I don’t think it’s an ea-specific problem, but ea is one of many groups that could improve percolation of coprotection patterns that can provide peer security against unsafe behaviors by getting better at secondhand amogus. I’m personally a fan of …[… copies comment so far into metaphor.systems …]
theanarchistlibrary.org/library/crimethinc-what-is-security-culture ← hmmm this seems worth debating with, but dang, anarchists sure are a certain way; I am very similar to anarchism but a lot more construction-focused than tit-for-tat-destruction focused. as such, I tend to disagree with this article, I would instead argue more for radical transparency, and the problem then is how to make that actually work.
uh this is several kinds of galaxy brain and probably relevant to ai safety in several possible ways, potentially both very positive and very negative depending on details of what universe we’re in: www.grassland.network/
I don’t think it’s an ea-specific problem, but ea is one of many groups that could improve percolation of coprotection patterns that can provide peer security against unsafe behaviors by getting better at secondhand amogus. I’m personally a fan of …[… copies comment so far into metaphor.systems …]
partial results of metaphor.systems query:
instagov.com/wiki/Peer-to-Peer_Accountability_Enforcement ← this is a stub, perhaps could use some of these refs on it?
valence-utilitarianism.com/posts/the-almighty-hive-will← I would actually submit this as a counterexample of what not to doovercomingbias.com/2020/01/nickname-court.html ← this is a solid take I think, but meh
contextualscience.org/heat_honorably_experiencing_anger_and_threat_protocol ← interesting
slatestarcodex.com/2013/05/11/raikoth-symbolic-beads/← nope not thisgordonbrander.com/pattern/soft-security ← interesting
makopool.com/better_space_with_wots.html ← this one is one I already knew to be cool cool but not what I sought
socialpatterns.adl.org/patterns/personal-bubble-setting/ ← huh straightforward description
lesswrong.com/posts/dfszfhtkEDhoQ5nH6/trustworthy-computing ← nnnnmmaybe?
openai.com/blog/amplifying-ai-training/
theanarchistlibrary.org/library/crimethinc-what-is-security-culture ← hmmm this seems worth debating with, but dang, anarchists sure are a certain way; I am very similar to anarchism but a lot more construction-focused than tit-for-tat-destruction focused. as such, I tend to disagree with this article, I would instead argue more for radical transparency, and the problem then is how to make that actually work.
neighborhoodanarchists.org/security-culture/ ← this is pretty much the same article.
localcircles.org/ ← verrry interesting!
intelligence.org/ ← nope heh
squadbox.org/ ← dead link
as followup, I tried only using the end part of the query—similar results, I skipped dupes:
interesting, but not quite what I’m looking for: en.wikipedia.org/wiki/Collective_efficacy
kinda relevant i guess: en.wikipedia.org/wiki/Security_culture
not amazed, but i guess it’s not the worst: nnscommunities.org/strategies/group-violence-intervention-v3/
huh, very interesting take on community inclusion currencies: groupcurrency.org/
huh interesting: medium.com/capabul/grassroots-insurance-8b353a1670f6
cool if it works: betterangels.github.io/buoy/
hmm interesting but not what I was looking for: en.wikipedia.org/wiki/Credible_messenger_program
interesting! very cool! agentofuser.com/groupdag/
oh hi vitalik arxiv.org/abs/1809.06421
verrryyy intersting en.wikipedia.org/wiki/Social_immunity
this is literally the exact topic we were on, in fact! what an interesting result en.wikipedia.org/wiki/Circles_of_Support_and_Accountability
uh this is several kinds of galaxy brain and probably relevant to ai safety in several possible ways, potentially both very positive and very negative depending on details of what universe we’re in: www.grassland.network/
oooh, this is very promising: www.protectivebehaviours.org/
huh, interesting concept to know about en.wikipedia.org/wiki/Complex_contagion
very interesting point, which is that random peer selection allows advantage of human natural coprotection to protect against agentic unfriendly behavior: www.schneier.com/blog/archives/2009/03/the_kindness_of.html
very interesting concept, idk that its quite what I was looking for en.wikipedia.org/wiki/Civil_inattention
cool as hell https://friendlysocieties.org/en/index.html