HunterJay

• HunterJay 30 Jul 2020 2:12 UTC

  1 point

  on: What a 20-year-lead in mil­i­tary tech might look like

  I re­ally en­joy’d this writeup! I’d prob­a­bly even go a lit­tle bit on the pes­simistic (op­ti­mistic?) side, and bet that al­most all of this tech­nol­ogy would be pos­si­ble with only a few years of de­vel­op­ment from to­day—though I sup­pose it might be 20 if de­vel­op­ment doesn’t start/​ramp up in earnest.

• HunterJay 2 Jan 2020 1:22 UTC

  1 point

  in reply to: Lucas2000’s comment on: Plau­si­ble A.I. Take­off Sce­nario Short Story

  Thanks!

• HunterJay 1 Jan 2020 12:07 UTC

  2 points

  in reply to: Rafael Harth’s comment on: Plau­si­ble A.I. Take­off Sce­nario Short Story

  That’s a good point, I’ll write up a brief ex­pla­na­tion/​dis­claimer and put it in as a foot­note.

• HunterJay 1 Jan 2020 11:29 UTC

  2 points

  in reply to: Rafael Harth’s comment on: Plau­si­ble A.I. Take­off Sce­nario Short Story

  Typo cor­rected, thanks for that.

  I agree, it’s more likely for the first AGI to be­gin on a su­per­com­puter at a well-fund­ing in­sti­tu­tion. If you like, you can imag­ine that this AGI is not the first, but sim­ply the first not effec­tively boxed. Maybe its pro­gram­mer sim­ply im­ple­mented a leaked al­gorithm that was de­vel­oped and pre­vi­ously run by a large pro­ject, but changed the goal and tweaked the safeties.

  In any case, it’s a story, not a pre­dic­tion, and I’d defend it as plau­si­ble in that con­text. Any story has a thou­sand as­sump­tions and events that, in se­quence, re­duce the prob­a­bil­ity to in­finites­i­mal. I’m just try­ing to give a sense of what a take­off could be like when there is a large hard­ware over­hang and no safety—both of which have only a small-ish chance of oc­cur­ring. That in mind, do you have an al­ter­na­tive sug­ges­tion for the ti­tle?

Plau­si­ble A.I. Take­off Sce­nario Short Story