10 posts I like in the 2018 Review

I see ba­si­cally ev­ery post that gets sub­mit­ted to LessWrong, whereas many users come in and read things more oc­ca­sion­ally, so I thought I’d list 10 posts I like in the 2018 Re­view that peo­ple might have missed.

I wasn’t able to write quick re­views of each of them, so this is more like a list of nom­i­na­tions. I’ve left out a few posts that I ex­pect will nat­u­rally be very pop­u­lar.

  1. Nam­ing the Name­less by Sarah Constantin

    • This points to a va­ri­ety of spe­cific mechanisms and gears by which cul­ture and aes­thet­ics af­fect our judg­ments and our choices, and com­bines them into an es­say that walked me much fur­ther than I was be­fore-hand in notic­ing these effects. I think it might be in my top 5 posts of 2018.

  2. Ex­plicit and Im­plicit Com­mu­ni­ca­tion by lionhearted

    • This has some pow­er­ful ar­gu­ments about when not to say things out loud—when not to make back­ground as­sump­tions ex­plicit—which I think is a re­ally pow­er­ful dat­a­point for a ra­tio­nal­ist to take on board. And the anti-Nazi ‘Sim­ple Sab­o­tage Field Man­ual’ is amaz­ing.

  3. Challenges to Chris­ti­ano’s ca­pa­bil­ity am­plifi­ca­tion pro­posal + Paul’s re­search agenda FAQ by Eliezer Yud­kowsky and zhu­keepa (re­spec­tively)

    • Th­ese to­gether are maybe the first time I re­ally un­der­stood what Paul’s ideas were. Really helpful. I’m mostly talk­ing about the Eliezer-Paul di­alogue, both on Eliezer’s post, and in the com­ments of Alex’s post. I ap­pre­ci­ate zhu­keepa putting in the ex­plana­tory work post as a nec­es­sary step for that di­alogue to con­tinue.

  4. Va­ri­eties of Ar­gu­men­ta­tive Ex­pe­rience by Scott Alexander

    • So. Many. Ex­am­ples. This is a strong step for­ward in con­cep­tu­al­is­ing ar­gu­ments and dis­agree­ments, writ­ten by some­one who’s read and been in­volved in an in­cred­ible amount of good (and bad) ones on the in­ter­net.

  5. Un­rol­ling so­cial metacog­ni­tion: Three lev­els of meta are not enough. by Academian

    • This very clearly lays out the iter­a­tive pro­cess by which so­cial emo­tions and at­ti­tudes are built up. Foun­da­tional for a lot of so­cial mod­el­ling.

  6. Act of Char­ity by Jessicata

    • A thought­ful di­alogue throw­ing into con­trast a lot of key in­tu­itions around de­cep­tion, self-de­cep­tion, and ethics.

  7. Public Po­si­tions and Pri­vate Guts by Vaniver

    • This post says a lot about how for­mal com­mu­ni­ca­tion ig­nores many messy parts of what it’s ac­tu­ally like be­ing a hu­man, that helps show how what com­mu­ni­ca­tion needs to be changes at scale.

  8. Op­ti­miza­tion Am­plifies by Scott Garrabrant

    • One of the most im­por­tant ideas IMO when think­ing about AI and AI al­ign­ment. Best sum­marised by the line: “I am not just say­ing that ad­ver­sar­ial op­ti­miza­tion makes small prob­a­bil­ities of failure large. I am say­ing that in gen­eral any op­ti­miza­tion at all messes with small prob­a­bil­ities and er­rors dras­ti­cally.”

  9. In­tel­li­gent So­cial Web by Valentine

    • Really well-writ­ten ex­pla­na­tion of a ba­sic way of un­der­stand­ing so­cial dy­nam­ics, that helps me un­der­stand a lot of my ex­pe­riences.

  10. Pre­dic­tion Mar­kets: When Do They Work by Zvi

    • This post com­mu­ni­cates a lot of Zvi’s taste and ex­pe­rience about how to de­sign ba­sic mar­kets—pre­dic­tion mar­kets, in par­tic­u­lar. Shaped a lot of my think­ing about this sub­ject, and has been a good guide when I’ve been (a lit­tle bit) in­volved in try­ing to build fore­cast­ing in­fras­truc­ture.

I’d be in­ter­ested to see posts that other users re­ally liked.

No comments.