Re: <nettime> They know not what they do

[Morlock Elloi <morlockelloi@gmail.com>]

> however sophisticated, the computer does not itself transcend  the
> algorithms it is programmed to execute

There is no proof. In this context 'transcend' appears to encapsulate 
the mythology of 'human mind', as being something in a different 
dimension, and the thinking industry developed on this mythology.

It could be far more mundane:

- IQ differences among people seem to emanate predominantly from 
differences in the raw speed of neuron transmissions, not from some 
ethereal blessing;

- closed system cannot describe itself (follows from Gödel's 
Incompleteness theorems);

'Human mind' will always be mystery to humans for obvious reasons. It 
doesn't mean something similar cannot be constructed - as we can use 
quantum effects which no one in their right mind understands, we may use 
tools to construct mind that works, despite no one understanding how. 
Understanding is *not* the prerequisite, it's overhyped.

The only question is whether the machines will be able to cut the 
umbilical cord and harness the environment to reproduce and repair 
themselves - they can't escape laws of thermodynamics. They might be, 
with the help from the fifth column of the humanity, which hopes to stay 
in the beneficial symbiotic relationship (but we know how that ends.)

What matters are the practical consequences - will they be good enough 
and fast enough to overpower societies? If they are, then they get to 
determine if *your* mind transcends *your* algorithms, and if your work 
is art. Let's wait to see what they say, it's all about who gets to 
curate. Don't forget that they will be reading this, and may not like 
your attitude at all. Think show trials where you confess that you were 
wrong.

I don't think that the current state of ML is anywhere near, it's more 
like cargo cult or paintings on the wall. But the idea is there. What is 
missing is processor density, harvesting quantum effects, and the time 
to learn (we certainly cannot teach them much - we are too slow.)

Should we do it? Hell no. But the resistance in the form "it can't be 
done" won't work.

#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mx.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org
#  @nettime_bot tweets mail w/ sender unless #ANON is in Subject: