From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Thu Aug 31 2006 - 00:02:41 MDT
Here's my view of recent events, down for the record.
At the end of August, i.e., at the end of tomorrow, this topic is
officially dead and killthreaded, so if you've got something you
absolutely need to get off your chest, get it off your chest before then.
May the Wise kindly take note:
Ben Goertzel is a nice fellow - so nice that he found nice things to say
about Mentifex, for God's sake. Saying of Mentifex: "In my view,
although Arthur sometimes *presents* his ideas in a somewhat kooky
way... the ideas themselves are significantly better than most of what
passes for cognitive science and AI." Goertzel later mostly retracted
this, then rephrased it and said it again (see the Mentifex FAQ). The
point is: if Ben wasn't willing to call Mentifex a crackpot, then I'm
sorry, Ben, but I'm going to take any nice things you say about
Loosemore with a rather large grain of salt.
At first I too thought that Loosemore was a polymath. However, there is
such a thing as updating past opinions in response to new evidence.
There are such things as crackpots, in this world. If you are not
unfortunate enough to be a very minor public figure, you probably do not
get emails from lunatics who want to talk to you about their perpetual
motion machine. I get those emails. I don't reply to them, but still -
bizarre as it may seem - even I am enough of a celebrity that I have to
deal with crackpots. How Ray Kurzweil ever gets to his legitimate
email, I don't know; I expect he has someone sort it for him.
Most crackpots cannot spell. A few crackpots are articulate and can be
mistaken for technical people. Since there are more technical people
than articulate crackpots, the prior probability is low, but it is not
zero, and one runs into them from time to time.
Crackpots share certain characteristics in common, which is why John
Baez's Crackpot Index is funny. It's more funny if you've had to
actually deal with crackpots. I don't get physics crackpots so much,
though I do get some; but that list is very, very true.
Is Loosemore a crackpot?
At first I thought Loosemore was a polymath, familiar with many fields,
as he claimed (it is not so implausible).
Then I thought Loosemore was a polymath who had been seduced by the
sucking black hole of Emergence Mysticism in the guise of Complex
Systems Theory. This is sad, but not a crime, and who knew but when
some subject would come up that might benefit from his expertise?
Then Loosemore declared forcefully that falsifiable predictions were not
the issue, and that any discussion must be conducted on the paradigm
level. And this was flagrantly stupid, but still not a crime, and still
not something that would prevent Loosemore from contributing to a
technical discussion that infringed on one of his polymath subjects. At
this time I still thought he was a polymath - rather unskilled in the
Art, but still a polymath who had just fallen into some very unfortunate
memetic attractors. I thought of the 50-point item on Baez's list, but
I did not think "crackpot".
So when did I undergo the figure-ground inversion? When did the vase
become two faces, the young lady become the old crone?
When Loosemore started to claim that he was already an expert
specifically in heuristics and biases, which I knew damn well he was
not, and which anyone in the audience can verify by reading up on the
subject; also you have Jeff Medina's word for it, and even Ben Goertzel
agrees on this.
If Loosemore would exaggerate his expertise in one field, why not another?
Suddenly his open rejection of falsifiable predictions took on a whole
different light.
And his evasiveness, and frequent changes of subject.
And the fact that he had never actually demonstrated any of the
technical knowledge he claimed.
And his swiftness to take insult when questioned.
If I needed any possible further confirmation, I had it when Loosemore
tried to pull rank and get everyone to stop questioning him - by
invoking, for God's sake, a Master's degree in psychology. *That* was
classic crackpot. (I am still willing to believe that Loosemore has a
Master's degree in psychology, though I wouldn't bet money on it. The
tale about the psychology department inventing a special distinction for
just his degree, seems rather less probable.)
And then, Loosemore told an outright untruth. He said that he had
known, at the time he wrote his first take on the Conjunction Fallacy,
of the research that refuted his suggestions.
When I say that the eye of the Wise is blinded to the gapingly obvious,
I am referring, specifically, to anyone who can look at Loosemore's
original post on the Conjunction Fallacy; read the paper I referenced by
Osherson; read Loosemore's later claim to have known about all that
research already and had it in mind at the time of writing his reply;
and still believe that Loosemore is telling the truth.
Now, of course, I could be wrong about all this; that is the hazard of
being mortal. Even more likely is that I am wrong about at least one
aspect of it; that is the hazard of telling a detailed tale in which
each detail is another chance to be wrong.
If I were more reluctant to change my beliefs; or if, desiring to appear
Wise, I felt more obligated to give people the benefit of the doubt and
believe them virtuous, even when a pure rationalist might deem it
improbable; then I might have invented a different tale. Then I might
have made yet another incremental modification to my beliefs about
Loosemore. So that I could keep to my old belief, rather than
abandoning it entirely, and so never admit that I had been fooled from
the beginning...
In this tale, Loosemore is a true and knowledgeable polymath...
...who got stuck in the black hole of emergence mysticism...
...who made the severe (but honest) mistake of believing that
falsifiable predictions were not important...
...who sometimes changes the subject a little, but so do we all...
...who, steeped in the halls of academia, is accustomed to a high degree
of courtesy from others, and swift to point out any departure from that
standard...
...who didn't recall any of the research on the Conjunction Fallacy when
he first encountered the example, but who did genuinely read Judgment
Under Uncertainty as part of a psychology course in 1987, and just
didn't remember it...
...who, later, after a quick online refresher course in heuristics and
biases, made an honest error of recollection as to how much he had
originally known, and what he had had originally in mind; for those who
study such matters will know that memories are oft all too pliable,
changing each time we recall them to mind; and perhaps his mismemory was
exaggerated a bit more by the heat of discussion, his desire to prove
himself wholly right from the beginning...
...who, when angry, made intemperate remarks about his degree, but this
was only in the heat of anger...
Yes, I can see this alternative hypothesis. Of course I see it. Even
if I did not invent it for myself, I would have seen the elements of
this picture in others' minds.
Yes, I assign this alternative hypothesis a nonzero probability - though
it is too detailed to be wholly true in all its parts. It is more
likely that Loosemore is a genuine polymath for whatever reason, than
that the entire explanation is exactly true. Perhaps the truth is some
third way entirely. In my experience so far, when it seems hard to
decide which of two hypotheses is true, there often is a third
hypothesis that is much simpler which I am just overlooking.
But in the end I must judge and act.
My guess was that Loosemore was a crackpot, or, even if not a crackpot,
someone who was rather unlikely to contribute usefully to SL4. I acted
accordingly.
-- 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 - 04:00:57 MDT