Greg Detre
Wednesday, April 09, 2003
claims that
Lisp is powerful (partly) because it�s so easy to define new data structures
rumour that
Lenat faked his results � people talked of micro-Lenats as a measure of how
much you should doubt someone
AM � found
numbers with only three divisors � the squares of primes
he lost the backups and was unable to show people the program running
Ken Hoss wrote a program based on Lenat�s
description, which did some of the things that Lenat had claimed � though it
didn�t find prime numbers
eventually, it too stopped discovering interesting things
someone
suggested that the Critic/Selector model is too brutal
the
Crit/Sel model doesn�t take into account the WtT you�re in now, does it???
he found
out that when you have longer than 1-page ideas, you have interactions between
ideas
�I can
resolve all the objections so far by weasel�
see SoM discussion of game
�W. does this by describing �language games�: the practices, activities, actions and reactions in characteristic contexts in which the rule-governed use of a word is integrated.� � method for disentangling conceptual confusions�. � quote from Sarah W notes
simpler example of blindsight � you can see that�s a chair, but you can�t say why you know it�s a chair J
IBM subcontracted the original PC-building
ban international travel, until we know more about genetics and viruses
$100b to build a pipeline to suck the oil out of Iraq
send
hydrogen
article
ethics
ontology
synopticon
the mutation rate of a virus is proportional to how much of it there is floating around � so ebola isn�t a threat if it�s killing off a village every so often, it won�t mutate and become dangerous � if it gets to a city, it�ll mutate really fast and probably wipe us all out
religions as persistent self-consistent meme institutions
religion would have affected the course of evolution
me: that�s why we�re so credulous J
nancy wants to know what goes in the brain during forgiveness
pah � presumably it�s evolved, or simply it�s because forgiving someone is easier than fighting them � vengeance is only worthwhile as a deterrent
robert wood � n-rays???
if you made a very smart machine, surely the first 1000 would go insane
do you see
a role for statistical methods at all???
is there a
qualitative difference between evolutionary and lifetime learning???
how does it
generate new Critics???
do you ever worry about the ethical implications of AI research???
ethics is just the business of prohibiting important things
what about legal definitions of personhood???