The Orion's Arm Universe Project Forums





<Link> U. Leicester Students find problems with uploading
#1
Physics students from the University of Leicester have made made a calculation that may not be very promising to the possibility of 'beaming' uploads and Mind uploading.

https://physics.le.ac.uk/journals/index....ew/558/380


Attached Files
.pdf   teleportation paper 558-2375-1-PB.pdf (Size: 395.09 KB / Downloads: 1)
Reply
#2
They're assuming that you need to scan the brain down to the quantum level and are looking at the maximum possible amount of data involved. Other than Roger Penrose, no other researcher in studying the brain has suggested that quantum effects are involved in the operation of brains. Nor is there any particular reason to think that a fully functional brain is using anything near the maximum possible amount of information (other than human conceit that we are so uber special that it just must require that much to characterize something as special as we are).

Bottom line: They are looking at an absolute worst case scenario that is highly unlikely to reflect the reality of how brains operate.

I don't see anything here to indicate that either uploading or 'beaming' is in any danger or is any less possible.

Todd
Reply
#3
That's right. Uploading and engeneration in OA transfers much less information than the entire quantum load. In that paper the total information required for a human body and mind is calculated to be 4.5x 10e42 bits; I think an engenerator from the late First Federation would only transfer about 10e20 bits, so it would take about ten seconds to transmit a human using the transmitters they describe. Probably the amount of information transmited would increase in later, more sophisticated versions of the technology; at the end of this page http://www.orionsarm.com/eg-article/4a3d78dc77e6f I mention that full, quantum teleportation is possible using transapient technology, but the bandwidth for this sort of transfer must be gigantic.

The sort of uploading imagined in OA assumes that a mind can be copied at several levels of fidelity, and that the essential details of a personality can be described in much less than 4.5x 10e42 bits. This means that an uploaded personality in OA is a copy of the original, rather than having the same identity. A low-fi simulation of a person's identity could probably be acheived using only a few petabits, but obviously the more information tranferred the better.

But there is a limit to how much of an improvement could be made by increasing the fidelity of the copy - brains are warm, organic entities, and after a certain level all you are copying is thermal noise and brownian motion in liquids and colloids.
Reply
#4
Some of the possible different levels of fidelity are described in this .pdf book by Bostrom and Sandberg
http://www.philosophy.ox.ac.uk/__data/as...report.pdf
10e20 bits is equivalent to von Neumann's 1958 estimate of the memories stored over a lifetime, but there are many other estimates there.
Reply
#5
(08-02-2013, 12:03 PM)stevebowers Wrote: at the end of this page http://www.orionsarm.com/eg-article/4a3d78dc77e6f I mention that full, quantum teleportation is possible using transapient technology, but the bandwidth for this sort of transfer must be gigantic.

Not necessarily. As I understand it, quantum teleportation is based on transmitting the results of a series of measurements, which are then carried out on matter at the 'destination'. It doesn't actually involve transmitting data about the quantum state of every atom involved or every possible bit of data encoded down to the quantum level.

(08-02-2013, 12:03 PM)stevebowers Wrote: The sort of uploading imagined in OA assumes that a mind can be copied at several levels of fidelity, and that the essential details of a personality can be described in much less than 4.5x 10e42 bits. This means that an uploaded personality in OA is a copy of the original, rather than having the same identity. A low-fi simulation of a person's identity could probably be acheived using only a few petabits, but obviously the more information tranferred the better.

But there is a limit to how much of an improvement could be made by increasing the fidelity of the copy - brains are warm, organic entities, and after a certain level all you are copying is thermal noise and brownian motion in liquids and colloids.

We actually presume that it takes about 1e18bits to characterize a human (or equivalent) mind in OA. As far as an upload being a copy rather than having the same identity, unless someone can provide a hard definition of what an 'identity' even is, I don't see how we can say that it isn't the same. Also OA comm lasers can transmit orders of magnitude more data than that found in a human mind. See here:

http://www.orionsarm.com/eg-article/48545a0f6352a

A one petawatt (1e15W) laser is described as being able to transmit 2.28e31 bits/sec through a micrometer wormhole link. That's about 22 trillion human minds per second

As you say, below a certain level you aren't copying/transmitting information, but just noise. This level seems to be far above the quantum level. It seems likely that minds are the product of molecular chemistry, not quantum physics.

Todd
Reply
#6
Quote:We actually presume that it takes about 1e18bits to characterize a human (or equivalent) mind in OA.

I tend to think of that as a starting point, rather than a fixed datum throughout the timeline. 10e18 bit uploads might seen as primitive by the citizens of the Current Era, who like to copy themselves at higher resolution. Even then there will be significant numbers of people who reject uploading and engeneration, because of the uncertainty over the identity question.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)