Mati_Roy(Mati Roy)

• Mati_Roy 19 Feb 2020 4:23 UTC

  7 points

  on: Dist­in­guish­ing defi­ni­tions of takeoff

  try­ing to put this in my own words to re­mem­ber it

  so differ­ent axes for take-off dy­nam­ics in­clude:

  • time span: phys­i­cal time, eco­nomic time, poli­ti­cal time, AI time (de­vel­op­ment speed of front run­ners over oth­ers)

  • shape of the take-off curve, ex.: ex­po­nen­tial, S-curves, lin­ear, etc.

  • mo­nop­o­lis­tic effect: do front run­ners be­come less likely to be out­com­peted as they grow? how many large play­ers will there be? also: how will this change? ex.: it could be that AI doesn’t have strong mo­nop­o­lis­tic effect un­til you reach a cer­tain level

  mea­sure­ment to quan­tify:

  • AI progress: GDP, de­ci­sive strate­gic advantage

  • AI progress speed: time un­til AI hav­ing more power than the rest of hu­man­ity /​ time un­til solv­ing the con­trol problem

  re­lated:

  • im­pact on fore­cast­ing capabilities

• Mati_Roy 19 Feb 2020 3:54 UTC

  1 point

  on: Dist­in­guish­ing defi­ni­tions of takeoff

  the frac­tion of global eco­nomic ac­tivity at­tributable to au­tonomous AI sys­tems will rise

  I thought a bit about this, but haven’t figured it out: how can this be mea­sured? if AI is com­modi­tized, AI com­pa­nies won’t make a profit from it. AI re­searchers might make more money, but likely not more than how­ever much it would cost to train more AI re­searchers (or some­thing like that). maybe we can see which in­dus­tries have their price re­duced be­cause of AI, and count this as a lower bound for the con­sumer sur­plus cre­ated by AI. what else?

[Question] What frac­tion of work time in the world is done at a com­puter?

• Mati_Roy 13 Feb 2020 3:23 UTC

  1 point

  in reply to: Andrew Jacob Sauer's comment on: What are the risks of hav­ing your genome pub­li­cly available?

  I per­son­ally would rather an FAI be able to bring me back than pre­vent­ing an UFAI from do­ing so

• Mati_Roy 12 Feb 2020 3:43 UTC

  3 points

  on: The housekeeper

  when we were more peo­ple at the Macro­scope (https://​​macro­scope­house.com/​​), the last tool we were us­ing was weighted re­verse auc­tions: https://​​matiroy.com/​​writ­ings/​​weighted-re­verse-auc­tion.html

[Question] What are the risks of hav­ing your genome pub­li­cly available?

• Mati_Roy 15 Jan 2020 23:55 UTC

  2 points

  in reply to: Richard_Kennaway's comment on: What plau­si­ble be­liefs do you think could likely get some­one di­ag­nosed with a men­tal ill­ness by a psy­chi­a­trist?

  Thanks for the clar­ify­ing ques­tion. It’s the im­per­sonal “you”; al­though I think speci­fi­cally the per­sonal “you” would be even more in­ter­est­ing. I up­dated the ques­tion.

[Question] How use­ful do you think par­ti­ci­pat­ing to the Hu­man Micro­biome Pro­ject would be?

• Mati_Roy 15 Jan 2020 11:14 UTC

  4 points

  on: What plau­si­ble be­liefs do you think could likely get some­one di­ag­nosed with a men­tal ill­ness by a psy­chi­a­trist?

  Believ­ing you’re likely in a simulation

  • data point: https://​​www.youtube.com/​​watch?v=sIt-w8BwvV4 (see de­scrip­tion as well)

  Var­i­ous lifestyles: be­liev­ing that polyamory can be an healthy re­la­tion­ship struc­ture; be­liev­ing prac­tic­ing var­i­ous kinks and sex­ual in­ter­ests can be healthy

[Question] What plau­si­ble be­liefs do you think could likely get some­one di­ag­nosed with a men­tal ill­ness by a psy­chi­a­trist?

• Mati_Roy 15 Jan 2020 8:14 UTC

  3 points

  in reply to: Kubera Imxal's comment on: What are be­liefs you wouldn’t want (or would feel ap­pre­hen­sive about be­ing) pub­lic if you had (or have) them?

  Thank you for your an­swer! I’d want you to ex­pand on all points, but will be more rea­son­able and pri­ori­tize my cu­ri­os­ity; re 2: does it cre­ates more failures than suc­cesses be­cause a) more peo­ple start a startup as a re­sult, think­ing it will be eas­ier than it is, b) harder to get profit be­cause com­pe­ti­tion is more fierce, or c) other rea­sons?

[Question] What are be­liefs you wouldn’t want (or would feel ap­pre­hen­sive about be­ing) pub­lic if you had (or have) them?

• Mati_Roy 4 Jan 2020 2:53 UTC

  4 points

  in reply to: davidad's comment on: Whole Brain Emu­la­tion: Look­ing At Progress On C. elgans

  For your in­for­ma­tion, the above two links were judged as wrong

  @davi­dad, any up­dates on your work?

• Mati_Roy 23 Dec 2019 3:30 UTC

  2 points

  on: What could a World Unifi­ca­tion In­dex track to mea­sure how unified the world is, was, and is be­com­ing?

  I have the im­pres­sion that the more unified the world is, a) the less likely wars are, b) the eas­ier it is to co­or­di­nate to solve global prob­lems. It would also cre­ate more eco­nomic pros­per­ity. But some forms of global­i­sa­tion could also make our civil­i­sa­tion less ro­bust. An­ders Sand­berg says the ex­tent a fea­ture is global­ize should be pro­por­tional to the ex­tent that fea­ture’s im­pact is global (https://​​www.youtube.com/​​watch?v=91FKgBURAV0).

• Mati_Roy 23 Dec 2019 3:22 UTC

  1 point

  in reply to: habryka's comment on: LessWrong 2.0 Fea­ture Roadmap & Fea­ture Suggestions

  awe­some, thanks a lot! :)

• Mati_Roy 23 Dec 2019 2:45 UTC

  1 point

  in reply to: Mati_Roy's comment on: LessWrong 2.0 Fea­ture Roadmap & Fea­ture Suggestions

  @habryka, I don’t see In­ter­com, but I’m still in­ter­ested in hav­ing my user­name changed; thank you!

• Mati_Roy 22 Dec 2019 6:03 UTC

  1 point

  in reply to: quanticle's comment on: If Antarc­tic be­came hos­pitable to hu­mans, and con­se­quently re­ceived a mass mi­gra­tion, what are likely ways the Antarc­tic le­gal sys­tem could evolve?

  ah, right, the premise of my ques­tion might be wrong! do you think colonies is a more likely out­come? why? (will up­date the ques­tion’s de­scrip­tion)

• Mati_Roy 22 Dec 2019 4:56 UTC

  4 points

  on: What could a World Unifi­ca­tion In­dex track to mea­sure how unified the world is, was, and is be­com­ing?

  Some vari­ables I would be in­ter­ested in see­ing tracked in­clude:

  • the dis­tri­bu­tion on the num­ber of peo­ple a per­son can com­mu­ni­cate with (with­out trans­la­tors; both fluently and at a ba­sic level)

  • how good and easy to use trans­la­tors are get­ting (in­clud­ing au­dio<>text con­ver­sion)

  • how much coun­tries are legally merg­ing (ex.: EU) or frag­ment­ing (ex.: UK?), in­clud­ing level of de­sire to sep­a­rate (https://​​en.wikipe­dia.org/​​wiki/​​Lists_of_ac­tive_sep­a­ratist_move­ments)

  • how much art (mu­sic, books, movies) are con­sumed in­ter­na­tion­ally vs locally

  • how much other cul­tural el­e­ments are get­ting globally adopted (ex.: food, cloth­ing)

  • how much in­ter­na­tional prod­ucts /​ ser­vices are get­ting consumed

  • what frac­tion of peo­ple are mul­ti­ra­cial, and mul­ti­ra­cial to which ex­tent on average

[Question] What could a World Unifi­ca­tion In­dex track to mea­sure how unified the world is, was, and is be­com­ing?

• Mati_Roy 22 Dec 2019 4:35 UTC

  3 points

  on: If Antarc­tic be­came hos­pitable to hu­mans, and con­se­quently re­ceived a mass mi­gra­tion, what are likely ways the Antarc­tic le­gal sys­tem could evolve?

  Epistemic sta­tus: arm chair thinking

  I won­der if an Antarc­tic mass mi­gra­tion would lead to the UN work­ing closer to­gether to de­velop in­ter­na­tional laws, which would con­se­quently in­crease in­ter­na­tional co­op­er­a­tion.