From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Fri May 03 2002 - 13:51:52 MDT
For me, one of the highlights of the Foresight Gathering was Christine
Peterson (President of Foresight) reminding everyone at the Foresight
Gathering of some important rules for how to avoid forming a clique. And I
listened to her and thought to myself: "Chris Peterson had to put effort
into doing this. It didn't just happen." Any group of people who have
something in common, and who have their own mailing list, will naturally
tend to form a clique unless specific steps are taken to avoid it. It's not
something that happens when we do something wrong; it's something that
happens by default.
The only really long-term solution that I know of is studying up on
evolutionary psychology and learning to see yourself as a modified primate,
after which you start to recognize the emotions relating to tribalism and
group polarization dynamics. Meanwhile, because the SL4 list isn't my own
private piece of territory, I expect there will always be people here who
disagree with me about evolutionary psychology, newcomers who haven't had
the chance or don't have the time to read up on ev-psych, and so on. So
here's a list of some of the ways I recognize groupishness.
1) Individual actions are taken as being representative of the group. This
hasn't happened on SL4 yet, but I've seen it happen on two occasions very
recently - once between the WTA and Extropy, and once between the Extropians
mailing list and the Church of Virus mailing list. Or to be more precise,
once between James Hughes (WTA) and several Extropians, and once between
Eugene Leitl and Hermit (CoV). Usually what happens is that someone from
Group A, speaking as an individual, says something bad about Group B, and
Group B takes this as an assault from Group A - or at least one individual
from Group B takes this as an assault from Group A. The really nasty part
of this situation is that often Group A, instead of issuing a disclaimer and
trying to sort the issue out, reacts in kind to Group B - or rather, several
people from Group A react in kind to people from Group B. This is how an
individual disagreement can escalate into a tribal war. Very nasty. In
both cases the issue was successfully sorted out, but only because several
people, including me, were there thinking: "Argh! Group polarization
dynamics!" and pounding the sand trying to get things sorted out, instead of
paying attention to the putative topic of dispute.
2) Writing for other group members is another symptom of cliqueishness -
this is another of the things Christine Peterson specifically highlighted.
Writing for people who are already convinced accomplishes nothing - why
would anyone even try to do it? Because it's easier and natural; it happens
by default unless someone tries to prevent it. The only acceptable proposal
I've ever heard for "writing for SL4 members" was a putative project to
summarize some of the major threads in the archives. Aside from that, there
are two kinds of writing; there's FAQs and other introductory literature
targeted at a mass audience, where assuming scientific literacy is the most
you can get away with; and there are very detailed explorations of a
specific topic where the intent is to say something new about the issue.
What could possibly be the point of writing introductory material for an SL4
audience?
3) Every now someone on SL4 accuses us of groupthink or accuses me of being
the local guru. Usually this person is also incapable of correct spelling
or structured thinking and gets kicked out for that reason; the ones who can
write a decent post stay and usually learn better after a while. Of course
this doesn't just happen on SL4; I recently witnessed a very comical
incident wherein someone in a chat channel who was spamming the channel
insisted that he was being banned from chats because chat ops were all
power-mad dictators. However, just because a lot of annoying people launch
throwaway accusations of this class doesn't mean that we can get careless to
the point of letting it become true. If you find yourself looking at a bad
post and thinking "Ew, how un-SL4," then that's probably groupthink.
Remember that there are plenty of non-SL4s out there who are scientifically
literate, extremely versed in the SL3 or whatever futurism, who are
excellent writers who could dismember many SL4 posters in the blink of an
eye, and who will be completely welcome on this mailing list if they ever
choose to show up. There are people who plain don't believe in the
Singularity who subscribe to SL4 and make perfectly good posts. So if
you're looking at a bad post and thinking "Ew, how un-SL4," then that's
groupthink. The real problem probably isn't that the person's
technological-futurism assumptions are different than yours, but that they
made some other kind of mistake entirely. Attributing a vague feeling of
"That's wrong" or "That makes no sense" to the poster being "un-SL4" is
groupishness. Attacking the poster for being un-SL4 is definitely
groupishness.
Will Pearson's problem is not that he is "un-SL4" or that he thinks
self-modifying goal systems inevitably drift - it's perfectly acceptable for
someone to show up and say, "You know, I'm going to have to disagree with
the Friendliness advocates about this one; I think that self-modifying goal
systems are inevitably going to drift." It's saying: "I have this
brilliant idea which you people have clearly never thought of:
Self-modifying goal systems are going to drift," that marks the speaker as
someone who needs to read up a bit more in order to learn, not so much the
specific technical counter to this argument, but rather that the SL4 list is
of reasonable collective intelligence and may have thought of this objection
already. "People who've spent their lives as the smartest person in the
room" syndrome.
Mike-and-Donna (whichever of them they are) are perfectly welcome to
disagree about whether morality requires post-Singularity revenge on
pre-Singularity criminals. You are welcome to be horrified by this (I am),
but in what specific way is it un-SL4? Attributing this to un-SL4-ishness
seems to me like groupthink.
Max Comess's previous crime was a series of one-line posts followed by the
spamming of people I'd like to talk to someday. But since then he raised
the quality of his posts and I have no objection to the recent ones. Why is
proposing a planetary soup of self-modifying heuristics stupid? Back in
1999/2000 that was my actual proposed strategy for implementing the
Singularity!
Well, that's pretty much what I wanted to say. Hopefully anyone subscribed
to the mailing list or reading the archives will remember that being the
moderator is not the same as being the tribal chief, and that if other SL4
list members feel that groupishness is the best way to bond into a group
capable of achieving its goals, there isn't anything I can do about it
except try to persuade them otherwise.
-- -- -- -- --
Eliezer S. Yudkowsky http://intelligence.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:38 MDT