<html>
<body>
<font face="arial" size=2>Good morning folks!<br><br>
One reason I like Greg's books is that he -- like other good fictions
writers -- <br>
"puts it together." There are lots of verbal theories espoused
in social<br>
science, especially, which, when translated into a picture... don't
translate.<br>
And when ideas are translated into a picture... it tells us something we
wouldn't have seen<br>
otherwise. That's a very important part of .. advancing our
thought.<br><br>
But, as with other compelling pieces of fiction, like Ayn Rand or most of
the Bible,<br>
we then may move on to ask which parts were real and essential, which
were interesting but uncertain,<br>
and which were just props. We don't always know.<br><br>
The core message that most of our DNA may be involved in a kind of
"intelligence" function,<br>
a kind of metagenetic system, is important ... and I think it is
inescapable.<br><br>
My views here are influenced heavily by what I have seen with
intelligence in neural network systems where,<br>
for good or ill, I would claim to have tangible scientific knowledge a
bit more than others.<br>
(I could elaborate but not this morning.)<br><br>
It is clear that an effective learning system needs to exercise more
resources to the learning itself than it does<br>
to the "crystallized" stimulus-memory-response patterns which
it learns. That's a general<br>
sort of reality, and it would apply to genetic systems as well. Certainly
there are times when slow learning or adaptation<br>
gets you killed. Thus there is a strong evolutionary advantage to having
systems which can adapt better.<br>
Also... on the whole... life today seems to evolve faster than the life
of the preCambrian era.<br>
It does indeed seem to have evolved an evolutionary capability. And
then the "junk DNA"<br>
fraction provides strong confirmation of what we should have expected.
<br><br>
Good learning abilities do not come cheap. And that's something I could
elaborate on later.<br><br>
But...<br><br>
What about stockpiling? That's a different question.<br><br>
I would think of it more as maintaining a kind of sphere of variance...
analogous to what people would<br>
do with... extended Kalman filtering or particle filtering or a system
called SEDP/TLRN (which is like particle filtering but more<br>
efficient)... well....<br><br>
There is a problem with stockpiling PIECES of a new state , when the
COMBINATION of pieces has never been tested together.<br><br>
Greg's novels do discuss some of the problems that can occur, actually,
in settling down when the rate of deployment of<br>
stockpiled variations rises very suddenly. There is a whole range of
plausible systems, ranging from<br>
continuous deployment and testing of new ... phenotypes?.... to what Greg
portrays, ultradiscontinuous...<br><br>
<br>
But... I am not an evolutionary theorist, and the clock just came
online...<br><br>
Best,<br><br>
Paul</font></body>
</html>