On 6/21/2018 6:42 PM, Jason Resch wrote:
On Thu, Jun 21, 2018 at 7:26 PM, Brent Meeker <[email protected]
<mailto:[email protected]>> wrote:
On 6/21/2018 3:29 PM, Jason Resch wrote:
On Thu, Jun 21, 2018 at 5:10 PM, Brent Meeker
<[email protected] <mailto:[email protected]>> wrote:
On 6/21/2018 7:47 AM, Jason Resch wrote:
On Thu, Jun 21, 2018 at 12:59 AM, Brent Meeker
<[email protected] <mailto:[email protected]>> wrote:
On 6/20/2018 9:48 PM, Jason Resch wrote:
On Tue, Jun 19, 2018 at 5:24 PM, Brent Meeker
<[email protected] <mailto:[email protected]>> wrote:
On 6/19/2018 7:10 AM, Jason Resch wrote:
On Tue, Jun 19, 2018 at 12:21 AM, Brent Meeker
<[email protected]
<mailto:[email protected]>> wrote:
On 6/18/2018 4:27 AM, Jason Resch wrote:
On Sun, Jun 17, 2018 at 9:57 PM, Brent Meeker
<[email protected]
<mailto:[email protected]>> wrote:
On 6/17/2018 4:43 AM, Telmo Menezes wrote:
On 17 June 2018 at 13:26,
<[email protected]
<mailto:[email protected]>> wrote:
On Sunday, June 17, 2018 at
10:15:05 AM UTC, Jason wrote:
On Sun, Jun 17, 2018 at 12:12
AM, <[email protected]
<mailto:[email protected]>>
wrote:
why do you prefer the
MWI compared to the
Transactional Interpretation?
I see both as absurd. so
I prefer to assume the wf
is just epistemic, and/or
that we have some holes
in the CI which have yet
to be resolved. AG
--
1. It's the simplest theory:
"MWI" is just the Schrodinger
equation,
nothing else. (it doesn't say
Schrodinger's equation only
applies sometimes,
or only at certain scales)
2. It explains more while
assuming less (it explains
the appearance of
collapse, without having to
assume it, thus is preferred
by Occam's razor)
3. Like every other
successful physical theory,
it is linear, reversible
(time-symmetric), continuous,
deterministic and does not
require faster than
light influences nor
retrocausalities
4. Unlike single-universe or
epistemic interpretations,
"WF is real" with
MWI is the only way we know
how to explain the
functioning of quantum
computers (now up to 51 qubits)
5. Unlike copenhagen-type
theories, it attributes no
special physical
abilities to observers or
measurement devices
6. Most of all, theories of
everything that assume a
reality containing
all possible observers and
observations lead directly to
laws/postulates of
quantum mechanics (see
Russell Standish's Theory of
Nothing, Chapter 7 and
Appendix D).
Given #6, we should revise
our view. It is not MWI and
QM that should
convince us of many worlds,
but rather the assumption of
many worlds (an
infinite and infinitely
varied reality) that gives
us, and explains all the
weirdness of QM. This should
overwhelmingly convince us of
MWI-type
everything theories over any
single-universe
interpretation of quantum
mechanics, which is not only
absurd, but completely devoid
of explanation.
With the assumption of a
large reality, QM is made
explainable and
understandable: as a theory
of observation within an
infinite reality.
Jason
You forgot #7. It asserts
multiple, even infinite copies of
an observer,
replete with memories, are
created when an observer does a
simple quantum
experiment. So IMO the alleged
"cure" is immensely worse than
the disease,
CI, that is, just plain idiotic. AG
It is important to make the
distinction between our intuition and
common sense and actual formal
reasoning. The former can guide the
latter very successfully, but the
history of science teaches us that
this is not always the case. You
don't provide an argument, you just
present your gut feeling as if it
were the same thing as irrefutable
fact.
I think Scott Aaronson has the right
attitude toward this:
https://www.scottaaronson.com/blog/?p=326
<https://www.scottaaronson.com/blog/?p=326>
As such a strong believer in quantum
computers (he's staked $100,000 of his own
money on the future construction of large
scale quantum computers), I would love to ask
Scott Aaronson what he thinks about running a
conscious AI on such a quantum computer. That
trivially leads to "many worlds" at least as
seen by that AI.
If it's so trivial maybe you can explain it.
1. A quantum computer is isolated from the
environment so as to remain in a super position of
many possible states.
2. Quantum computers are Turing universal,
anything that can be programmed on a classical
computer can be programmed on a quantum computer
3. Assuming Computational Theory of mind, a
quantum computer can execute the same conscious
program as "Brent Meeker's Brain"
4. The quantum computer can be arranged to
entangle an unmeasured particle with Brent
Meeker's quantum brain emulation,
a) by feeding in spin up as an auditory tone in
Brent Meeker's left auditory nerve
b) by feeding in spin down as an auditory tone in
Brent Meeker's right auditory nerve
5. The quantum brain simulation, being isolated
from the environment, remains in a super position
of the Brent Meeker brain emulation hearing an
auditory tone in his left and right ears.
You can repeat this process 30 times, with 30
different measurements of different electrons, and
end up with over 1 billion Brent Meeker brain
emulations, each remembering a different pattern
of auditory tones.
For the Brent Meeker quantum brain emulation, many
worlds is definitely true.
No. If decoherence occurs when there a many
degrees of freedom in which to disperse
entanglements then my brain is plenty big enough to
decohere the signal; and you seem to assume this
when supposing that I form different memories.
Otherwise I wouldn't form any definite memory, my
memory would merely exist in a superposition of a
billion different patterns.
That is the whole point of (and difficulty) of making a
quantum computer. Its qubits must remain isolated from
the rest of the environment such that it does not
decohere while it is computing something. You seem to
be postulating some upperbound on how large quantum
computers can get. This is the exact thing Scott
Aaronson has staked $100K on (that large scale quantum
computers /can/ be built), which is why I find his
antipathy towards MWI so paradoxical. If they can be
built, then we can create many-experiences by running
an AI emulation on a quantum computer, where some of
the qubit registers are prepared in an undetermined state.
How will you know it has many experiences?
If computationalism is true (which Aarson has defended), it
will have an experience.
An experience is not many experiences. And what does
computationalism mean; it gets used sloppily on this list,
sometimes meaning only that "saying yes to the doctor" is
justified, other times meaning that Bruno's whole theory is true?
By computationalism, I mean there exists a computation that if
performed/implemented by any machine, it is sufficient to
instantiate a given conscious experience.
By the way, I e-mailed Scott Aaronson asking about the thought
experiment I gave above regarding running a conscious AI/brain
emulation on a quantum computer that enters a superposition, and
just got a reply.
He said that he agrees that if consciousness is inherent to a
particular computation (including a computation restricted to
only one branch of the wave function), and if large-scale quantum
computers are possible, then you could have a superposition of
different conscious experiments and this appears to force one to
a many-worlds picture. He said he has made this same point before
in some of his blog posts and writings.
What he thought what least clear from my proposed thought
experiment concerned what is necessary for a conscious
computation, and gave the examples of alternate theories of
consciousness which required coherence, or irreversibility, for
example.
However, I think from a plain "computational theory of mind",
which is free from any quantum mechanical/physical definitions,
he is in general agreement that Computationalism + Quantum
computers large enough to run conscious programs, yields many-worlds.
It won't be able to say what they are.
Sure it can, within its virtual reality it can say or do
anything. Whether or not it can tell us what it sees is
another question. I would say if we decide to cause the
quantum computer to decohere and entangle ourselves with its
state, we will hear what it is saying (but in each branch we
will hear it say only one thing).
Exactly my point.
But if those other computations (which we know must exist, as
they are necessary to explain the functioning of quantum
computers) to say they are not conscious is to abandon
Computationalism.
I don't think that follows. I think consciousness (and thought in
general) is a classical phenomenon.
We are talking about classical computations through. (The same form
that classical computers implement). The only difference is, we are
using a quantum computer to perform those classical computations in
parallel.
OK. That's different. Except I don't think you can do that. You can
only do classical computations in parallel on a quantum computer in the
same sense you can do parallel computations on a classical computer.
But to the extent you can, I would agree that they would produce some
cosciousness (although ultimately I think conscious computations require
an environment to give them meaning).
So a superposition of different computations would only be a
bunch of parallel consciousnesses (which I think Scott discounts)
if each existed in a parallel classical world.
I am not sure I follow what you are saying here. I agree a super
position of different computations would only be a bunch of parallel
consciousnesses.
No. A superposition is different from just a collection of classical
states. A system that is in a superposition of states is sort of in all
of them at once and not in any one of them.
(Scott seemed to agree this would be the result under standard
computationalism, but might not under some non-standard form of
computationalism that additionally required some necessary bit of
physicalism--I would call this physicalism rather than computationalism).
But in that case only one could communicate with us.
Well, unless we too decide to join the superposition (which I argue we
do when the quantum computer decoheres). Again, no one consciousness
gets to interview all the others (beyond whatever interference might
permit).
It won't be able to act intelligently in more than one
world. Scott also notes that quantum computers solve
problems by having destructive interference zero out the
probability of incorrect solutions...which means
computation all happens in the same world.
That is when quantum computers are used to obtain a single
definite result in all branches. This is what can make
quantum computers more powerful. But I am not using this, I
am merely riding off the quantum computer's ability to
maintain a large scale superposition by virtue of a quantum
computer's ability to remain isolated from its environment
while it computes what it does.
But then it doesn't actually compute anything. In the words
of Schroedinger it is jellified.
It must have computed that trace, and by computationalism, that
execution trace in that branch must have been conscious. You
can't explain quantum computers without assuming those executions
in the other branches all exist.
/*Unless*/ you think wave function collapse has the power to
delete an /already-had/ experience out of existence,
No, I'm saying there is no experience except classical
experiences...we never experience superpositions.
Each branch of the superposition represents a classical computation.
So they are not a superposition they're a mixture.
Are you saying this computation doesn't exist, or that a computation
can only be conscious if it is constantly interacting with its
environment? If I isolate my laptop computer from the rest of the
environment for a few hours while it works on some problem, and then
check it after it is done, will it not still have computed the
expected result? Do you reject the reality of the intermediate steps
of the computation?
In the real world "collapse" is continuous except in carefully
contrived experiments. Those other branches exist in a quantum
computer because they interfere, all in this world and produce one
classical result. I see no reason to suppose there is any
experience apart from classical results.
Either you reject the reality and efficacy of the intermediate
per-branch computations (which is an anti-realist position that fails
to explain how quantum computers can produce final results, if the
intermediate per-branch computational states don't compute what
they're supposed to such that they interfere correctly at the end to
yield the expected result)
But if it's a quantum computer the intermediate branches are not
classical and so do not instantiate experiences.
Or you are rejecting computationalism and replacing it with your own
theory, which says, conventional computation is insufficient. Only a
computation performed by a physical computer that frequently (how
frequently?) interacts with its environment (how big does the
environment need to be, how do you define where the computer ends and
the environment begins?) can be conscious. This theory seems a bit ad
hoc, and whose sole motivation is to zombify the AIs in the branches
of the quantum computer.
That's one on my complaints about Bruno's step 8 "proof", he
claims to have shown computationalism inconsistent with
materialism. But it only seems that way because he ignores the
necessity of having a whole physical environment to have
experience in. But then you recognize that the physical world is
a necessary component and must exist to make computationalism
meaningful.
I will let Bruno reply to this, but I think we need to be very careful
about the subtle role physics plays. To be clear Bruno does not
reject the existence of a physical reality, but says its existence is
made a theorem in arithmetic and becomes a necessary role in relating
experiences between different machines.
There is also the difference between the appearance of the physical
world and the infinity of computations below us competing with each
other in that appearance. If we look at the Boltzmann bran
computations, some of those are larger programs that may contain many
conscious sub-computations, and if in one of those computations we
build a computer which itself manifests as a sub-routine of that
larger computation you could say we have realized a new conscious
entity. But is conscious computation exists as arithmetical
computations, and appears infinitely often in the UD, so we should not
confuse a local incarnation of a computation with the observer, whose
future experiences are not bound to the evolution of the
locally-realized computation. I think this is why Bruno says we
cannot physically create an observer, only locally instantiate one
relatively to us. But the two forms: the single deterministic machine
instantiation, and the observers first person point of view supported
by infinite computations, remain distinct.
out of the past, in a manner such that it never existed, and we
were wrong to believe that it ever did exist. But this seems
magical to me.
I know. But having worlds coming into existence on a continuum
seems just fine.
So you reject QM on that basis?
No, I'm not rejecting anything, I'm suggesting you consider other
possibilities which are no crazier than MWI.
Brent
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
