Re: [sl4] Alan Turing's results are profound

From: Robin Lee Powell (rlpowell@digitalkingdom.org)
Date: Thu Oct 15 2009 - 11:10:43 MDT


On Thu, Oct 15, 2009 at 09:57:14AM -0700, John K Clark wrote:
> > A brain doesn't work on a fixed set of inputs
>
> And neither does a Turing Machine.

0.o

Go read a book on formal computation. *PLEASE*? This is, in fact,
a very important part of the Turing machine formalism.

A Turing machine, on each run, takes a single input, at the
beginning, and that's all. This is part of where the halting
problem comes from. This is what is meant by "fixed set of inputs".
There's no such thing as "interaction" with a Turing machine; you
give it a program, and an input, and you run it; you never interact
with it again until it halts. That's the whole point of the Turing
machine formalism.

Humans, and computers in the real world, routinely receive new and
interesting inputs from the outside world, whether they want to or
not. This rather handily makes the halting problem essentially
irrelevant: given new input to respond to, an infinite loop can be
avoided, unless it's so bad that it leads to ignoring all input
entirely, in which case someone was a pretty crappy programmer, and
that's why we have reset switches (which, again, are input from the
outside, nicely circumventing the halting problem). For humans, we
have things like adrenalin and antipsychotics and shock therapy,
although to be fair those don't always work. But then, the reset
button doesn't always work.

Anyways, it's only a real problem if every thinking being on earth
halts at the same time. I'll lay my odds against that one, thanks.

-Robin

-- 
They say:  "The first AIs will be built by the military as weapons."
And I'm  thinking:  "Does it even occur to you to try for something
other  than  the default  outcome?"  See http://shrunklink.com/cdiz
http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:05 MDT