“EA residencies” as an outreach activity

[This was par­tially in­spired by some ideas of Claire Za­bel’s. Thanks to Jes­sica McCurdy, Neel Nanda, Kuhan Jeyapra­gasan, Re­becca Baron, Joshua Mon­rad, Claire Za­bel, and the peo­ple who came on my Slate Star Codex road­trip for helpful com­ments.]

A few months ago, some EAs and I went on a trip to the East Coast to go to a bunch of Slate Star Codex mee­tups. I’m go­ing to quote that en­tire post here (with a cou­ple ed­its):

Our goals are:
- to meet promis­ing peo­ple at the SSC mee­tups and move them into the EA re­cruit­ing pipeline
- to spend some time with promis­ing new EAs, eg those at stu­dent groups, in the hope that spend­ing a few hours of fo­cused one-on-one time with one of us will help get them more into EA. Like, I think 80K finds peo­ple who are ex­cited about AI safety stuff but aren’t very knowl­edge­able about it yet; I think that those peo­ple can maybe get a lot out of a few hours’ con­ver­sa­tion with a few peo­ple who have worked pro­fes­sion­ally on this stuff.
- to visit EAs who are “in hold­ing” do­ing things like PhD or EtG tech jobs, with pos­si­ble good out­comes be­ing that they’ll be fired up wrt EA and more likely to do re­ally im­pact­ful EA stuff on a timescale of like a year, or that their im­proved (Bay Area/​pro­fes­sional EA) con­nec­tions make it eas­ier for them to spot good op­por­tu­ni­ties or move into do­ing more im­pact­ful work.
- (less pri­mary) to talk to hard­core EAs and swap ar­gu­ments and get to know each other better
Here’s why I think it’s worth us talk­ing to var­i­ous promis­ing new EAs and en­thu­si­as­tic EAs who haven’t worked in the EA scene full time:
- There are a lot of ac­cu­mu­lated ar­gu­ments about EA top­ics which I think it’s re­ally helpful to think about but which are hard to ac­cess when you only know EAs on the in­ter­net, be­cause those ar­gu­ments haven’t been writ­ten up clearly or at all, or be­cause their write­ups are hard to find and rely on back­ground knowl­edge that you don’t know how to ac­quire.
- A lot of the time, EAs pre­sent ver­sions of ar­gu­ments that are strong enough to con­vince you to ten­ta­tively think that it’s worth en­gag­ing se­ri­ously with the pos­si­bil­ity that the con­clu­sion might be true, but which have a bunch of holes in them that re­quire sub­stan­tial think­ing to fill. Some­times EAs (eg me) make the mis­take of con­flat­ing these two lev­els of strength of ar­gu­ment, and act as if peo­ple should be per­suaded by the ini­tial sketch. One way that I no­tice when I’m mak­ing this mis­take is by get­ting in ar­gu­ments with peo­ple who’ve thought about stuff more than me. I hope that talk­ing to more knowl­edge­able EAs might help some of the EAs we hang out spot holes in their un­der­stand­ing that might help im­prove their un­der­stand­ing and their epistemics.
Here is a rea­son that I think that hav­ing SF Bay Area EAs talk to ra­tio­nal­ists in these cities at SSC mee­tups is plau­si­bly worth­while:
When smart peo­ple are skep­ti­cal of some of my weird be­liefs, eg that AI x-risk is re­ally im­por­tant, or that they should con­sider work­ing on EA stuff, or that long term we should con­sider rad­i­cally re­struc­tur­ing the world to make it bet­ter for an­i­mals, a lot of the time their dis­agree­ment stems from some­thing true about the world that the ar­gu­ments they’ve seen didn’t ad­dress. This is hard to avoid be­cause if you try to write an ar­gu­ment that ad­dresses all the po­ten­tial con­cerns, it will be in­cred­ibly long. But this makes me think that it’s of­ten re­ally high im­pact for peo­ple who have thought a lot about these ar­gu­ments to talk to peo­ple who have heard of them but felt very un­per­suaded.

My pre­dic­tions mostly matched my im­pres­sions of what hap­pened.

But I think you might be able to get many of these benefits more effi­ciently by do­ing some­thing more like a res­i­dency, where you spend a rel­a­tively long time in each city, com­pared to do­ing a tour, for a few rea­sons:

  • If you want to spend 20 hours talk­ing to peo­ple in a city, you can do this more effi­ciently if you choose the best twenty hours in a ten day pe­riod, rather than hav­ing to do 20 con­tigu­ous hours. We missed out on a bunch of good op­por­tu­ni­ties be­cause we were only in each city for less than 24 hours. It meant that if peo­ple weren’t free right when we were in the city, we couldn’t talk to them.

  • Spend­ing more time in a city means that your travel costs are amor­tized over more hours.

  • If you only do a few hours of talk­ing to peo­ple each day, you might be able to do it with your so­cial time bud­get, and not sac­ri­fice many hours of nor­mal work.

Here’s an ex­am­ple plan:

  • For a month, Alice does her usual EA work from the US East Coast.

  • She stays mostly with EAs who she wants to spend more time get­ting to know.

  • She tries to get quite a lot of nor­mal work done while she’s there.

  • She tries to meet a bunch of EAs/​EA-rele­vant-peo­ple in her evenings. She goes to mee­tups and par­ties. Maybe she tries to get peo­ple to host a few ex­tra mee­tups and par­ties. (If she has some kind of rep­u­ta­tion that al­lows her to draw a crowd, this can also benefit the lo­cal EA groups.)

  • She stays mostly in Bos­ton, but she also spends a week­end in New York and spends a few days at Yale or at other places with a bunch of promis­ing EAs. (I think it can be nice to be in a place twice, so that the first time you meet peo­ple and the sec­ond time you meet the peo­ple those peo­ple wanted you to meet, and you can see some peo­ple twice, which I think is of­ten pretty helpful, so she does.)

Here are the main costs:

  • Lo­gis­ti­cal inconveniences

    • travel time

      • hope­fully this isn’t too bad if you can take trains (on which you can work), which you can do if you’re not in a hurry. One les­son from the SSC road­trip is that travel is sub­stan­tially more in­con­ve­nient when you have time pres­sure.

    • or­ga­niz­ing events and con­ver­sa­tions with people

  • Work effi­ciency penalties.

    • You might sleep worse. You could get around this by sleep­ing in ho­tels, or be­ing vigilant about en­sur­ing a good sleep en­vi­ron­ment when sleep­ing at the houses of EAs.

    • Some types of work are a lot more effi­cient with in-per­son col­lab­o­ra­tion. Maybe your job doesn’t in­volve that kind of work, or maybe you could ar­range to not do that kind of work for a month.

    • You might have a worse work en­vi­ron­ment. You could get around this if you made sure you had eg a large mon­i­tor ev­ery­where you were go­ing to be try­ing to work. Plan­ning to work out of a WeWork or some­thing could also be helpful for this.

  • Most peo­ple don’t like be­ing far away from home for long

Here are the prop­er­ties that I think make some­one a good can­di­date for this, other than be­ing an EA with in­ter­est­ing ex­pe­rience or per­spec­tives or similar:

  • It seems good if you’re gen­er­ally like­able, and in­ter­est­ing in con­ver­sa­tion, and have a good sense of so­cial ap­pro­pri­ate­ness in un­fa­mil­iar situ­a­tions. Neel Nanda, a Cam­bridge stu­dent who IMO has good judge­ment about EA out­reach, said: “I have a slight ap­pre­hen­sion that an out­sider from a more es­tab­lished com­mu­nity try­ing to do out­reach in smaller ones could come across as pa­tro­n­is­ing/​ar­ro­gant. I’d also be very con­cerned about the per­son com­ing across as weird (es­sen­tially the un­fa­mil­iar situ­a­tions point)”. I think this is a cru­cial point.

    • One ob­vi­ous sub­tlety here is that differ­ent peo­ple vibe differ­ently well with each other. Neel says, “I’d also ex­pect cul­tural differ­ences to be quite im­por­tant and a po­ten­tial source of failure modes, es­pe­cially if some­one has well cal­ibrated so­cial skills in their so­cial con­text but doesn’t prop­erly ac­count for the change in con­text, eg a Bay Area EA com­ing to the UK would prob­a­bly come across as very out­go­ing, po­ten­tially offputtingly weird/​ar­ro­gant/​con­fi­dent, and I think the baseline level of so­cial con­fi­dence eg re­quest­ing 1 on 1s prob­a­bly differs a lot”.

      • You could try to ad­dress this con­cern by go­ing with some­one with com­ple­men­tary so­cial skills

    • I think it’s also much eas­ier to pull this off if you un­der­stand things about the lo­cal EA scene like their in­ter­ests and cul­ture, and if you know some lo­cal mem­bers rea­son­ably well already, so that you can get rapid feed­back from them about how this is go­ing.

    • I think that this prob­a­bly works much bet­ter if the EA in res­i­dency isn’t try­ing to rep­re­sent all of EA, they’re just try­ing to rep­re­sent them­selves, as an EA who has opinions about things, and they make it clear that they are not a rep­re­sen­ta­tive of all of EA. If you do this, you’re less mak­ing a claim about your own le­gi­t­i­macy, you make it clearer that you’re not speak­ing for all of EA (which frees you up to share your non­stan­dard EA opinions), and peo­ple might jump less to the con­clu­sion that all EAs have the same be­liefs as you.

  • It’s good if talk­ing to strangers is more of­ten fun and fulfilling than stress­ful or tiring. You could try to set it up so that you were mostly talk­ing to peo­ple in ways that aren’t as tiring, by for ex­am­ple mostly talk­ing to peo­ple who you feel you can re­lax about.

  • Hav­ing work that you can do fairly effec­tively re­motely.

  • Hav­ing a lot of fa­mil­iar­ity with the EA com­mu­nity and EA cause ar­eas, and a lot of en­thu­si­asm for talk­ing and think­ing about this stuff.

  • I think that be­ing in­ter­ested in EA out­reach is some­what helpful, be­cause a lot of the peo­ple you talk the most to do EA out­reach work (eg run­ning EA groups) and are in­ter­ested in talk­ing about it.

  • Be­ing good at quickly get­ting a sense of peo­ple, so that you can eg spot whether this per­son should be in­tro­duced to a par­tic­u­lar per­son you know (be­ing well-con­nected also helps for this).

I think that it might be good for peo­ple to do res­i­den­cies like this. If you’re in­ter­ested in do­ing this, I’m in­ter­ested in talk­ing to you about it. I’m also in­ter­ested in talk­ing to peo­ple who live in places which they think would benefit from this type of thing.