­cause a spe­cific skill is the wrong case study.

  I think by the time some­one de­cides on a spe­cific skill they can have already baked in re­ally crit­i­cal mis­takes.

  They aim for a goal that only vaguely fits what they re­ally care about, so their clar­ity and mo­ti­va­tion bleed out as they spend time on it. (I want to stop feel­ing like an im­poster! Guitar play­ers are ob­jec­tively im­pres­sive, let’s learn gui­tar!)

  Or they fail to pri­ori­tize effec­tively, so open 50 pro­jects and make scant progress in any of them.

  Or they have a bro­ken un­der­stand­ing of what learn­ing looks like, so in­ten­tion­ally try­ing to learn mostly stifles the ac­tual pro­cess of dis­cov­ery and in­te­gra­tion. (I must ace this course, or must mem­o­rize all the syn­tax and best prac­tices, or must dive straight into com­pli­cated Real Pro­ject use.)

  Or they for­get to have a gears model of suc­cess­fully ac­quiring the skill at all, and in­stead half-ass­edly hope it will fall out of the sky if they perform the right ges­tures. (If I give my­self a few hours of ex­po­sure ther­apy to em­barass­edly drop­ping clubs, I’ll learn to jug­gle right?)

  The right class looks more like:

  You feel like an im­poster. All your suc­cesses are flukes and light­ning won’t strike twice, but on this day you’re set­ting your­self up to repli­cate the con­di­tions for suc­cess any­ways. All your failures were in­her­ent and in­evitable, but on this day you’re giv­ing your­self a chance to no­tice small things that vary the ex­pe­rience.

  YOU WANT TO DO ALL THE THINGS! When you clean up the space and leave out only the things you need to­day, that’s still a lot of things! When spend­ing dou­ble
  time on ev­ery­thing, you also wind up get­ting to fewer of them than you ex­pected. After a dozen iter­a­tions of this you have prob­a­bly got­ten the hint that your cal­ibra­tion is off. There’s ob­vi­ous solu­tion av­enues you can try once you have a way to gauge the prob­lem—hide ev­ery­thing by de­fault, limit your­self to a few solid pro­jects you pick ev­ery time, tighten your stan­dards of pick­ing, use a new se­lec­tion strat­egy en­tirely, ac­cept the bias and work around it.

  You have cleared your desk­top, phys­i­cal and digi­tal, and closed all your open tabs. You’re work­ing on this course for a half hour. Yay cel­e­brate, maybe take a breather. You’re go­ing to work on this course for an­other half hour.

  You’ve been go­ing over syn­tax for 10 min­utes. Wow this re­ally sucks and you hate it. What could you do that would be more re­ward­ing and still, in some sense, be equiv­a­lent to go­ing over syn­tax for 10 min­utes? Well… what was the im­por­tant part here? Was it the spe­cific in­for­ma­tion—can you read in a differ­ent way, switch up note-tak­ing styles, make anki cards? Was it the re­view­ing—could you use differ­ent sources, pri­ori­tize differ­ent in­for­ma­tion, go over some other topic en­tirely? Was it un­der­stand­ing the lan­guage bet­ter—maybe you could go through a tu­to­rial, or read source code, or mess around in a REPL in­stead.

  After a cou­ple dozen cases of beat­ing your head against a big pro­ject plan you don’t un­der­stand enough to work on, then be­ing faced with the prospect of re­peat­ing the ex­er­cise know­ing you still can’t make progress, you start to ap­pre­ci­ate the value of lev­er­ag­ing prior knowl­edge and perform­ing small em­piri­cal ex­per­i­ments.

  Your friends are rav­ing about lan­guage X, so you read an in­tro on it. You read a differ­ent per­spec­tive. It does sound neat, and you’ve got time to mess around. You go through a ba­sic pro­ject setup to get a feel for it. You set it up again and get slightly more com­fortable. Maybe you end there, satis­fied, and ex­plore other things… maybe you want to try to do some­thing a bit more com­pli­cated. Maybe you try and strug­gle at both at­tempts so you shift to read­ing docs or work­ing from an ex­am­ple. You stay fo­cused on spe­cific ac­tion­able goals at ev­ery step, keep­ing in tune with they pan out.

  You spend way too much time stuck on the task ‘get rid of dis­trac­tions’ and don’t get around to do­ing any­thing use­ful. Re­mov­ing dis­trac­tions takes un­sus­tain­able amounts of up­front en­ergy, or there’s a de­bil­i­tat­ing back­log of clean­ing and or­ga­niz­ing to get through. Of course this wasn’t less true any other time that you needed to be free of dis­trac­tion. So, you pick a stan­dard to care about and painstak­ingly fa­mil­iarize your­self with the ac­tual costs and benefits of reach­ing it. Work­ing through the back­log in pieces will just be what Dou­bleDay is for you un­til that’s no longer your bot­tle­neck.

  You’re to­tally in a half-ass, satis­fice mind­set about this prob­lem. But you’ve already satis­ficed a nice en­vi­ron­ment and are mak­ing plau­si­bly de­ni­able dummy at­tempts re­gard­less, so you could stand to oc­ca­sion­ally throw in a sincere try. Do some­thing that has a real chance of work­ing whether or not it fits the nar­ra­tive. No need for any­one to know your se­cret au­dac­ity.

• hamnox 30 Nov 2017 3:36 UTC

  43 points

  on: Ex­am­ples of Miti­gat­ing As­sump­tion Risk

  The most com­mon­sense ex­am­ple of mak­ing as­sump­tions ir­rele­vant I’ve heard of is from weapons safety: always act as if the gun is loaded.

  See also

  • Cut away from yourself

  • Rule #2: Dou­ble tap

  • Eat­ing Mealsquares in­stead of Soylent

• hamnox 28 Oct 2017 6:32 UTC

  8 points

  in reply to: ESRogs's comment on: The MTG Color Wheel

  Slack feels dis­tinct from keep­ing a buffer—Black would definitely think of gath­er­ing ex­tra re­sources to overde­ter­mine their vic­tory in a way that they might not think to in gen­eral take the less-than-max­i­mal path to­wards ac­com­plish­ing their stated goals. Gather­ing re­sources is just gath­er­ing re­sources.

  I pro­mote Red Slack. Break your chains, be­fore they break you.

• hamnox 27 Oct 2017 16:53 UTC

  7 points

  in reply to: Rossin's comment on: A Thou­sand “Heads” In A Row

  Track­ing helps avoid some bias.

  If you for­get that the data col­lec­tion hap­pens through se­lec­tive ac­tion and the data’s mean­ing is seen through a flawed lens, though, then your ‘ob­jec­tive view’ can wind up more sharply skewed than your vague gut feels.

• hamnox 27 Oct 2017 16:39 UTC

  8 points

  on: Refer­ences & Re­sources for LessWrong

  Want to note: I no­ticed the cat­e­gory “memetic haz­ard” and started im­me­di­ately skim­ming the page to find ev­ery­thing la­beled as such. Some­thing is wrong with my rea­son­ing here—

  It wasn’t the worst im­pulse to fol­low af­ter all, since the cat­e­gory means some­thing like con­tro­ver­sial or fic­tional. Ex­cept… “memetic haz­ard” is a mean­ingful warn­ing. I would pre­fer it keeps its value as a sig­nal.