There are two components to this. The first ignores the economic issues to invoke a castrophe, i.e., that information requires energy to create and destroy it. And this energy equates to mass, that at current rates of information growth would equal half the earth's mass in 150 years. This ignores the economic constraints that would put a brake on any such process long before it can become a catastrophe.
The second is, umm, ridiculous on the face of it to put it mildly, namely, that information literally has mass all by itself, and that a storage medium full of coherent data weighs more than the same media in an erased state. This is from Vopson's paper: https://aip.scitation.org/doi/full/10.1063/1.5123794
He further extends this to propose that dark matter is information mass. I can't even...
The obvious problem here is that "erased state" means nothing. From the standpoint of information, there is no difference between an "erased" and "non-erased" storage medium. Both contain exactly the same amount of information. If there is an information component to mass, it is already implicit in the mass of baryonic matter. It cannot be added to or taken away. It can only be transformed from one information state to another, and thus any such implicit mass would be preserved.
> Both contain exactly the same amount of information.
I would think erased data would have higher entropy, assuming it's relative to the interpreter which would contextualize the meaning of erased or valid data. "Erased" would mean the action of raising the entropy on some arrangement of matter.
The word "erased" is undefined. Erased could be contiguous 0's, 1's, random data, or some pattern. And in every case, those are all information in and of themselves. The amount of information entropy does not matter. Even if one postulates that entropy and mass are inversely related, that makes no difference on a storage media. The underlying media itself has the same entropy from the standpoint of its baryonic mass regardless of the information state it hosts.
Even the concept of baryonic entropy contributing a mass component is rather ludicrous. This is akin to the idea that either the Higgs field, and/or the quark and gluon energy within a hadron respond to the specific arrangements of the hadrons to which they impart mass. I.e., that the mass of the particles changes in some proportion relative to their physical relationships to one another in spacetime. That's a rather extraordinary claim that is nowhere in the math.
I don't know if erase is undefined... it takes work to erase something, and every time you do the entropy surely is higher. This is why the word itself is also a metaphor for destruction, to erase someone or something would be to destroy it. It's akin to blowing something up, one second there is a complex structure of some kind i.e. a military facility (low entropy), and the next second it's a pile of rubble i.e. high entropy.
From your example of storage medium, I think you're missing that someone or something has to do the erasing in said storage medium (by definition, there is an interpreter interpreting the information in that example). So "erase" would be defined by the interpreter, however by any definition it would involve destroying information, and that would require work.
Yes, I understand that work is involved, but the energy involved in storing and altering information is not the mass equivalence this paper is talking about. That is a separate question Vopson addresses, the one which ignores the economic factors, and the one that relies on simple mass-energy equivalence. It's some supposed "information catastrophe" that just isn't possible.
But the radical claim of this paper, and the one I am addressing, is that information, in and of itself, contributes a mass component to a storage media regardless of the media involved. That's a completely different proposition.
All baryonic matter has an information component. But the theorems concerning information stipulate that information is never destroyed. Entropy does not destroy it either. This paper violates the central theorem of information science. That is why "erased" here is meaningless. Entropy state is irrelevant.
Further, it violates the conservation of mass-energy. The work required to "erase" a given medium is the same as the work required to record information on it. Yet this paper claims the one process adds mass, and the other removes it. The whole thing is just nonsense.
6 comments
[ 4.8 ms ] story [ 25.9 ms ] threadThe second is, umm, ridiculous on the face of it to put it mildly, namely, that information literally has mass all by itself, and that a storage medium full of coherent data weighs more than the same media in an erased state. This is from Vopson's paper: https://aip.scitation.org/doi/full/10.1063/1.5123794
He further extends this to propose that dark matter is information mass. I can't even...
The obvious problem here is that "erased state" means nothing. From the standpoint of information, there is no difference between an "erased" and "non-erased" storage medium. Both contain exactly the same amount of information. If there is an information component to mass, it is already implicit in the mass of baryonic matter. It cannot be added to or taken away. It can only be transformed from one information state to another, and thus any such implicit mass would be preserved.
I would think erased data would have higher entropy, assuming it's relative to the interpreter which would contextualize the meaning of erased or valid data. "Erased" would mean the action of raising the entropy on some arrangement of matter.
Even the concept of baryonic entropy contributing a mass component is rather ludicrous. This is akin to the idea that either the Higgs field, and/or the quark and gluon energy within a hadron respond to the specific arrangements of the hadrons to which they impart mass. I.e., that the mass of the particles changes in some proportion relative to their physical relationships to one another in spacetime. That's a rather extraordinary claim that is nowhere in the math.
From your example of storage medium, I think you're missing that someone or something has to do the erasing in said storage medium (by definition, there is an interpreter interpreting the information in that example). So "erase" would be defined by the interpreter, however by any definition it would involve destroying information, and that would require work.
But the radical claim of this paper, and the one I am addressing, is that information, in and of itself, contributes a mass component to a storage media regardless of the media involved. That's a completely different proposition.
All baryonic matter has an information component. But the theorems concerning information stipulate that information is never destroyed. Entropy does not destroy it either. This paper violates the central theorem of information science. That is why "erased" here is meaningless. Entropy state is irrelevant.
Further, it violates the conservation of mass-energy. The work required to "erase" a given medium is the same as the work required to record information on it. Yet this paper claims the one process adds mass, and the other removes it. The whole thing is just nonsense.