From: John K Clark (johnkclark@fastmail.fm)
Date: Sun Oct 11 2009 - 10:52:32 MDT
On Sun, 11 Oct 2009 "Mu In Taiwan" <mu.in.taiwan@gmail.com> said:
> we have strange mysterious devices, whose abilities surpass those of turing machines
> already! They're called desktop PCs.
And then he said:
> I just wanted to point out there really is a lot of crap being written on
> this list currently
I agree with you 100%.
John K Clark
and it would be good if it this situation were to
> improve rapidly.
>
> 1. "Fixed minds". What rot. Unless you happen to have a working
> artificial
> mind already built, perhaps it is rather arrogant of you to discuss the
> design limitations of such a thing. Heck, it's not unknown for even
> apparent
> 'physical law' to be overturned occasionally when reality asserts itself
> over theory in the form of empirical data. As for mathematical theorems:
> well, maths is beautiful but it's only a constraint when the axioms of
> your
> mathematics coincide accurately with the axioms that reality has chosen
> for
> itself.
>
> 2. Comments on here about turing machines and their consequences for
> artificial minds. Wow, so much complete crap to wade through. In case you
> haven't noticed, there is not one single Turing machine on this planet.
> Not
> one.
>
> Everything we do today, we do with finite state machines that do passable
> imitations of 'almost' PDAs and 'almost' TMs. Indeed, They have things like clock-driven
> interrupts,
> physical random number generators and state reset buttons, features that
> would make a Turing Machine weep with envy. A turing machine is actually
> just a handy model we use for some theory and programming - when it suits
> us.
>
> Has anyone here actually experienced a desktop PC going into a loop that
> couldn't be halted? I doubt it. Mr Reset Button is your Friend. As for
> general algorithms and coding, use a timer driven interrupt that resets
> the
> state of the entire machine after a fixed time if you want to be
> absolutely,
> positively, certain that any arbitrary routine will certainly halt within
> a
> known time. It might not complete - but it will damn well halt.
>
> 3. Ever noticed a model of time in a turing machine? A model of truly
> non-deterministic behaviour? Well, perhaps making appeals to Turing's
> work
> in your discussion of future artificial minds is a bit stupid, in that
> case,
> and you should stop doing it.
>
> 4. In case you missed it in (2): Artificial minds will not be turing
> machines, because there are no turing machines. They don't exist. If you
> disagree, then please write your response on a piece of infinitely long
> tape
> and post it to this address.
>
> Thank you for listening
>
> Mu
-- John K Clark johnkclark@fastmail.fm -- http://www.fastmail.fm - mmm... Fastmail...
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:04 MDT