Eigenmorality • Hacker News Board

Re: Eigenmorality

#31

Post by mrb » Sat, Jun 21, 2014, 9:58 PM UTC

Wow: "The mathematical verdict of both eigenmoses and eigenjesus is unequivocal: the 98% are almost perfectly good, while the 2% are almost perfectly evil." The author says this diverge violently from most people’s moral intuitions, but actually this result is PRECISELY what moral relativism predicts. See, there are 2 school of thoughts attempting to explain where morality comes from:

- either morality is an absolute concept (things are inherently good or evil, theists might say this good/evil is defined by a god or gods). This is http://en.wikipedia.org/wiki/Moral_absolutism

- or morality is relative, defined by people, defined by cultures (what one culture might consider immoral, another culture will consider it moral, and nobody is inherently right or wrong). This is http://en.wikipedia.org/wiki/Moral_relativism

If moral relativism is right, it would be absolutely expected that the 98% are "almost perfectly good", since they do things that the majority consider good. What a fantastic essay...

Re: Eigenmorality

#32

Post by blauwbilgorgel » Sat, Jun 21, 2014, 10:09 PM UTC

derefr wrote: post #8

As Aaronson points out, PageRank has a few edge-cases when used to do this analysis, basically because it treats its graph as a closed, internally-solipsistic system--it has no definition of morality other than what each of its nodes prefer of one-another. This works if you have a diverse spectrum of preference functions distributed among the nodes (the result tends toward a "live and let live" meta-ethics), but if y…

Agreed. Once a network becomes too homogeneous, its collective problem solving abilities go way down. Effective networks can adapt to and benefit from a large number of different fluid participants. It is a cheap way to get complexity/variety from simple individual nodes, and to be more robust, safe-guarded against variance and overfit.

The author aims to converge upon a group with a good morality (even though its members may have been in the minority). But what is to say that all-good morality is good for a group's decision making as a whole? Would it be sustainable from an energy viewpoint? Or collapse and destroy everything? Won't we need a supply of villains for our moral heros [1]? That to converge on an optimal answer as a network we need wildly opposing views? That to be stable an ecosystem needs variety, decay and destruction?

More philosophically: Is a program like this moral cooperation plan based on pay-it-forward-currency even moral in itself (as it clearly discriminates)?

Attempts at a supermorality have stunned philosophers and logicians for ages. If a rich benefactor would give a 1000 people in a room a 1000$, and if you ask for more you get 1$ more, but 999 people would get 2$ deducted, people would leave that room with some having enough to buy a cup of coffee. 1 million dollars wasted by the greedy individualistic game theory that seems to be in place in animals: I want energy for me and my family first, forget the network. Contests to give a $ amount equal to the lowest unique number send in, would have perfectly rational players roll a dice with the number of contestants, all submit 1 trillion, and the one person who rolls a 1 submits a billion. Instead they receive bids as low as single cents in a manner of days.

[1] http://www.sciencedirect.com/science/article/pii/S0022519311... "The joker effect: Cooperation driven by destructive agents"

Re: Eigenmorality

#33

Post by Eliezer » Sat, Jun 21, 2014, 10:13 PM UTC

mrb wrote: post #31

Wow: "The mathematical verdict of both eigenmoses and eigenjesus is unequivocal: the 98% are almost perfectly good, while the 2% are almost perfectly evil." The author says this diverge violently from most people’s moral intuitions, but actually this result is PRECISELY what moral relativism predicts. See, there are 2 school of thoughts attempting to explain where morality comes from: - either morality is an absolute…

That's... kinda not a very good description of the major contemporary schools of thought on metaethics. I don't know if any respected analytic philosophers take moral relativism seriously as philosophy; you can't ground the foundational meaning of the word 'good' as 'different cultures think different things are good', since there's no base case for the recursion. Well-known mainstream positions in metaethics hold that moral language is not meant to express statements which are either true or false, i.e., it is not semantic or truth-apt; but I have no idea what it would mean for 'good' to be defined as 'different cultures think different things are good'. What's the difference between 'good' and 'fzoom', then?

This appears both well-written and standard: http://cdn.preterhuman.net/texts/thought_and_writing/philoso...

I'd refer you to my own writings on the subject but I don't think they've been very productive in practice of understanding, so I'll leave you with a reference to the standard literature, and remark that the correct analysis (using standard nomenclature, which is somewhat misleading) is obviously moral cognitivism::strong cognitivism::moral realism::naturalist reductionism.

Re: Eigenmorality

#34

Post by gjm11 » Sat, Jun 21, 2014, 10:15 PM UTC

cma wrote: post #20

Needy babies are moral monsters according to many of these models..

Needy babies are also stupid according to many notions of intelligence, ugly according to many notions of beauty, etc. That doesn't mean there's anything wrong with those notions; it means they aren't designed for assessing babies.

Re: Eigenmorality

#35

Post by nabla9 » Sat, Jun 21, 2014, 10:22 PM UTC

jonnathanson wrote: post #30

Please note that what follows can be interpreted as criticism, but it's not intended as such. I found this article quite interesting, and for me, it was the starting point for a lot of different thoughts about game-theoretical approximations of morality. So what follows is a somewhat tangential addition to the article, and not a critique of it. My problem is not with the "eigenmorality" concept, nor with the various…

Prisoner's Dilemma (PD) is good abstraction. The power for game theory comes from abstraction. If you want to understand the implications, you must understand how this abstraction works.

If you can make humans really play PD and they are rational, nothing unexpected happens. The problem with these human experiments is not that PD can't model decision making, it's in the leaking implementation where payoff does not quantify the utility for the players.

Re: Eigenmorality

#36

Post by aruggirello » Sat, Jun 21, 2014, 10:23 PM UTC

MarkPNeyer wrote: post #3

my friends and i had started on this already. i had a hard time explaing to people why it was valuable; looks like scott has done it for us. please help us! https://github.com/neyer/dewDrop right now all we have is a way to state which facebook users a person trusts. there's a chrome extension to help with this. it's extremely basic. i have a server running at https://dewdrop.neyer.me - we need a lot more help! i'm j…

Cool... After reading that, I realised that Eigenmorality is to social networks what Pagerank is to search engines - great to see somebody already working on this!

Re: Eigenmorality

#37

Post by sage_joch » Sat, Jun 21, 2014, 10:26 PM UTC

JackFr wrote: post #18

knowtheory wrote: post #2

This is very long but worth reading. The modeling exercise herein is basically attempting to use a game theoretic model to test out some really dumb/simplified models of cooperation and whether the behaviors observed approximate anything approaching what our intuitions might say is moral behavior, up to and including an 'eigenjesus' and 'eigenmoses' up against tit_for_tat bots and the like.

To each his own. I actually think its overly long, wordy, rambling, poorly conceived and short any novel or important ideas.

He is exploring how the Internet might be used to save civilization. What constitutes an "important idea" in your book?

Re: Eigenmorality

#38

Post by yason » Sat, Jun 21, 2014, 10:31 PM UTC

There is no right or wrong, just acts with unescapable consequences and your freedom to learn something from your choices.

Re: Eigenmorality

#39

Post by blauwbilgorgel » Sat, Jun 21, 2014, 10:31 PM UTC

knowtheory wrote: post #15

rm445 wrote: post #5

Earlier quoted context omitted.

It's very long but written in a light, readable style. What I took away from it was the point late-on about, if we had a PageRank-type system for establishing trust in issues that drive policy (such as climate change), then it would show where a group had left the consensus and formed a separate talking-shop that was trying to shout down the greater consensus. I think this actually applies back to the Iterated Prison…

Yep. My gut reaction to the phrase "eigenmorality" is to immediately turn around and ask "have you ever seen an eigenface?" (see: https://www.google.com/search?q=eigenface&tbm=isch ) They are a certainly a mathematically accurate way to represent a distribution of human faces, but no one would ever confuse it with an actual human face. And this is one way i view political/policy compromises (which is ultimately where…

I found your post interesting, because I think there is always a way for a group to make smarter decisions than any of its individuals. Let's say a middle-ground compromise is chosen on immigration reform: Deport new illegals, offer stricter official ways to become an American. Naturalize people who have worked or studied in the US for a certain time period.

Even if the individuals all agree the collective made a bad decision, that is not to say that the group decision itself was bad: It may have very well been the perfect lesser of all evils. A system can converge to an optimal solution, while individual participants do not realize this. Likewise, a participant may spot a problem in need of a solution, where in reality there is no such problem or manageable solution.

I do agree that, next to making more optimal decisions, a committee can make poor collective decisions that any individual members would never make. If the crowd is too frantically opposed, not willing to give in, then that crowd or system itself may be broken and dysfunctional, and will always produce inferior solutions. It has a bigger problem than a single circumstance.

Re: Eigenmorality

#40

Post by gone35 » Sat, Jun 21, 2014, 10:32 PM UTC

cscurmudgeon wrote: post #25

JackFr wrote: post #18

Earlier quoted context omitted.

To each his own. I actually think its overly long, wordy, rambling, poorly conceived and short any novel or important ideas.

Exactly. Imagine his anger if someone had written something amateurish about quantum computation or complexity without acknowledging any work done since the beginning of these fields. I seriously hope most of the article is in jest. "Hey all, I have a new way of measuring how much resources computer programs take..." If I remember correctly, he banned a professor from commenting on his blog over what seemed to be a r…

If I remember correctly, he banned a professor from commenting on his blog over what seemed to be a routine academic debate.

Not really: for the record, Scott "sentenced" John Sidles repeatedly to a 3-month ban (later "commuted" to two weeks) due to his increasingly derailing and nearly trollish behavior [1], but soon he "dismissed" it for "time served" [2].

[1] http://www.scottaaronson.com/blog/?p=1478#comment-84734

[2] http://blog.computationalcomplexity.org/2013/09/tldr.html?sh...