KatjaGrace

• KatjaGrace 22 Apr 2020 21:22 UTC

  9 points

  on: My dad got stung by a bee, and mildly aller­gic. What are the trade­offs in­volved in de­cid­ing whether to have him go the the emer­gency room?

  In such a case, you might get many of the benefits with­out the covid risks from driv­ing to very close to the ER, then hang­ing out there and not go­ing in and risk­ing in­fec­tion un­less worse symp­toms de­velop, but be­ing able to act very fast if they do.

Dis­con­tin­u­ous progress in his­tory: an update

Atari early

• KatjaGrace 19 Feb 2020 1:11 UTC

  LW: 3 AF: 2

  AF

  in reply to: Daniel Kokotajlo’s comment on: Soft take­off can still lead to de­ci­sive strate­gic advantage

  1) Even if it counts as a DSA, I claim that it is not very in­ter­est­ing in the con­text of AI. DSAs of some­thing already al­most as large as the world are com­mon­place. For in­stance, in the ex­treme, the world minus any par­tic­u­lar per­son could take over the world if they wanted to. The con­cern with AI is that an ini­tially tiny en­tity might take over the world.

  2) My im­por­tant point is rather that your ’30 year’ num­ber is spe­cific to the start­ing size of the thing, and not just a gen­eral num­ber for get­ting a DSA. In par­tic­u­lar, it does not ap­ply to smaller things.

  3) Agree in­come doesn’t equal tak­ing over, though in the mod­ern world where much pur­chas­ing oc­curs, it is closer. Not clear to me that AI com­pa­nies do bet­ter as a frac­tion of the world in terms of mil­i­tary power than they do in terms of spend­ing.

• KatjaGrace 18 Feb 2020 6:56 UTC

  LW: 4 AF: 2

  AF

  on: Soft take­off can still lead to de­ci­sive strate­gic advantage

  The time it takes to get a DSA by grow­ing big­ger de­pends on how big you are to be­gin with. If I un­der­stand, you take your 30 years from con­sid­er­ing the largest coun­tries, which are not far from be­ing the size of the world, and then use it when talk­ing about AI pro­jects that are much smaller (e.g. a billion dol­lars a year sug­gests about ¹⁄_100,000 of the world). If you start from a situ­a­tion of an AI pro­ject be­ing three dou­blings from tak­ing over the world say, then most of the ques­tion of how it came to have a DSA seems to be the ques­tion of how it grew the other sev­en­teen dou­blings. (Per­haps you are think­ing of an ini­tially large coun­try grow­ing fast via AI? Do we then have to imag­ine that all of the coun­try’s re­sources are go­ing into AI?)

• KatjaGrace 8 Dec 2019 8:12 UTC

  15 points

  on: LW For Ex­ter­nal Com­ments?

  This sounds great to me, and I think I would be likely to sign up for it if I could, but I haven’t thought about it for more than a few min­utes, am par­tic­u­larly un­sure about the im­pli­ca­tions for cul­ture, and am maybe too en­thu­si­as­tic in gen­eral for things be­ing ‘well or­ga­nized’.

• KatjaGrace 13 Nov 2019 8:11 UTC

  5 points

  in reply to: gwern’s comment on: Pie­ces of time

  Oh yeah, I think I get some­thing similar when my sleep sched­ule gets very out of whack, or for some rea­son when I moved into my new house in Jan­uary, though it went back to nor­mal with time. (Po­ten­tially rele­vant fea­tures there: bed­room didn’t seem very sep­a­rated from com­mon ar­eas, at first was sleep­ing on a pile of yoga mats in­stead of a bed, didn’t get out much.)

Pie­ces of time

Eth­i­cal experimentation

For the metaphors

Self polic­ing for self doubt

Per­sonal qual­ity experimentation

Nor­ma­tive reductionism

To­tal horse takeover

Poli­tics is work and work needs breaks

Skill and leverage

But ex­actly how com­plex and frag­ile?

Do­ing things in a pop­u­lated world

• KatjaGrace 31 Oct 2019 17:08 UTC

  9 points

  in reply to: jacobjacob’s comment on: ja­cob­ja­cob’s Short­form Feed

  I think ran­dom ob­jects might work in a similar way. e.g. if talk­ing in a restau­rant, you grab the ketchup bot­tle and the salt to rep­re­sent your point. I’ve only ex­per­i­mented with this once, with ul­ti­mately quite an elab­o­rate set of condi­ments, table­ware and fries in­volved. It seemed to make things more mem­o­rable and fol­low­able, but I wasn’t much in­clined to do it more for some rea­son. Pos­si­bly at that scale it was a lot of effort be­yond the con­ver­sa­tion.

  Things I see around me some­times get in­volved in my thoughts in a way that seems re­lated. For in­stance, if I’m think­ing about the in­ter­ac­tions of two orgs while I’m near some trees, two of the trees will come to rep­re­sent the two orgs, and my thoughts about how they should in­ter­act will echo ways that the trees are in­ter­act­ing, with­out me in­tend­ing this.

Halloween