Liron

• Liron 26 Oct 2020 13:59 UTC

  17 points

  on: Does play­ing hard to get work? AB test­ing for romance

  Cofounder of Re­la­tion­ship Hero here. There’s a sound un­der­ly­ing prin­ci­ple of courtship that PHTG taps into: That if your part­ner mod­els you as some­one with a high stan­dard that they need to put in effort to meet, then they’ll be more at­tracted to you.

  The prob­lem with try­ing to ap­ply any dat­ing tac­tic, even PHTG, is that courtship is a com­plex game with a lot of state and con­text. It’s very com­mon to be un­cal­ibrated and ap­ply a tac­tic that back­fires on you be­cause you weren’t aware of the over­all men­tal model that your part­ner had of you. So I’d have to ob­serve your in­ter­ac­tions and con­firm that “be­ing too easy” is a suffi­ciently ac­cu­rate one-di­men­sional pro­jec­tion of where you cur­rently stand with your part­ner, be­fore recom­mend­ing this one par­tic­u­lar tac­tic.

  In­stead of rely­ing on a toolbag of one-di­men­sional tac­tics, my recom­mended ap­proach is to fo­cus on un­der­stand­ing your part­ner’s men­tal model of you, and of their re­la­tion­ship with you, and of the re­la­tion­ship they’d want. Then you can strate­gize how to get the re­la­tion­ship you want, as­sum­ing it’s com­pat­i­ble with a kind of re­la­tion­ship they’d also want.

• Liron 6 Oct 2020 0:43 UTC

  11 points

  on: Ra­tion­al­ity and Cli­mate Change

  I agree with the other an­swers that say cli­mate change is a big deal and risky and worth a lot of re­sources and at­ten­tion, but it’s already get­ting a lot of re­sources and at­ten­tion, and it’s pretty low as an ex­is­ten­tial threat.

  Also, my im­pres­sion is that there are im­por­tant facts about how cli­mate change works that are al­most never men­tioned. For ex­am­ple, this claim that there are diminish­ing green­house effects to CO2: https://​​wattsup­with­that.com/​​2014/​​08/​​10/​​the-diminish­ing-in­fluence-of-in­creas­ing-car­bon-diox­ide-on-tem­per­a­ture/​​

  Also, I think most of the ac­tivism I see around cli­mate change is dumb and coun­ter­pro­duc­tive and mor­al­iz­ing, e.g. en­courag­ing per­sonal lifestyle sac­ri­fices.

• Liron 1 Oct 2020 13:30 UTC

  4 points

  in reply to: romeostevensit’s comment on: “Zero Sum” is a mis­nomer.

  I think they mean that ad tech (or per­haps a more con­sen­sus ex­am­ple is nukes) is a pris­oner’s dilemma, which is nonzero sum as op­posed to pos­i­tive/​nega­tive/​con­stant/​zero sum.

• Liron 1 Oct 2020 12:28 UTC

  4 points

  on: Open & Wel­come Thread—Septem­ber 2020

  Golden raises $14.5M. I wrote about Golden here as an ex­am­ple of the most com­mon startup failure mode: lack­ing a sin­gle well-formed use case. I’m con­fused about why some­one as savvy as Mark An­dreessen is tripling down and join­ing their board. I think he’s mak­ing a mis­take.

• Liron 28 Sep 2020 15:30 UTC

  10 points

  on: What are good ra­tio­nal­ity ex­er­cises?

  I was think­ing that if the se­quences and other LW clas­sics were a high school class, we could make some­thing like an SAT sub­ject test to check un­der­stand­ing/​fluency in the sub­ject, then that could be a badge on the site and po­ten­tially a good cre­den­tial to have in your ca­reer.

  The kinds of ques­tions could be like:

  1.

  If a US cit­i­zen has a le­gal way to save $500/​year on their taxes, but it re­quires spend­ing 1 hour/​day filling out bor­ing pa­per­work on 5 days of ev­ery week, should they do it?

  a. Vir­tu­ally ev­ery­one should do it

  b. A sig­nifi­cant frac­tion (10-90%) of the pop­u­la­tion should do it

  c. Vir­tu­ally no one should do it

  2.

  With suffi­cient ev­i­dence and a ra­tio­nal de­liber­a­tion pro­cess, is it pos­si­ble to be­come sure that the Loch Ness Mon­ster does/​doesn’t ex­ist?

  a. We CAN po­ten­tially be­come sure ei­ther way

  b. We CAN’T po­ten­tially be­come sure ei­ther way

  c. We can only po­ten­tially be­come sure that it DOES exist

  d. We can only po­ten­tially be­come sure that it DOESN’T exist

• Liron 26 Sep 2020 21:57 UTC

  3 points

  in reply to: Vanilla_cabs’s comment on: Honor­ing Petrov Day on LessWrong, in 2020

  Be­cause the driver ex­pects that the con­se­quences of run­ning you over will be asym­met­ri­cally bad for them (and you), com­pared to the rest of hu­man­ity. Ac­tions that take hu­man­ity down with you per­versely feel less ur­gent to avoid.

• Liron 26 Sep 2020 21:44 UTC

  16 points

  in reply to: habryka’s comment on: Honor­ing Petrov Day on LessWrong, in 2020

  Yeah I was off base there. The Nash Equil­ibrium is non­triv­ial be­cause some play­ers will challenge them­selves to “win” by trick­ing the group with but­ton ac­cess to push it. Plus prob­a­bly other rea­sons I haven’t thought of.

• Liron 26 Sep 2020 8:48 UTC

  11 points

  on: Honor­ing Petrov Day on LessWrong, in 2020

  Right now it seems like the Nash equil­ibrium is pretty sta­ble at ev­ery­one not press­ing the but­ton. Maybe we can simu­late adding in some lower-pri­or­ity yet still com­pel­ling pres­sure to press the but­ton, analo­gous to Petrov’s need to fol­low or­ders or the US’s need to pre­vent Rus­si­ans from sta­tion­ing nu­clear mis­siles in Cuba.

• Liron 24 Sep 2020 14:43 UTC

  2 points

  in reply to: ryan_b’s comment on: Are air­craft car­ri­ers su­per vuln­er­a­ble in a mod­ern war?

  Wow A+ an­swer, thanks!

[Question] Are air­craft car­ri­ers su­per vuln­er­a­ble in a mod­ern war?

• Liron 17 Sep 2020 22:25 UTC

  6 points

  on: Covid 9/​17: It’s Worse

  Now that so much of Cal­ifor­nia has burned, does that mean we’re in good shape for a few years of mild fire sea­sons?

• Liron 17 Sep 2020 10:57 UTC

  6 points

  on: Book Re­view: Work­ing With Contracts

  Thanks for the in­for­ma­tive and eas­ily-read­able sum­mary! This makes me wish Blinkist would add a check­box to en­able good episte­mol­ogy in their book sum­maries. Or more plau­si­bly, makes me want to con­tribute some sum­maries here too.

• Liron 29 Aug 2020 18:32 UTC

  3 points

  in reply to: Neel Nanda’s comment on: How to teach things well

  Re ex­am­ples of toy ex­am­ples with mov­ing parts:

  Andy Grove’s clas­sic book High Out­put Man­age­ment starts with the ex­am­ple of a diner that has to pro­duce break­fasts with cooked eggs, and keeps refer­ring to it to teach man­age­ment con­cepts.

  Minute Physics in­tro­duces a “Space­time Globe” to vi­su­al­ize space­time (the way a globe vi­su­al­izes the Earth’s sur­face) and refers to it of­ten start­ing at 3:25 in this video: https://​​youtu.be/​​1rLWVZVWfdY

• Liron 29 Aug 2020 18:24 UTC

  7 points

  on: How to teach things well

  My fa­vorite part was the ad­vice to high­light what’s im­por­tant, and it helped that you ap­plied your own ad­vice by high­light­ing that the most im­por­tant part of your les­son is the ad­vice to high­light the most im­por­tant part of your les­son.

  I’ve pre­vi­ously at­tempted to elab­o­rate on why ex­am­ples are helpful for teach­ing: https://​​www.less­wrong.com/​​posts/​​CD2kRisJcdBRLhrC5/​​the-power-to-teach-con­cepts-better

• Liron 24 Aug 2020 18:56 UTC

  2 points

  in reply to: gilch’s comment on: How to Pick the Wrong Side

  You can make money from an out-of-the-money short call even if the stock goes up

  Oh so in this case you’re sel­l­ing a call, but you can’t be said to be “short­ing the stock” be­cause you still benefit from a higher price?

• Liron 24 Aug 2020 18:54 UTC

  2 points

  in reply to: SatvikBeri’s comment on: What posts on fi­nance would your find helpful or in­ter­est­ing?

  Nice ones. The first is prob­a­bly the one that most ac­counts for funds like Ti­tan mar­ket­ing them­selves mis­lead­ingly (IMO), but the oth­ers are still im­por­tant caveats of the defi­ni­tion and good to know.

• Liron 24 Aug 2020 12:31 UTC

  2 points

  on: How to Pick the Wrong Side

  You are al­lowed to be bear­ish at times, but it’s bet­ter to sell calls or buy an­ti­cor­re­lated bonds and con­tinue to col­lect the risk pre­mium, than to short the stocks and be on the hook for the div­i­dends or buy­outs.

  Doesn’t “sell calls” mean the same thing as “short the stocks”?

• Liron 23 Aug 2020 14:08 UTC

  8 points

  on: What posts on fi­nance would your find helpful or in­ter­est­ing?

  I’ve been won­der­ing what are the caveats with rely­ing on Sharpe ra­tio to mea­sure how much risk was taken to get an in­vest­ment’s re­turns.

  For ex­am­ple, Ti­tan touts a high Sharpe ra­tio, and frames its mar­ket­ing like it’s bet­ter than the S&P in ev­ery way with no down­side: see https://​​www.less­wrong.com/​​posts/​​59oPYfFJjYn3BBBwi/​​titan-the-wealth­front-of-ac­tive-stock-pick­ing-what-s-the

  But doesn’t EMH im­ply that all Sharpe ra­tios long term will tend to the same av­er­age value, i.e. no one can have a suffi­ciently repli­ca­ble strat­egy that gives more re­turns with­out more risk?

  And in the case of Ti­tan, is the “catch” to their Sharpe ra­tio that they have higher down­side ex­po­sure to mo­men­tum re­ver­sal and mul­ti­ple con­trac­tion?

• Liron 9 Aug 2020 11:36 UTC

  6 points

  in reply to: Rafael Harth’s comment on: In­ner Align­ment: Ex­plain like I’m 12 Edition

  Hm ya I guess the causal­ity be­tween sex and ba­bies (even sex and visi­ble preg­nancy) is so far away in time that it’s tough to make a brain want to “make ba­bies”.

  But I don’t think com­pu­ta­tion­ally in­tractabil­ity of how ac­tions effect in­clu­sive ge­netic fit­ness is quite why evolu­tion made such crude heuris­tics. Be­cause if a brain un­der­stood that it was try­ing to max­i­mize that quan­tity, I think it could figure out “have a lot of sex” as a heuris­tic ap­proach with­out evolu­tion hard-cod­ing it in. And I think hu­mans ac­tu­ally do have some level of in-brain goals to have more de­scen­dants be­yond just hav­ing more sex. So I think these things like sex plea­sure are just perfor­mance op­ti­miza­tions to a men­tally tractable challenge.

  E.g. snakes quickly trig­ger­ing a fear reflex

• Liron 8 Aug 2020 23:22 UTC

  10 points

  on: In­ner Align­ment: Ex­plain like I’m 12 Edition

  Thanks for the ELI12, much ap­pre­ci­ated.

  evolu­tion’s ob­jec­tive of “max­i­mize in­clu­sive ge­netic fit­ness” is quite sim­ple, but it is still not rep­re­sented ex­plic­itly be­cause figur­ing out how ac­tions af­fect the ob­jec­tive is com­pu­ta­tion­ally hard

  This doesn’t seem like the bot­tle­neck in many situ­a­tions in prac­tice. For ex­am­ple, a lot of young men feel like they want to have as much sex as pos­si­ble, but not father as many kids as pos­si­ble. I’m not sure ex­actly what the rea­son is, but I don’t think it’s the com­pu­ta­tional difficulty of rep­re­sent­ing hav­ing kids vs. hav­ing sex, be­cause hu­mans already build a world model con­tain­ing the con­cept of “my kids”.

  It seems to me that one un­der-ap­pre­ci­ated as­pect of In­ner Align­ment is that, even if one had the one-true-util­ity-func­tion-that-is-all-you-need-to-pro­gram-into-AI, this would not, in fact, solve the al­ign­ment prob­lem, nor even the in­tent-al­ign­ment part. It would merely solve outer al­ign­ment (pro­vided the util­ity func­tion can be for­mal­ized).

  Damn, yep I for one un­der-ap­pre­ci­ated this for the past 12 years.

  What else have peo­ple said on this sub­ject? Do folks think that sce­nar­ios where we solve outer al­ign­ment most likely in­volve us not hav­ing to strug­gle much with in­ner al­ign­ment? Be­cause fully solv­ing outer al­ign­ment im­plies a lot of deep progress in al­ign­ment.