[ExI] How not to make a thought experiment

[John Clark]

Since my last post Gordon Swobe has posted 7 times.

> I have yet to meet a conscious computer

One can't help but wonder how Swobe knows this, and how he knows if he has ever met a conscious human being.

> some people believe, naively, that if we can compute x then a computation of x = x.

So if we can compute that 2+2=4 then it's naive to think that a computation of 2+2 is 4. Huh?
> 
> the event itself does not equal a computation.

That depends entirely on what the "event" is; if it involves moving atoms then Swobe is right, but if it involves things other than nouns or in making decisions then he is not.
> 
> Anybody who says "A digital simulation of a brain will equal a brain" says exactly that a computation of x = x where x = a brain.

Swobe trots out this sad old straw man yet again! Of course a simulated brain is not identical with a biological brain, but a simulated mind is, just as a simulated 4 (whatever the hell that means) is identical with a real 4. 

> Except in the special case in which x = a digital program or computer, the digital simulation of x will not even look like the originals

Except for that Mrs. Lincoln, how did you like the play?

> the man in the  [Chinese] room has no access to sense data from the outside world

Yet another reason why the Chinese room just may be the most useless thought experiment ever devised; a real AI would certainly have sense data about the outside world, far more the we do in fact.

> when we look at how computers get sense data, we see that sense data also amounts to nothing more than meaningless patterns of 1's and 0's.

One can't help but wonder why Swobe doesn't erase the hard drive on his computer, it is after all full of  nothing but a meaningless pattern of 1's and 0's. It's also puzzling how Swobe makes a living, he says he does it by generating patterns of 1's and 0's, but who would pay him for such a meaningless activity?

> go find your magnifying glass and take closer look at that apple on your monitor. Compare it to what you see when you hold the glass to an actual apple. That's not an apple on your screen after all, now is it?

Swobe may be the very first person to make such a brilliant observation!

> Computationally speaking, on this view, you can make a "brain" that functions just like yours and mine out of cats and mice and cheese or levers or water pipes or pigeons or anything else [e.g., beer cans and toilet paper] provided the two systems are, in Block's sense, "computationally equivalent" .

True but trivially obvious.

> You would just need an awful lot of cats, or pigeons or waterpipes, or whatever it might be. 

And if you wanted to make a brain out of neurons you'd need an awful lot of them too.

> the man can learn only that 'squoogles' follow after 'squiggles'. 

Wow, now I see the error of my ways! It's a pity Swobe didn't say that two months and several hundred posts ago, think of the time we could have saved. Oh wait he did.

 John K Clark


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20100221/2399e4a9/attachment.html>