[ExI] Public draft of my book 'Tales of the Turing Church

John Clark johnkclark at gmail.com
Thu Oct 18 18:11:16 UTC 2018


This is my second try at posting this, the first time bounced.

On Tue, Oct 16, 2018 at 10:06 PM Stuart LaForge <avant at sollegro.com> wrote:

>>  Quantum Mechanics can't be the ultimate scientific model of reality
>> because it says nothing about gravity, and we know nothing about Dark
>> Energy and Dark Matter except that its 95% of reality.
>
>
> > *MWI also has the advantage of being compatible with general relativity
> in a way that Copenhagen and its ilk cannot be: no FTL wave function
> collapse*


Copenhagen has many faults but I don't think FTL wave function collapse
violates relativity because it can't be used to transmit a  signal faster
than light.

>*That was the part of Everett's interpretation that I disliked the most
> until I realized that the math works out the same if universe doesn't split
> because all the possible universes are all already out there *


That is only from the viewpoint of somebody outside the multiverse looking
in at it, and that viewpoint does not exist.

>
> *If all possible causal cells (Everett branches) exist on the same
> infinite n-dimensional manifold, then the overall state of universe itself
> does not change*


Does not change in what dimension? If the spacetime in the multiverse did
not change in any direction the Multiverse could only be an eternal
unbounded infinitely large homogeneous lattice, and anything that simple
and dull could never have something as complex as life in it or anything
interesting.

>
> *Nothing splits because there are infinite versions of every
> observer observing every possible outcome except for the ones for which
> they are not present. Nothing changes when you observe a quantum state
> except your knowledge. You are just narrowing down which causal cell you
> have beenresiding in all this time.*
>

When speaking about the multiverse and Many Worlds the use of personal
pronouns like "you" and "your" can easily become ambiguous.

>>  I don't see how  locality could be wrong. If things were non-local a
>> change anywhere would instantly change everything everywhere and before
>> you could understand anything you'd have to understand everything. We
>> certainly don't know everything but we do know a few things and I don't
>> see how we could if things were non-local.
>
>
> > *String theorists propose as many as 14 dimensions. All it takes is for
> one of those extra dimensions to actually exist and opposite sides of
> the observable universe could be an inch apart through that extra
> dimension.*


But then it's hard to understand why classical physics can do such a good
job at predicting the tides when it only takes into account the moon and
the sun, it ignores all the movements of planets a billion light years away
and assumes that things are local. And classical physics gets away with it
and I don't see how it could if things were not local.


>
>
> *> Something like this must happen since opposite sides of the
> observableuniverse are in thermal equilibrium with one another despite
> having beencausally separated from one another since the big bang.*
>

Alan Guth's Inflation theory explains that very nicely, the distant parts
of the universe we see with our largest telescopes are not causally
connected now but at one time there were, and then the universe expanded
much faster than light and that's why they are in thermal equilibrium now.
During inflation the expansion of the universe was exponential which means
it had a fixed doubling time, in this case every 10^-37 seconds the
diameter of the universe doubled. In 10^-35 seconds it doubled a hundred
times and it probably continued doubling for much longer than 10^-35
seconds. That also explains why at the largest scales spacetime looks flat,
the inflation smoothed out any curvature it may have had. And it explains
why we don't see any magnetic monopoles which should have been produced in
abundance in the era before inflation, the inflation thinned them out so
that now they are super rare.
>
>
>> Hoyle glosses over what he means by "choosing" but there are only 2 things
>> it could mean, you made the choice you did for a reason or you didn't,
>> so you're either a Cuckoo Clock or a  roulette wheel, but we're not
>> going to learn much from that.
>
>
> > *Cuckoo clocks are a bad example here. Cuckoo clocks don't
> "choose" anything,*


What exactly does "choose" mean?  I've asked people this question before
and I usually get an answer like "picking an action of your own free will",
but then I ask what "free will" means and they say the ability to choose.
And round and round we go. But the fact is you chose to do X rather than Y
because you prefer X. Why do you prefer X ? There are only 2 possibilities,
 there was a reason for you preferring it in which case you are in the
realm of cause and effect, or there was no cause for your preference in
which case it was random. So it's always a cuckoo clock or a roulette wheel.


> >
> *Conway wrote Life so that every state is used to compute a successor
> state, thus every state issolely a function of the previous state.*


The Life universe is deterministic so you outside the Life universe can
figure out its future if you know its present state, although a intelligent
being in that universe couldn't because before he finished calculating what
the future state will be the future would have already arrived. And nobody
can go in the other direction, nobody inside or outside the Life universe
could deduce its history just from the present state because more than one
past state could have produced the same present state.

>
> >
> *Furthermore Conway published theorems regarding free will which defined
> it as the ability to makechoices that are not a function of the past.*


That's an event without a cause, and that's the definition of random.


> >
> *Therefore, the cellular automata of Life do not have free will as defined
> by their creator.*


The only definition of free will that I know of that isn't gibberish is the
inability to always know for sure what you're going to do next until you
actually do it, and I find that no more mysterious and profound than the
fact that you don't know what the results of a calculation will be until
you've finished the calculation.

> *Whether Conway's definition of free will is correct or not is
> debatable, but I can deliberately choose to say a non-sequitur or do
> something unpredicatbly spontaneous.*


I agree, there is no law of logic that demands every event have a cause,
randomness is possible. And there is another name for doing things for no
reason, irrational.
>
>
>> We know you can make a Turing Machine in the Life universe and if you cando
>> that then you can make make a intelligent machine and if you did that you've
>> got a conscious machine, or at least you do unless Darwin was wrong. I
>> don't think he was wrong.
>>
>
> > *That does not follow at all.*


I sure don't see why that doesn't follow! Evolution can't see consciousness
so it certainly could not have selected for it, and yet it produced me and
I am conscious. So Evolution must have selected for something that it can
see, like intelligence, and consciousness just rode in on its coattails.


> *>A Turing Machine is a convenient abstract mathematical object *


A Turing Machine is the simplest form of computer but it is not an
abstraction , it can be and has been built:

https://www.youtube.com/watch?v=E3keLeMwfHY&t=39s

And it is a common misconception that a Turing Machine requires an infinite
tape but that is not true, it only needs a sufficient tape, so if it runs
low you give it more tape, but at any finite time it has only used a finite
amount of tape,  and certainly if it halts and provides an answer to a
calculation only a finite amount of tape was used.

> that is horribly innefficient compared to real computers.


Being so simple it's not surprising that its very slow and impractical, but
the logical design of any computer can always be mapped to a Turing
Machine, if a Turing Machine can't do something in a finite amount of time,
such as find the 7918th Busy Beaver number (it probably can't even find the
5th Busy Beaver number and all we'll ever know are the first 4), then no
computer can ever find it, not even a quantum Computer.

> Furthermore a Turing Machine is like a combination of hardware and
> software


All Turing Machines have the same hardware, only the software changes, and
some have more internal states than others.

*> so every possible algorithm is its own Turing Machine*


Yes, and when it has halted and solved that problem the Turing Machine will
be in a particular state, and that state could be part of a larger Turing
Machine that contains more states. One Turing Machine can act as a
subroutine for a larger Turing Machine that contains more states.

> *So if I can implement a Turing Machine to add integers in Life, it does
> not necessarily mean that I can use Life to run Windows or John Clark.*


No, if Life is Turing complete, and it is, then it necessarily does mean it
can emulate Windows or a Mac or the actions of John Clark. Of course to do
all that the Turing Machine used would need to able to go into many many
different states, but that's just a question of software, the hardware
wouldn't change.
>
>
> *Just because given an infinite amount of time, an intelligence could
> be computed on an abacus does not qualify the abacus as intelligent*


If it would take a infinite amount of  time (not just astronomically large
but infinite) then its a  non-computable problem and no computer can solve
it, and no human can either.

John K Clark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20181018/6475e2a5/attachment.html>


More information about the extropy-chat mailing list