Comments on: The failure mode of abundance http://trilema.com/2017/the-failure-mode-of-abundance/ Moving targets for a fast crowd. Thu, 14 May 2026 00:40:52 +0000 http://polimedia.us hourly 1 By: The Bitcoin learning tournament. on Trilema - A blog by Mircea Popescu. http://trilema.com/2017/the-failure-mode-of-abundance/#comment-143337 The Bitcoin learning tournament. on Trilema - A blog by Mircea Popescu. Wed, 08 Jan 2020 10:19:42 +0000 http://trilema.com/?p=73308#comment-143337 [...] by discussion on the previous article, it's my great pleasure to give you the world-famous Bitcoin Learning [...] [...] by discussion on the previous article, it's my great pleasure to give you the world-famous Bitcoin Learning [...]

]]> By: Stanislav Datskovskiy http://trilema.com/2017/the-failure-mode-of-abundance/#comment-122198 Stanislav Datskovskiy Wed, 07 Jun 2017 17:54:17 +0000 http://trilema.com/?p=73308#comment-122198 I don't have a Eulora-sized hole in my calendar, sadly. I don't have a Eulora-sized hole in my calendar, sadly.

]]> By: Mircea Popescu http://trilema.com/2017/the-failure-mode-of-abundance/#comment-122197 Mircea Popescu Wed, 07 Jun 2017 17:22:29 +0000 http://trilema.com/?p=73308#comment-122197 Challenge accepted. Because somehow the full blown item that's Eulora isn't good, gotta make schematics on napkins as part of life in the Petri dish or somesuch. Challenge accepted. Because somehow the full blown item that's Eulora isn't good, gotta make schematics on napkins as part of life in the Petri dish or somesuch.

]]> By: Stanislav Datskovskiy http://trilema.com/2017/the-failure-mode-of-abundance/#comment-122196 Stanislav Datskovskiy Wed, 07 Jun 2017 17:20:15 +0000 http://trilema.com/?p=73308#comment-122196 Chess board doesn't much resemble battlefield from actual life either, sadly. And -- I don't doubt that you could conceive of a more interesting tourney than the EU ( yes, the subj was an official project thereof! believe or not ). I'd play. Chess board doesn't much resemble battlefield from actual life either, sadly.

And -- I don't doubt that you could conceive of a more interesting tourney than the EU ( yes, the subj was an official project thereof! believe or not ). I'd play.

]]> By: Mircea Popescu http://trilema.com/2017/the-failure-mode-of-abundance/#comment-122195 Mircea Popescu Wed, 07 Jun 2017 17:16:56 +0000 http://trilema.com/?p=73308#comment-122195 The problem I see right off is the specification of the environment. <blockquote>The simulated environment for our tournament was a ‘multi-armed bandit’ (18), analogous to the ‘one-armed bandit’ slot machine but with multiple ‘arms’. In the tournament, the bandit had 100 arms, each representing a different behavior, and each with a distinct payoff drawn independently from an exponential distribution.</blockquote> The bias in favour of "social learning" bullshit is screamingly obvious, not that it's surprising from someone who "publishes" his research online. Here's an idea : suppose the number of arms were a) not known in advance and <b>most importantly</b> they were not always presented in the same god damned slots to all comers. Yes n-armed bandit may have arm 2, but it won't appear in slot fucking #2, conveniently labeled so as to save ~100% of the fucking cost of social "learning", which is to say global namespace. The "tournament" as presented is no more than an exercise in taking a winding (and perhaps pictoresque, especially if you're new at this) path from the author's pet "truths" back to the author's own pet "truths". The problem I see right off is the specification of the environment.

The simulated environment for our tournament was a ‘multi-armed bandit’ (18), analogous to the ‘one-armed bandit’ slot machine but with multiple ‘arms’. In the tournament, the bandit had 100 arms, each representing a different behavior, and each with a distinct payoff drawn independently from an exponential distribution.

The bias in favour of "social learning" bullshit is screamingly obvious, not that it's surprising from someone who "publishes" his research online. Here's an idea : suppose the number of arms were a) not known in advance and most importantly they were not always presented in the same god damned slots to all comers. Yes n-armed bandit may have arm 2, but it won't appear in slot fucking #2, conveniently labeled so as to save ~100% of the fucking cost of social "learning", which is to say global namespace.

The "tournament" as presented is no more than an exercise in taking a winding (and perhaps pictoresque, especially if you're new at this) path from the author's pet "truths" back to the author's own pet "truths".

]]>