I will quickly remark that some aspects of this comment seem to betray a non-info-theoretic point of view. From the perspective of someone like me, the key question for cryonics are “Do two functionally different start states (two different people) map onto theoretically indistinguishable molecular end states?” You are not an expert on the future possibilities of molecular nanotechnology and will not be asked to testify as such, but of course we all accept that arbitrarily great physical power cannot reconstruct a canister of ash because the cremation process maps many different possible starting people to widely overlapping possible canisters of ash. It is this question of many-to-one mapping alone on which we are interested in your expertise, and I would ask you to please presume for the sake of discussion that the end states of interest will be distinguished to molecular granularity (albeit obviously not to a finer position than thermal noise, let alone quantum uncertainty).
That said, I think we will all be interested if you can expand on
many aspects of synaptic strength and connectivity are irretrievably lost as soon as the synaptic membrane gets distorted
and whether you mean this in the customary sense of “it won’t boot back up when you switch it on” or in the info-theoretic sense of “this process will map functionally different synapses to exactly similar molecular states, or a spread of such states, up to thermal noise”. You are not being asked to overcome a burden of infinite proof either—heuristic argument is fine, we’re not asking for particular proofs you can’t possibly provide—we just want to make sure that what is being argued is the precise question we are interested in, that of many-to-one mappings onto molecular end states up to thermal noise.
EDIT: Oops, didn’t realize this was an old comment.
There, I’ve just caused you to scramble a vast array of concentration gradients, proteins tumbling around in a merry free-for-all. I have thus killed the old king, vive le roi nouveau!
If that’s not enough, consider the forces inertia exerts on all those precious gradients and precise molecule orientations when on a rollercoaster. Uh oh.
The molecular states may not have to be exactly similar after all to yield functional equivalency within an acceptable margin, a margin we deem acceptable throughout our lives. Let’s not worry about a few non-injective transformations?
I was initially swayed by Kalla724′s arguments (as well as a little molecular biology background), but it’s missing the forest for the trees.
I tried that. Now I’m a whole different combination proteins and chemicals. And this new me doesn’t understand how the point you are trying to illustrate relates to the grandparent any better than the old me.
Is it just tangential expression of your own position on broadly the same subject in loose agreement with Eliezer or is there an additional point you are trying to make?
I tried that. Now I’m a whole different combination proteins and chemicals. And this new me doesn’t understand how the point you are trying to illustrate relates to the grandparent any better than the old me.
You can’t simply replace unfolded proteins, since their relative position and concentration (and modification, and current status in several different signalling pathways) determines what happens to the signals that go through that synapse
I’m saying such a scrambling happens all the time anyways, and that preserving the exact relative position, concentration, folding status etcetera may not be all that important to the cognitive system at large, at least we don’t fret about it when shaking our heads. Does that help?
Edit to respond to your edit:
Is it just tangential expression of your own position on broadly the same subject in loose agreement with Eliezer or is there an additional point you are trying to make?
I’d say it’s an important point (naturally). If there were only one-to-one mappings, that would certainly be sufficient to establish that information-theoretically the original state information isn’t lost. But that’s a red herring, since we don’t even need that strong a claim to argue for the theoretical viability of cryonics:
When voluntarily shaking your head (you madman!) you were content with a much, much more forgiving standard. That’s the one that should be used for such discussions.
I’m saying such a scrambling happens all the time anyways, and that preserving the exact relative position, concentration, folding status etcetera may not be all that important to the cognitive system at large, at least we don’t fret about it when shaking our heads.
That seems like a reasonable and important point.
When voluntarily shaking your head (you madman!) you were content with a much, much more forgiving standard.
I’d probably go as far as to use a standard that accepted even a severe concussion or three’s worth of damage and information loss. Especially given that the ongoing functional impairment (albeit not the identity related loss) can presumably be trivially repaired.
There, I’ve just caused you to scramble a vast array of concentration gradients, proteins tumbling around in a merry free-for-all.
You probably haven’t actually, anymore then when you shake your hands vigorously back and forth the germs fly off. The force applied for so small a time is unlikely to have much of an effect on the cells which are dominated by chemical interactions,osmotic pressures,etc. Things don’t scale the way you’d like them to. Your whole argument is just invalid.
Edit to incorporate a point made below: which is good as if you scrambled the proteins in your brain, you’d die.
False analogy; there is a change in medium going from the oiled up dermis to air.
Take a glass of water with a large number of tagged proteins. That is the model of e.g. presynaptic vesicles filled with neurotransmitters, swimming in the cytoplasmic matrix (which is mostly water). A significant amount of them is not attached to anything. I didn’t say the folding structure would be scrambled, I said that the concentration gradients would be influenced, and that the orientation of non-attached proteins would change.
Shake the glass of water. What happens?
Shake the cytoplasmic matrix. What happens? What does such a rearrangement probably entail? That’s right, some change in concentration gradients, and a host of non-attached proteins tumbling around in a merry free-for-all.
Your whole argument is just invalid.
We can arrange to bet on our beliefs about this, say a 4 figure sum going to a charity of the other’s choice, with mutually agreed upon authorities in the field being the arbiters? If so send me a PM, and we can make the result public once we’re done.
EHeller was correct in so far as physical accelerations as occurring in every-day life do not have an effect on proteins and other small cell components which exceeds thermal noise.
I did win the bet since EHeller committed to a statement saying there would be no effect (on at least the order of magnitude of thermal noise) on any component of the cell, and as calculated by a referee, it turns out that larger cell organelles such as mitochondria are affected to such a degree (assuming 0.5g over 10s, 0.5g occurs e.g. when taking a car from 0 to 60mph in 6 seconds, so the 0.5g over 10s would occur for example when taking a car from 0 to 100mph over 10 seconds). Referee statement see here.
The wager goes to Miri, as chosen by Kawoomba, with thanks for all the fish.
I thank EHeller for his professional conduct and his charitable interpretations of my claims, resulting in the, well, result. Had he held me to my initial statements as made, he would have won.
I should have made my point in an entirely different way from the start: Namely, consider you took a psychoactive drug, such as an SSRI. Naturally, all sorts of channel configurations, concentration gradients and the such would be affected. But consider what happens after you stop taking it: you’d regress to some semblance (but not precisely the same state!) of the state you had before.
If someone paid you a large sum of money for taking SSRIs for 1 month, wouldn’t you do so? I probably would. I wouldn’t consider myself to be committing informational suicide, even though even post-SSRI-taking my state would still be different from my state now. That’s my point; many-to-one transformations—variances between the original state and the reconstructed state—don’t need to be functionally significant, and even when they are, it doesn’t follow that we couldn’t consider ourselves to be reconstituted, just as we don’t consider ourselves to be committing identity-suicide when taking SSRIs for a month.
Just to verify- we made the bet and unfortunately I allowed the burden of proof to shift to myself, rather than to Kawoomba, and the statement we used for the bet was effectively “everyday scale accelerations will have an affect on proteins or other biologically important compounds at most an order of magnitude below thermal noise.” This statement is technically not true, because fairly strong (~g), long-time-scale accelerations (10 s of seconds) will have an effect on the largest structures of just-about-thermal-noise (so basically if you get in one of those centrifuge-style-carnival-rides, you can affect your mitochrondia to order thermal noise).
A car doing 0-60 in 6 seconds won’t quite do it, it needs to be 10s of seconds.
Also this should underscore that bets might not be the best way to go about truth seeking :)
Also, to respond to the SSRI point- I doubt Kawoomba would take a large sum of money to take ANY psychoactive drug for 1 month- rather he knows the specific affects of SSRIs. Some changes in brain chemistry can be more important than others.
Hmm, for sufficiently large values of ‘large sum of money’, as long as the drug was randomly (not maliciously) chosen from a pool of FDA-approved psychoactive drugs, I would. Wouldn’t you, for one trillion dollars? If so, we’re just haggling about price.
The original example used 0.5g, just to give an impression that would be on the scale of going 0-60 in 6 seconds, for 10 seconds just keep up the acceleration for 4 seconds longer.
Also, given “one magnitude less than thermal noise” (as in the statement), apparently normal rollercoasters are, after all, sufficient to affect even smaller structures such as lysosomes at >58 (fifty-eight) nm, see this amendment.
We can arrange to bet on our beliefs about this, say a 4 figure sum going to a charity of the other’s choice, with mutually agreed upon authorities in the field being the arbiters? If so send me a PM, and we can make the result public once we’re done.
To be clear, you want to bet that shaking your head side-to-side vigorously causes the same qualitative type of changes to protein concentrations along cell membranes as introducing a large concentration of toxic cryoprotectants?
I want to emphasize along cell membranes, as I believe those are the concentrations kalla was referring to, its where the information will be lost, and is why I referred to osmotic pressure and chemical interactions being very dominant, and such forces are not present in your glass of water analogy (other than shaking the glass will NOT cause the proteins to precipitate out), so I want to make sure we are talking about the same thing.
To be clear, you want to bet that shaking your head side-to-side vigorously causes the same qualitative type of changes to protein concentrations along cell membranes as introducing a large concentration of toxic cryoprotectants?
...
No, I am not saying that shaking your head will rupture many of your phospholipid bilayers and in effect kill you.
Glad we could clear that up.
(What I am saying is “after the damage (e.g. cell membranes) has been repaired, certain information was not recoverable” not being a death knell for cryonics, since we accept a considerable margin of error concerning e.g. specific neurotransmitter vesicle locations in daily life without considering ourselves a different person just because we’ve effected a host of changes in distributions and arrangements. An example other than shaking your head, which also strongly affects cross-membrane distributions: SSRIs)
I agree that bioactive drugs will have effects on distributions, as thats what they exist for, they also cause (sometimes mild, sometimes extreme) personality changes. What I’m interested in is how wrong can we get things and still have a “working” brain.
I’m not sure Kawoomba would really like brains to be scrambled by head shaking. I expect that all else being equal he would prefer as little disruption as possible.
Shaking your head applies relatively uniform forces, which elasticity and natural repair mechanisms can deal with. Even then it can be a close thing; people have been known to take permanent damage from trivial-seeming head injuries.
Freezing applies non-uniform forces. It’s the difference between riding an elevator and hopping into a blender.
I’m not interested in damage so much as in changes in e.g. “the exact orientation of presynaptic vesicles” not being integral to our personal identity.
If e.g. someone said “we have no scanning techniques that can tell us how exactly each molecule was oriented, it could have been one of many ways (since you’re e.g. trying to reverse a non-injective function”, I’d say “well, we constantly change that orientation by random body movements, yet don’t mind. So we can assume that change is not identity-constituting.”
Edit: Even a uniform force will affect soluble elements differently from other elements. The stiffness is different for various body elements, which will lead to all sorts of tiny changes: Erythrocytes being pressed against the cell wall etcetera. That’s not a significant change so that we’d say “we leave the elevator a different person”, but that’s precisely the point in the comparison with certain information that can’t be read from a cell: It may not make all that much of a difference.
I will quickly remark that some aspects of this comment seem to betray a non-info-theoretic point of view. From the perspective of someone like me, the key question for cryonics are “Do two functionally different start states (two different people) map onto theoretically indistinguishable molecular end states?” You are not an expert on the future possibilities of molecular nanotechnology and will not be asked to testify as such, but of course we all accept that arbitrarily great physical power cannot reconstruct a canister of ash because the cremation process maps many different possible starting people to widely overlapping possible canisters of ash. It is this question of many-to-one mapping alone on which we are interested in your expertise, and I would ask you to please presume for the sake of discussion that the end states of interest will be distinguished to molecular granularity (albeit obviously not to a finer position than thermal noise, let alone quantum uncertainty).
That said, I think we will all be interested if you can expand on
and whether you mean this in the customary sense of “it won’t boot back up when you switch it on” or in the info-theoretic sense of “this process will map functionally different synapses to exactly similar molecular states, or a spread of such states, up to thermal noise”. You are not being asked to overcome a burden of infinite proof either—heuristic argument is fine, we’re not asking for particular proofs you can’t possibly provide—we just want to make sure that what is being argued is the precise question we are interested in, that of many-to-one mappings onto molecular end states up to thermal noise.
EDIT: Oops, didn’t realize this was an old comment.
Shake your head. Vigorously. (Do it.)
There, I’ve just caused you to scramble a vast array of concentration gradients, proteins tumbling around in a merry free-for-all. I have thus killed the old king, vive le roi nouveau!
If that’s not enough, consider the forces inertia exerts on all those precious gradients and precise molecule orientations when on a rollercoaster. Uh oh.
The molecular states may not have to be exactly similar after all to yield functional equivalency within an acceptable margin, a margin we deem acceptable throughout our lives. Let’s not worry about a few non-injective transformations?
I was initially swayed by Kalla724′s arguments (as well as a little molecular biology background), but it’s missing the forest for the trees.
I tried that. Now I’m a whole different combination proteins and chemicals. And this new me doesn’t understand how the point you are trying to illustrate relates to the grandparent any better than the old me.
Is it just tangential expression of your own position on broadly the same subject in loose agreement with Eliezer or is there an additional point you are trying to make?
:-D
Take for example:
I’m saying such a scrambling happens all the time anyways, and that preserving the exact relative position, concentration, folding status etcetera may not be all that important to the cognitive system at large, at least we don’t fret about it when shaking our heads. Does that help?
Edit to respond to your edit:
I’d say it’s an important point (naturally). If there were only one-to-one mappings, that would certainly be sufficient to establish that information-theoretically the original state information isn’t lost. But that’s a red herring, since we don’t even need that strong a claim to argue for the theoretical viability of cryonics:
When voluntarily shaking your head (you madman!) you were content with a much, much more forgiving standard. That’s the one that should be used for such discussions.
That seems like a reasonable and important point.
I’d probably go as far as to use a standard that accepted even a severe concussion or three’s worth of damage and information loss. Especially given that the ongoing functional impairment (albeit not the identity related loss) can presumably be trivially repaired.
You probably haven’t actually, anymore then when you shake your hands vigorously back and forth the germs fly off. The force applied for so small a time is unlikely to have much of an effect on the cells which are dominated by chemical interactions,osmotic pressures,etc. Things don’t scale the way you’d like them to. Your whole argument is just invalid.
Edit to incorporate a point made below: which is good as if you scrambled the proteins in your brain, you’d die.
False analogy; there is a change in medium going from the oiled up dermis to air.
Take a glass of water with a large number of tagged proteins. That is the model of e.g. presynaptic vesicles filled with neurotransmitters, swimming in the cytoplasmic matrix (which is mostly water). A significant amount of them is not attached to anything. I didn’t say the folding structure would be scrambled, I said that the concentration gradients would be influenced, and that the orientation of non-attached proteins would change.
Shake the glass of water. What happens?
Shake the cytoplasmic matrix. What happens? What does such a rearrangement probably entail? That’s right, some change in concentration gradients, and a host of non-attached proteins tumbling around in a merry free-for-all.
We can arrange to bet on our beliefs about this, say a 4 figure sum going to a charity of the other’s choice, with mutually agreed upon authorities in the field being the arbiters? If so send me a PM, and we can make the result public once we’re done.
Postmortem for the bet:
EHeller was correct in so far as physical accelerations as occurring in every-day life do not have an effect on proteins and other small cell components which exceeds thermal noise.
I did win the bet since EHeller committed to a statement saying there would be no effect (on at least the order of magnitude of thermal noise) on any component of the cell, and as calculated by a referee, it turns out that larger cell organelles such as mitochondria are affected to such a degree (assuming 0.5g over 10s, 0.5g occurs e.g. when taking a car from 0 to 60mph in 6 seconds, so the 0.5g over 10s would occur for example when taking a car from 0 to 100mph over 10 seconds). Referee statement see here.
The wager goes to Miri, as chosen by Kawoomba, with thanks for all the fish.
I thank EHeller for his professional conduct and his charitable interpretations of my claims, resulting in the, well, result. Had he held me to my initial statements as made, he would have won.
I should have made my point in an entirely different way from the start: Namely, consider you took a psychoactive drug, such as an SSRI. Naturally, all sorts of channel configurations, concentration gradients and the such would be affected. But consider what happens after you stop taking it: you’d regress to some semblance (but not precisely the same state!) of the state you had before.
If someone paid you a large sum of money for taking SSRIs for 1 month, wouldn’t you do so? I probably would. I wouldn’t consider myself to be committing informational suicide, even though even post-SSRI-taking my state would still be different from my state now. That’s my point; many-to-one transformations—variances between the original state and the reconstructed state—don’t need to be functionally significant, and even when they are, it doesn’t follow that we couldn’t consider ourselves to be reconstituted, just as we don’t consider ourselves to be committing identity-suicide when taking SSRIs for a month.
Just to verify- we made the bet and unfortunately I allowed the burden of proof to shift to myself, rather than to Kawoomba, and the statement we used for the bet was effectively “everyday scale accelerations will have an affect on proteins or other biologically important compounds at most an order of magnitude below thermal noise.” This statement is technically not true, because fairly strong (~g), long-time-scale accelerations (10 s of seconds) will have an effect on the largest structures of just-about-thermal-noise (so basically if you get in one of those centrifuge-style-carnival-rides, you can affect your mitochrondia to order thermal noise).
A car doing 0-60 in 6 seconds won’t quite do it, it needs to be 10s of seconds.
Also this should underscore that bets might not be the best way to go about truth seeking :)
Also, to respond to the SSRI point- I doubt Kawoomba would take a large sum of money to take ANY psychoactive drug for 1 month- rather he knows the specific affects of SSRIs. Some changes in brain chemistry can be more important than others.
Hmm, for sufficiently large values of ‘large sum of money’, as long as the drug was randomly (not maliciously) chosen from a pool of FDA-approved psychoactive drugs, I would. Wouldn’t you, for one trillion dollars? If so, we’re just haggling about price.
The original example used 0.5g, just to give an impression that would be on the scale of going 0-60 in 6 seconds, for 10 seconds just keep up the acceleration for 4 seconds longer.
Also, given “one magnitude less than thermal noise” (as in the statement), apparently normal rollercoasters are, after all, sufficient to affect even smaller structures such as lysosomes at >58 (fifty-eight) nm, see this amendment.
To be clear, you want to bet that shaking your head side-to-side vigorously causes the same qualitative type of changes to protein concentrations along cell membranes as introducing a large concentration of toxic cryoprotectants?
I want to emphasize along cell membranes, as I believe those are the concentrations kalla was referring to, its where the information will be lost, and is why I referred to osmotic pressure and chemical interactions being very dominant, and such forces are not present in your glass of water analogy (other than shaking the glass will NOT cause the proteins to precipitate out), so I want to make sure we are talking about the same thing.
...
No, I am not saying that shaking your head will rupture many of your phospholipid bilayers and in effect kill you.
Glad we could clear that up.
(What I am saying is “after the damage (e.g. cell membranes) has been repaired, certain information was not recoverable” not being a death knell for cryonics, since we accept a considerable margin of error concerning e.g. specific neurotransmitter vesicle locations in daily life without considering ourselves a different person just because we’ve effected a host of changes in distributions and arrangements. An example other than shaking your head, which also strongly affects cross-membrane distributions: SSRIs)
I agree that bioactive drugs will have effects on distributions, as thats what they exist for, they also cause (sometimes mild, sometimes extreme) personality changes. What I’m interested in is how wrong can we get things and still have a “working” brain.
I’m not sure Kawoomba would really like brains to be scrambled by head shaking. I expect that all else being equal he would prefer as little disruption as possible.
Shaking your head applies relatively uniform forces, which elasticity and natural repair mechanisms can deal with. Even then it can be a close thing; people have been known to take permanent damage from trivial-seeming head injuries.
Freezing applies non-uniform forces. It’s the difference between riding an elevator and hopping into a blender.
I’m not interested in damage so much as in changes in e.g. “the exact orientation of presynaptic vesicles” not being integral to our personal identity.
If e.g. someone said “we have no scanning techniques that can tell us how exactly each molecule was oriented, it could have been one of many ways (since you’re e.g. trying to reverse a non-injective function”, I’d say “well, we constantly change that orientation by random body movements, yet don’t mind. So we can assume that change is not identity-constituting.”
Edit: Even a uniform force will affect soluble elements differently from other elements. The stiffness is different for various body elements, which will lead to all sorts of tiny changes: Erythrocytes being pressed against the cell wall etcetera. That’s not a significant change so that we’d say “we leave the elevator a different person”, but that’s precisely the point in the comparison with certain information that can’t be read from a cell: It may not make all that much of a difference.