jacobjacob

• jacobjacob 13 Aug 2019 18:11 UTC

  1 point

  in reply to: ChristianKleineidam's comment on: Could we solve this email mess if we all moved to paid emails?

  How would this be an “in­ter­nal prac­tice”? The only way this would work would be to have peo­ple pub­li­cally post their earn ad­dresses.

  “In­ter­nal” in the sense of be­ing pri­mar­ily in­tended to solve in­ter­nal co­or­di­na­tion pur­poses and pri­mar­ily used in mes­sag­ing within the com­mu­nity.

  I think you un­der­rate the cost of weird­ness.

  You gave a par­tic­u­lar ex­am­ple of a causal path­way by which weird­ness leads to bad stuff, but it doesn’t re­ally cause me to change my mind be­cause I was already aware of it as a failure mode. What makes you think I un­der­rate the cost in com­par­i­son to the benefits of co­or­di­na­tion?

  While the kind of his sta­tus EA that might be con­tacted this way might get more emails then they pre­fer, it’s im­por­tant for them to be eas­ily con­tacted by out­siders be­cause that al­lows for valuable in­ter­ac­tions to hap­pen.

  They’d still have a nor­mal email. Though there is a risk of mov­ing to an equil­ibrium for non-paid emails get no at­ten­tion, and I haven’t thought that through in de­tail.

• jacobjacob 13 Aug 2019 17:39 UTC

  8 points

  in reply to: richard_ngo's comment on: Could we solve this email mess if we all moved to paid emails?

  It’s not clear to me that we are in a mess.

  Well, that’s why I’m post­ing this—to get some data and find out :)

  (I guess the ti­tle seemed to have turned a few peo­ple off, though)

  In hind­sight, I should have made the in­tended use-cases clearer in the post. I op­ti­mised for ship­ping it fast rather than not at all, but that had its costs.

  The rea­son I wrote this was ba­si­cally en­tirely mo­ti­vated by prob­lems I’ve en­coun­tered my­self.

  For ex­am­ple, I’ve spent this year try­ing to build an AI fore­cast­ing com­mu­nity, and faced the awk­ward prob­lem of need­ing a crit­i­cal mass of users, but at the same time re­cruit­ing from a base with high op­por­tu­nity costs and at­ten­tion value (largely EA). This usu­ally in­volves a pain-stak­ing pro­cess of think­ing care­fully about who we mes­sage and how much, and be­ing quite risk-averse and rather not mes­sag­ing peo­ple at all when we’re un­cer­tain. I would have loved the abil­ity to send paid emails, such that if we did hap­pen to spam peo­ple, they could just claim some com­pen­sa­tion. More­over, this is a scal­able strat­egy which would avoid the failure mode where pro­ject’s like ours which think a lot about at­ten­tion costs get de­pri­ori­tised in favour of pro­jects which don’t.

  As an­other ex­am­ple, I’ve con­sid­ered unilat­er­ally launch­ing ini­ti­a­tives that seemed im­por­tant and that no one was do­ing (like this!), but that very busy peo­ple might have reser­va­tions/​opinions about. This put me in a spot of mak­ing awk­ward trade-offs along the lines analysed above.

  In ad­di­tion to that, I added on some prob­lem that I’ve not per­son­ally ex­pe­rienced but which seemed like they should hap­pen due to ba­sic microe­co­nomics.

• jacobjacob 12 Aug 2019 12:22 UTC

  3 points

  in reply to: rosiecampbell's comment on: Could we solve this email mess if we all moved to paid emails?

  This is su­per helpful, thanks (and that’s a re­ally awe­some list of email hy­giene tips, I’ve saved it).

  I won­der whether ed­u­cat­ing and en­courag­ing good email hy­giene could be an eas­ier solu­tion (at least ini­tially).

  I think it would im­prove things on the mar­gin, and also has a much smaller risk of land­ing us in a worse equil­ibrium, so it seems ro­bustly good for peo­ple to do.

  Still, I’m not su­per ex­cited be­cause if you be­lieve that the ini­tial mess is a co­or­di­na­tion prob­lem, the solu­tion is not for in­di­vi­d­u­als to put in lots of effort to be helpful; but for ev­ery­one to jointly move to an­other game where the low-effort/​in­cen­tivised ac­tion is to co­op­er­ate rather than defect.

• jacobjacob 12 Aug 2019 11:09 UTC

  0 points

  in reply to: Khorton's comment on: Could we solve this email mess if we all moved to paid emails?

  On the topic of weird­ness: I ex­pect that if what I’m point­ing to is a real prob­lem, and paid emails would help the situ­a­tion, then the benefits from be­com­ing more effec­tive at co­or­di­nat­ing in­ter­nally would mas­sively out­weigh rep­u­ta­tional risks from in­creased weird­ness.

  I find it some­what hard to elu­ci­date the rea­sons I be­lieve this (though could try if you’d want me to), but some hand-wavy ex­am­ples are Paul Gra­ham’s thoughts that it’s al­most always a mis­take for star­tups to worry about com­peti­tors as op­posed to fo­cus­ing on build­ing a good product (see para­graph 4); as well as ex­tremely suc­ces­ful or­gani­sa­tions with pretty weird in­ter­nal prac­tices (e.g. Bridge­wa­ter, Ama­zon).

• jacobjacob 12 Aug 2019 10:59 UTC

  1 point

  in reply to: Khorton's comment on: Could we solve this email mess if we all moved to paid emails?

  I think the way to an­swer the ques­tion is: “given the dis­tri­bu­tion of equil­ibria we ex­pect fol­low­ing this change, what are the ex­pected costs and benefits, and how does that com­pare with the costs and benefits un­der the cur­rent equil­ibrium?” (as well as con­sid­er­ing strate­gic heuris­tics like avoid­ing ir­re­versible ac­tions and unilat­er­al­ist ac­tion.)

  I don’t up­date much on your com­ment since it feels like it’s just point­ing out a bunch of costs un­der a par­tic­u­lar new equil­ibrium, with­out en­gag­ing enough with how likely this is or what the benefits would be. [1]

  For ex­am­ple:

  -If Ju­lia Wise were pri­ori­tis­ing paid emails in her role re­gard­ing com­mu­nity health, is she more likely to miss emails from peo­ple on the periph­ery of EA or who have less money, who are po­ten­tially very vuln­er­a­ble?

  Here, by as­sump­tion, Ju­lia Wise already gets so many emails that she misses some/​has to pri­ori­tise. So the ques­tion is: what gets pri­ori­tised cur­rently, and would get pri­ori­tised un­der the new sys­tem? There would likely be a shift to­wards peo­ple with more money be­ing more able to get their is­sues heard—but I’d ex­pect it to be very small (e.g. ini­tial email costs of $5-$25 might be enough). It might also al­low her to find out about stuff she oth­er­wise wouldn’t (“I don’t know if this is worth your time, though it might be, and if it wasn’t, here’s $10 to offset the at­ten­tion cost”).

  Though to be clear, I’ve not thought a lot about com­mu­nity health mat­ters, and it’s not the area where I would pi­lot this.

  [1] To be clear, I’m not claiming you should do the en­tire anal­y­sis, this would be an iso­lated de­mand for rigor. Just to en­gage more with op­pos­ing points and say why they’re not con­vinc­ing.

• jacobjacob 12 Aug 2019 10:42 UTC

  2 points

  on: Could we solve this email mess if we all moved to paid emails?

  This was cross­posted to LessWrong, re­plac­ing all the men­tions of “EA” with “ra­tio­nal­ity”, mu­tatis mu­tan­dis.

• jacobjacob 12 Aug 2019 10:34 UTC

  1 point

  in reply to: aarongertler's comment on: Could we solve this email mess if we all moved to paid emails?

  I’m post­ing this as a first step to­wards col­lect­ing data. Poll is a good idea, thanks!

[Question] Could we solve this email mess if we all moved to paid emails?

AI Fore­cast­ing Dic­tionary (Fore­cast­ing in­fras­truc­ture, part 1)

• jacobjacob 4 Aug 2019 13:20 UTC

  1 point

  in reply to: Lukas_Finnveden's comment on: Con­ver­sa­tion on fore­cast­ing with Vaniver and Ozzie Gooen

  Thanks, fixed!

• jacobjacob 4 Aug 2019 13:19 UTC

  2 points

  in reply to: imben's comment on: Con­ver­sa­tion on fore­cast­ing with Vaniver and Ozzie Gooen

  I’m un­for­tu­nately only pub­lish­ing the tran­script at this time. The au­dio con­tains some sec­tions that were ed­ited out for pri­vacy rea­sons.

Con­ver­sa­tion on fore­cast­ing with Vaniver and Ozzie Gooen

• jacobjacob 30 May 2019 9:50 UTC

  1 point

  in reply to: aarongertler's comment on: Build­ing a suc­cess­ful econ­omy for col­lab­o­ra­tive cog­ni­tive work with high externalities

  Thanks, that’s great to hear.

  The prize has been go­ing on for a while, which seems im­por­tant, and I think the trans­parency of the Prize post is re­ally im­por­tant for mak­ing com­mon knowl­edge of what kind of work there is de­mand for. So over­all it’s pretty great.

  The struc­ture of feed­back looks to me like: “here’s the ob­ject-level con­tent of the post, and here are 2-3 rea­sons we liked it”. I think you could be more clear about what you want to in­cen­tivise. More pre­cisely, the cur­rent struc­ture doesn’t an­swer:

  • How strong were the rea­sons rel­a­tive to each other? (e.g. maybe re­mov­ing Rea­son A would make the per­son win 2nd prize in­stead of 1st, but re­mov­ing Rea­son B might make them win no prize)

  • Were the rea­sons only jointly suffi­cient to merit the prize, or might ac­com­plish­ing only one of them have worked?

  • What other prop­er­ties did the post dis­play, which did not merit the prize? For ex­am­ple, maybe prize-mer­it­ing posts tend to be quite long—even though length is not some­thing you want to in­cen­tivise on the mar­gin.

  • Why did the posts end up or­dered the way they did? Beyond “the black-box vot­ing pro­cess gave that ver­dict” :) Cur­rently I don’t know why SHOW was judged as de­serv­ing 4x the prize money of “The Case for the Ho­tel”, for ex­am­ple.

Build­ing a suc­cess­ful econ­omy for col­lab­o­ra­tive cog­ni­tive work with high externalities

• jacobjacob 30 Jan 2019 14:23 UTC

  6 points

  on: Invit­ing EA fore­cast­ers to AI fore­cast­ing platform

  [Note: I dou­ble-checked with the mod­er­a­tors be­fore post­ing this to en­sure it was not too “mar­ket­ingy”.]

Invit­ing EA fore­cast­ers to AI fore­cast­ing platform

• jacobjacob 16 Mar 2017 13:47 UTC

  2 points

  in reply to: MichaelPlant's comment on: Oxford Pri­ori­ti­sa­tion Pro­ject: Ver­sion 0

  1. When I and Tom came up with that, I don’t think we meant “be­lief” to be im­bued with the usual philo­soph­i­cal con­no­ta­tions. Rather, we in­tended it to mean some­thing like “ac­tion-guid­ing, in­tro­spec­tively ac­cessible rep­re­sen­ta­tion of a state of af­fairs ex­ist­ing in­de­pen­dently of whether it is queried”.

  When peo­ple ask me what I think about the world, I can of­ten come up with lots of in­tel­li­gent sound­ing an­swers—but it is un­for­tu­nately more rare that my ac­tual ac­tions, plans and nor­ma­tive eval­u­a­tions are some­how suit­ably hooked up to, and cru­cially de­pend upon, those an­swers.