I’m not suggesting all of society is ready for this. I’m suggesting we work toward it among highly rational peers and allies. This is how, and much of why, my close social circles work. Now that I’m used to it, I’d have it no other way.
I wouldn’t say “among rational peers” so much as “among EA-oriented peers”. For our specific community, there is significant overlap in the Venn diagram depicting those two qualities, but those two are very much distinct qualities nonetheless.
A community of HPMOR!Quirrell variations would have your very post in main, with plenty of upvotes, all the while secretly whetting their blades. Perfectly rational.
The more established the trust culture, the more vulnerable it would be to a traitor, a cunning red-pill bastard who plays the trust-network like a fiddle to the tune of his/her egotistical agenda.
Trust—the quintessential element of your so-called “tell culture”—and vulnerability are two sides of the same coin.
When the social circle is small enough as to resemble an expanded family unit, a clan, it may work. A strong sense of ties that bind to keep the commitment to honesty honest would tend to keep a “tell culture”′ social circle’s cardinality well below Dunbar’s number.
Trust—the quintessential element of your so-called “tell culture”—and vulnerability are two sides of the same coin.
That’s true in general. In network security circles, a trusted party is one with the explicit ability to compromise you, and that’s really the operational meaning of the term in any context.
A community of HPMOR!Quirrell variations would have your very post in main, with plenty of upvotes, all the while secretly whetting their blades. Perfectly rational.
I really don’t think so. A community of Briennes, which is not a community of HPMOR!Quirrells but shares some relevant features, would recognize the overwhelming benefit of coordination. Any given individual would be much stronger if she had the knowledge of all the other individuals, or if she could count on them as external memory. And because she would be stronger that way, she knows that they would be stronger if she also remains trustworthy. Her being trustworthy allows her to derive greater benefit from the rest of the community. Other people are useful, you see. With Tell culture in place, you can do things like feed your model of the world into someone else’s truth-checker and get back a more info-rich version. You only defect if the expected utility of doing so outweighs the expected utility of the entire community to your future plans.
I’d love to hear what culture Eliezer thinks an entire community of Quirrells would create.
If they all started off in a symmetrical position, they’d use Unbreakable Vows to keep from killing each other and then proceed to further affairs, not necessarily cooperatively.
Wouldn’t this require one Quirrell to agree to sacrifice a part of his power before any other Quirrell does? (Assuming that all of the vow rituals taking place at the same time would require each Quirrell to take part in more than one ritual simultaneously, which doesn’t seem possible.) It seems to me that a Quirrell wouldn’t agree to this.
You don’t have to sacrifice your own power for that, the bonder sacrifices power. And the Unbreakable Vow could be worded to only come into force once all Vows were taken.
But, in this case, the bonder is another Quirrell picked from this all-Quirrel community, right?
Of course, if we allow the ritual to depend on the completion of other rituals, then the problem is moot.
It strikes me that this conversation really hinges on just how evil HPMOR’s Quirrell turns out to be, which is problematic since you know a few chapters more plot than I do...
(Also, since I find myself having a conversation with you, might I say that I very much like HPMOR, and that I would like it even more if you were to amend chapter 19 so that Quirrell didn’t perpetuate one or two myths about martial arts, a subject on which I focus a certain amount of my own nerdly attentions? I posted a review under “James”, but the short version is that (1) martial arts aren’t particularly Asian, and (2) “I’m a sixth dan” means no more than “I once got a B- in a class whose subject I won’t divulge except to say that it was ‘Math’.”)
The great-grandparent comment did make me consider unbreakable vows as a theory of what happened on Halloween. E.g. to prevent one of his Horcruxes from later killing him, Voldemort made an unbreakable vow not to magically interact with his alter egos (this causing Harry’s sense of Doom around Quirrell). Doesn’t seem necessary, though.
A community of Briennes, [...], would recognize the overwhelming benefit of coordination.
But it would pay the price Tell comes with. And the Briennes wouldn’t need it because they know all their rules and could easily use the more efficient Guess.
“You only defect if the expected utility of doing so outweighs the expected utility of the entire community to your future plans.” These aren’t the two options available, though: you’d take into account the risk of other people defecting and thus reducing the expected utility of the entire community by an appreciable amount. Your argument only works if you can trust everyone else not to defect, too—in a homogenous community of Briennes, for instance. In a heterogenous community, whatever spooky coordination your clones would use won’t work, and cooperation is a much less desirable option.
I’m not suggesting all of society is ready for this. I’m suggesting we work toward it among highly rational peers and allies. This is how, and much of why, my close social circles work. Now that I’m used to it, I’d have it no other way.
Tricky (like most anything).
I wouldn’t say “among rational peers” so much as “among EA-oriented peers”. For our specific community, there is significant overlap in the Venn diagram depicting those two qualities, but those two are very much distinct qualities nonetheless.
A community of HPMOR!Quirrell variations would have your very post in main, with plenty of upvotes, all the while secretly whetting their blades. Perfectly rational.
The more established the trust culture, the more vulnerable it would be to a traitor, a cunning red-pill bastard who plays the trust-network like a fiddle to the tune of his/her egotistical agenda.
Trust—the quintessential element of your so-called “tell culture”—and vulnerability are two sides of the same coin.
When the social circle is small enough as to resemble an expanded family unit, a clan, it may work. A strong sense of ties that bind to keep the commitment to honesty honest would tend to keep a “tell culture”′ social circle’s cardinality well below Dunbar’s number.
That’s true in general. In network security circles, a trusted party is one with the explicit ability to compromise you, and that’s really the operational meaning of the term in any context.
I really don’t think so. A community of Briennes, which is not a community of HPMOR!Quirrells but shares some relevant features, would recognize the overwhelming benefit of coordination. Any given individual would be much stronger if she had the knowledge of all the other individuals, or if she could count on them as external memory. And because she would be stronger that way, she knows that they would be stronger if she also remains trustworthy. Her being trustworthy allows her to derive greater benefit from the rest of the community. Other people are useful, you see. With Tell culture in place, you can do things like feed your model of the world into someone else’s truth-checker and get back a more info-rich version. You only defect if the expected utility of doing so outweighs the expected utility of the entire community to your future plans.
I’d love to hear what culture Eliezer thinks an entire community of Quirrells would create.
If they all started off in a symmetrical position, they’d use Unbreakable Vows to keep from killing each other and then proceed to further affairs, not necessarily cooperatively.
Wouldn’t this require one Quirrell to agree to sacrifice a part of his power before any other Quirrell does? (Assuming that all of the vow rituals taking place at the same time would require each Quirrell to take part in more than one ritual simultaneously, which doesn’t seem possible.) It seems to me that a Quirrell wouldn’t agree to this.
You don’t have to sacrifice your own power for that, the bonder sacrifices power. And the Unbreakable Vow could be worded to only come into force once all Vows were taken.
But, in this case, the bonder is another Quirrell picked from this all-Quirrel community, right?
Of course, if we allow the ritual to depend on the completion of other rituals, then the problem is moot.
It strikes me that this conversation really hinges on just how evil HPMOR’s Quirrell turns out to be, which is problematic since you know a few chapters more plot than I do...
(Also, since I find myself having a conversation with you, might I say that I very much like HPMOR, and that I would like it even more if you were to amend chapter 19 so that Quirrell didn’t perpetuate one or two myths about martial arts, a subject on which I focus a certain amount of my own nerdly attentions? I posted a review under “James”, but the short version is that (1) martial arts aren’t particularly Asian, and (2) “I’m a sixth dan” means no more than “I once got a B- in a class whose subject I won’t divulge except to say that it was ‘Math’.”)
The great-grandparent comment did make me consider unbreakable vows as a theory of what happened on Halloween. E.g. to prevent one of his Horcruxes from later killing him, Voldemort made an unbreakable vow not to magically interact with his alter egos (this causing Harry’s sense of Doom around Quirrell). Doesn’t seem necessary, though.
But it would pay the price Tell comes with. And the Briennes wouldn’t need it because they know all their rules and could easily use the more efficient Guess.
“You only defect if the expected utility of doing so outweighs the expected utility of the entire community to your future plans.” These aren’t the two options available, though: you’d take into account the risk of other people defecting and thus reducing the expected utility of the entire community by an appreciable amount. Your argument only works if you can trust everyone else not to defect, too—in a homogenous community of Briennes, for instance. In a heterogenous community, whatever spooky coordination your clones would use won’t work, and cooperation is a much less desirable option.