I don’t think that’s a good analogy; IIRC organs (eg rabbit kidneys) have been successfully frozen and revived (good enough to implant), so it’s more a matter of whether that can be extended to human brains (which, sure, may be more delicate) rather than being something inherently absurd.
Rabbit kidneys are much smaller than human brains.
The square-cube law is the main showstopper: you can remove heat form a thing at a rate proportional to its surface area, while its heat capacity is proportional to mass and thus to volume. Therefore, maximum attainable cooling speed decreases with size (if you try to cool any faster, youl’ll just crack the surface).
Rabbit kidneys can be vitrified without using a toxic concentration of cryprotectants, moreover, IIUC the circulatory system of a kidney allows higher flow and pressure (a kidney is just a blood filter, after all), making cryoprotectant perfusion easier. Even then, cryopereservation isn’t perfect: microscopic damage has been observed.
Thanks for the information, but that suggests that preserving a human brain will be difficult and may require more advanced techniques than currently used, not necessarily that it’s some crazy impossible thing that shouldn’t even be thought about. Hell, maybe it would be possible to carefully cut up the brain into smaller chunks before freezing it (a sharp cut along the right line being perhaps not so damaging compared to bad freezing).
A lot of it is going to come down to exactly how memories are stored and how redundant they are. Last time I checked this wasn’t yet fully understood. If they really do depend on fine details of molecules that are inevitably irreversibly scrambled by freezing, then it probably is impossible after all.
I don’t think anybody is claiming that viable brain preservation will be necessarily forever impossible. The claim is that brain preservation as currently offered by cryonics comapanies is probably flawed and unlikely to maintain the relevant aspects of somebody’s personal identity.
Fair enough, though I do think this opinion is sometimes expressed a bit over-confidently given that the physical basis of memory is not yet well understood.
Fair enough, though I do think this opinion is sometimes expressed a bit over-confidently given that the physical basis of memory is not yet well understood.
Suppose that I were sceptical that boiling a head in a cooking pot for 2 hours followed by freezing preserved the information… you could say exact same words and they would be equally relevant (or irrelevant).
The “not yet well understood” does not mean it is warranted to plug in some entirely unspecified magic. We know that how-ever it is stored, it must affect transmission of the signals between brain cells. Which leaves us with receptor densities, positioning of receptors, adhesion of molecules to receptors, states of receptors, and the like.
Ultimately, when we do not have an actual reason to believe some procedure works, and only assume it might work from ignorance and introduction of too much magic, it falls to the background of considerations such as “what if donating that amount of money to best charities will make it more likely that the future people will use look-into-the-past wormholes to resurrect you?”. Or all the variations on the theme of living in a simulator (which may well use your brain’s data for something provided it fits some criteria). You need evidence to elevate one such idea above the milliard others.
There’s already enough evidence to locate the hypothesis—eg experiments on animal organs and small animals. Therefore your assertion that this is some crazy idea plucked out of nowhere doesn’t hold.
The other hypotheses were located in similar manner, though. And far too much has to be ignored to generalize from said animal organs or said microscopic animals. Generalization from kidneys to brains is particularly dubious. Especially as great many of said frozen corpses are frozen corpses precisely because of their brain’s unusual fragility with regards to loss of blood supply.
edit: The issue is not with cryonics in principle. The issue is with cryonics as it is. Similar to the people jumping off towers with some wings vs an airplane. Seeing a bird fly really doesn’t make for a case that you can fly by your own muscular power with some homemade birdlike wings, with the early prototype, in fact it makes for the opposite case (as none of the birds are as heavy as you are). Evidence against a hypothesis can locate it too.
edit2: And with regards to information theory, it is absolutely trivial and clear cut: given that there is a mapping from a larger phase space, to a smaller phase space, meaning that some information is irretrievably lost. It is not there anymore for any super-intelligence to deduce. Just that. This summarizes everything information theory has to say about the issue. Whenever that information is important or not, that is a question of neurobiology.
With regards to the future cryonics, there’s two possibilities:
1: Revival. If we can cryopreserve brain tissue, revive it, and it retains learning, that would be an indication that the procedure works.
2: Fixatives. The opposite of revival. If we find out how exactly the memories are stored, we can develop a fixative mix that would lock those proteins in place by cross linking, i.e. adding strong chemical bonds in place of weak intermolecular bonds. That is a drastic measure which would require pumping the brain full of highly toxic, carcinogenic chemicals such as formaldehyde. (This may even permit room-temperature storage, or may require cooling). Without knowledge of how it is stored one can make a shot in the dark and hope that particular fixatives would work.
Current cryonics is neither, and is hence not taken seriously by experts in any fields expertise in which is actually necessary for evaluating whenever the lost information is relevant. I’ve a nagging suspicion that an effective procedure for future uploading would end with a brain diced into small slices and stored at room temperature in a jar of some cheap cocktail of fixatives. With all molecules neatly cross linked in their original places rather than unravelled and detached by solvents.
Well, I’m mostly interested in brain/mind-preservation in general—I don’t care if it’s by fixatives or freezing. I’ve heard discussion of “plastination” which seems similar to your point #2. Even aside from whether it’s more likely to actually work or not, it seems like it could be cheaper and more practical as well. I’m all in favour of more research along those lines.
(Earlier you gave me the impression that you thought the entire concept of preserving a brain was some wild fantasy not even worth thinking about (ie the typical opinion of Very Serious People), but this seems more like you were disagreeing with the effectiveness of cyroprotectants specifically, which I don’t have strong grounds for an opinion on).
Keep in mind that cryonics generally advertises the possibility of in-place repair of some kind, biological revival. The uploading possibilities are not optimized for. The kind of compounds you would want to add to preserve information (to avoid loss due to denaturation) are very toxic at much lower concentrations.
With regards to plastination, it has the extra destructive steps of trying to get a solid in the end, and to avoid cutting it into pieces.
ie the typical opinion of Very Serious People
I doubt it’s the typical opinion, really. If by very serious people you mean top scientists and the like… they have more complex opinions because due to the training and intelligence they can relatively effortlessly hold complicated relations in the heads. Opinions could be “no future technology will permit revival of [currently] frozen corpses”, “freezing and biological revival is unlikely to ever be workable”, and so on.
And on a tangent, correct opinions about such topics are a matter of knowledge and capability involved in simulating said processes in your head.
To deal with a simpler example without cryoprotectants (e.g. as described here). When a scientist with relevant expertise considers dropping a head into liquid nitrogen, within mere seconds they do a lot of work in their heads—they correctly estimate the cooling rate inside the head (going to take many minutes to freeze), they picture the ice crystals growing, everything other than pure water (up until −18 celsius) getting squished into inter-crystal spaces and getting ripped and scraped in the process (irreversibly losing a lot of information due to many to one transitions—and no redundancy will help you when the ‘redundant’ storage is subject to same destruction), chemical damage due to high salinity (also irreversibly losing information), and so on and so forth.
The rationalists on the other hand seem not to even realize that such considerations are required, let alone occur. It does not matter for how long you are going to think about it if your thought is not even simulating any destructive processes that occur.
With cryoprotectants, the issue is considerably more complicated, with lesser possibility for a simple conclusive disproof, but no better reason to expect it to work (pumping brain full of solvents at denaturing concentrations doesn’t seem like a good idea, and all those references to rabbit kidneys at much lower concentrations mostly serve as evidence of bad faith rather than evidence that it works). The one perhaps big advantage though is that at least the connectivity map would be readable for sure, that’s provided that the cryo-protectants actually do reach most of the brain, which is uncertain as well.
It’s clear at this point that your opinion is not as extreme as I the impression I originally got (unlike some people) and I don’t really disagree much with what you say here. I too am skeptical of current methods (and I’m not signed up for this and other reasons), but I’d like to see further work on both traditional cyropreservation and other methods such as plastination, taking into account any new research on memory formation and storage. The idea being to get to a point where we can preserve an animal brain and check to see that the important information seems to be preserved (even if we can’t read it back out yet).
I am an absolute amateur, but wasn’t vitrification about replacing the ice-crystal-generating water in the brain/body with a liquid that turns into a glass when cooled? If you can get that liquid into the furthest reaches of the brain, wouldn’t you also be able to distribute coolant through its interior, turning the effective cooling surface area proportional to the volume?
In this case cooling speed would be limited by the coolant flow and its thermal capacity and conductivity. You would have to use the cryoprotectant has a coolant. IIUC typical cyroprotectants are not good coolants at that temperature range. Nothing can be a good coolant close to its own glass transition temperature, since by definition their viscosity becomes very high (solid-like) at that temperature.
It’s an exaggeration, but not far off. The information seems pretty damn fragile. From the linked thread: “The damage that is occurring—distortion of membranes, denaturation of proteins (very likely), disruption of signalling pathways. Just changing the exact localization of Ca microdomains within a synapse can wreak havoc, replacing the liquid completely? Not going to work.”
The counterarguments appear to be “but do we really need all that detail for a good-enough copy of the person?” Which is a “prove my negative”—the people arguing that don’t know either.
That argument seems to me to be based on an incredibly oversimplified view of what the recovery process would look like. It’s not going to involve restoring operation to the system.
It’s a double-edged negative… not only do we not know how good the copy will be, we don’t know how good is good enough. (Of course, if our standards for “good enough” are sufficiently low, then they can be satisfied by other people being born.)
Ultimately the cryonics argument is that the value to me of someone who meets my standards for being me existing in the future is so high that any increase, however small, in the chance of that happening has a higher expected value than anything else I could do with the resources consumed by post-mortem cryonic preservation of my brain (or at least, higher EV than many things I am currently doing with them, which I should therefore give up doing in favor of cryonics).
I don’t think that’s a good analogy; IIRC organs (eg rabbit kidneys) have been successfully frozen and revived (good enough to implant), so it’s more a matter of whether that can be extended to human brains (which, sure, may be more delicate) rather than being something inherently absurd.
Rabbit kidneys are much smaller than human brains.
The square-cube law is the main showstopper: you can remove heat form a thing at a rate proportional to its surface area, while its heat capacity is proportional to mass and thus to volume. Therefore, maximum attainable cooling speed decreases with size (if you try to cool any faster, youl’ll just crack the surface).
Rabbit kidneys can be vitrified without using a toxic concentration of cryprotectants, moreover, IIUC the circulatory system of a kidney allows higher flow and pressure (a kidney is just a blood filter, after all), making cryoprotectant perfusion easier. Even then, cryopereservation isn’t perfect: microscopic damage has been observed.
Thanks for the information, but that suggests that preserving a human brain will be difficult and may require more advanced techniques than currently used, not necessarily that it’s some crazy impossible thing that shouldn’t even be thought about. Hell, maybe it would be possible to carefully cut up the brain into smaller chunks before freezing it (a sharp cut along the right line being perhaps not so damaging compared to bad freezing).
A lot of it is going to come down to exactly how memories are stored and how redundant they are. Last time I checked this wasn’t yet fully understood. If they really do depend on fine details of molecules that are inevitably irreversibly scrambled by freezing, then it probably is impossible after all.
I don’t think anybody is claiming that viable brain preservation will be necessarily forever impossible. The claim is that brain preservation as currently offered by cryonics comapanies is probably flawed and unlikely to maintain the relevant aspects of somebody’s personal identity.
Fair enough, though I do think this opinion is sometimes expressed a bit over-confidently given that the physical basis of memory is not yet well understood.
Suppose that I were sceptical that boiling a head in a cooking pot for 2 hours followed by freezing preserved the information… you could say exact same words and they would be equally relevant (or irrelevant).
The “not yet well understood” does not mean it is warranted to plug in some entirely unspecified magic. We know that how-ever it is stored, it must affect transmission of the signals between brain cells. Which leaves us with receptor densities, positioning of receptors, adhesion of molecules to receptors, states of receptors, and the like.
Ultimately, when we do not have an actual reason to believe some procedure works, and only assume it might work from ignorance and introduction of too much magic, it falls to the background of considerations such as “what if donating that amount of money to best charities will make it more likely that the future people will use look-into-the-past wormholes to resurrect you?”. Or all the variations on the theme of living in a simulator (which may well use your brain’s data for something provided it fits some criteria). You need evidence to elevate one such idea above the milliard others.
There’s already enough evidence to locate the hypothesis—eg experiments on animal organs and small animals. Therefore your assertion that this is some crazy idea plucked out of nowhere doesn’t hold.
The other hypotheses were located in similar manner, though. And far too much has to be ignored to generalize from said animal organs or said microscopic animals. Generalization from kidneys to brains is particularly dubious. Especially as great many of said frozen corpses are frozen corpses precisely because of their brain’s unusual fragility with regards to loss of blood supply.
edit: The issue is not with cryonics in principle. The issue is with cryonics as it is. Similar to the people jumping off towers with some wings vs an airplane. Seeing a bird fly really doesn’t make for a case that you can fly by your own muscular power with some homemade birdlike wings, with the early prototype, in fact it makes for the opposite case (as none of the birds are as heavy as you are). Evidence against a hypothesis can locate it too.
edit2: And with regards to information theory, it is absolutely trivial and clear cut: given that there is a mapping from a larger phase space, to a smaller phase space, meaning that some information is irretrievably lost. It is not there anymore for any super-intelligence to deduce. Just that. This summarizes everything information theory has to say about the issue. Whenever that information is important or not, that is a question of neurobiology.
With regards to the future cryonics, there’s two possibilities:
1: Revival. If we can cryopreserve brain tissue, revive it, and it retains learning, that would be an indication that the procedure works.
2: Fixatives. The opposite of revival. If we find out how exactly the memories are stored, we can develop a fixative mix that would lock those proteins in place by cross linking, i.e. adding strong chemical bonds in place of weak intermolecular bonds. That is a drastic measure which would require pumping the brain full of highly toxic, carcinogenic chemicals such as formaldehyde. (This may even permit room-temperature storage, or may require cooling). Without knowledge of how it is stored one can make a shot in the dark and hope that particular fixatives would work.
Current cryonics is neither, and is hence not taken seriously by experts in any fields expertise in which is actually necessary for evaluating whenever the lost information is relevant. I’ve a nagging suspicion that an effective procedure for future uploading would end with a brain diced into small slices and stored at room temperature in a jar of some cheap cocktail of fixatives. With all molecules neatly cross linked in their original places rather than unravelled and detached by solvents.
Well, I’m mostly interested in brain/mind-preservation in general—I don’t care if it’s by fixatives or freezing. I’ve heard discussion of “plastination” which seems similar to your point #2. Even aside from whether it’s more likely to actually work or not, it seems like it could be cheaper and more practical as well. I’m all in favour of more research along those lines.
(Earlier you gave me the impression that you thought the entire concept of preserving a brain was some wild fantasy not even worth thinking about (ie the typical opinion of Very Serious People), but this seems more like you were disagreeing with the effectiveness of cyroprotectants specifically, which I don’t have strong grounds for an opinion on).
Keep in mind that cryonics generally advertises the possibility of in-place repair of some kind, biological revival. The uploading possibilities are not optimized for. The kind of compounds you would want to add to preserve information (to avoid loss due to denaturation) are very toxic at much lower concentrations.
With regards to plastination, it has the extra destructive steps of trying to get a solid in the end, and to avoid cutting it into pieces.
I doubt it’s the typical opinion, really. If by very serious people you mean top scientists and the like… they have more complex opinions because due to the training and intelligence they can relatively effortlessly hold complicated relations in the heads. Opinions could be “no future technology will permit revival of [currently] frozen corpses”, “freezing and biological revival is unlikely to ever be workable”, and so on.
And on a tangent, correct opinions about such topics are a matter of knowledge and capability involved in simulating said processes in your head.
To deal with a simpler example without cryoprotectants (e.g. as described here). When a scientist with relevant expertise considers dropping a head into liquid nitrogen, within mere seconds they do a lot of work in their heads—they correctly estimate the cooling rate inside the head (going to take many minutes to freeze), they picture the ice crystals growing, everything other than pure water (up until −18 celsius) getting squished into inter-crystal spaces and getting ripped and scraped in the process (irreversibly losing a lot of information due to many to one transitions—and no redundancy will help you when the ‘redundant’ storage is subject to same destruction), chemical damage due to high salinity (also irreversibly losing information), and so on and so forth.
The rationalists on the other hand seem not to even realize that such considerations are required, let alone occur. It does not matter for how long you are going to think about it if your thought is not even simulating any destructive processes that occur.
With cryoprotectants, the issue is considerably more complicated, with lesser possibility for a simple conclusive disproof, but no better reason to expect it to work (pumping brain full of solvents at denaturing concentrations doesn’t seem like a good idea, and all those references to rabbit kidneys at much lower concentrations mostly serve as evidence of bad faith rather than evidence that it works). The one perhaps big advantage though is that at least the connectivity map would be readable for sure, that’s provided that the cryo-protectants actually do reach most of the brain, which is uncertain as well.
It’s clear at this point that your opinion is not as extreme as I the impression I originally got (unlike some people) and I don’t really disagree much with what you say here. I too am skeptical of current methods (and I’m not signed up for this and other reasons), but I’d like to see further work on both traditional cyropreservation and other methods such as plastination, taking into account any new research on memory formation and storage. The idea being to get to a point where we can preserve an animal brain and check to see that the important information seems to be preserved (even if we can’t read it back out yet).
I am an absolute amateur, but wasn’t vitrification about replacing the ice-crystal-generating water in the brain/body with a liquid that turns into a glass when cooled? If you can get that liquid into the furthest reaches of the brain, wouldn’t you also be able to distribute coolant through its interior, turning the effective cooling surface area proportional to the volume?
In this case cooling speed would be limited by the coolant flow and its thermal capacity and conductivity. You would have to use the cryoprotectant has a coolant. IIUC typical cyroprotectants are not good coolants at that temperature range. Nothing can be a good coolant close to its own glass transition temperature, since by definition their viscosity becomes very high (solid-like) at that temperature.
It’s an exaggeration, but not far off. The information seems pretty damn fragile. From the linked thread: “The damage that is occurring—distortion of membranes, denaturation of proteins (very likely), disruption of signalling pathways. Just changing the exact localization of Ca microdomains within a synapse can wreak havoc, replacing the liquid completely? Not going to work.”
The counterarguments appear to be “but do we really need all that detail for a good-enough copy of the person?” Which is a “prove my negative”—the people arguing that don’t know either.
That argument seems to me to be based on an incredibly oversimplified view of what the recovery process would look like. It’s not going to involve restoring operation to the system.
It’s a double-edged negative… not only do we not know how good the copy will be, we don’t know how good is good enough. (Of course, if our standards for “good enough” are sufficiently low, then they can be satisfied by other people being born.)
Ultimately the cryonics argument is that the value to me of someone who meets my standards for being me existing in the future is so high that any increase, however small, in the chance of that happening has a higher expected value than anything else I could do with the resources consumed by post-mortem cryonic preservation of my brain (or at least, higher EV than many things I am currently doing with them, which I should therefore give up doing in favor of cryonics).
Quite. We don’t know, so what are the chances?
They don’t need to be very high for cryonics to be an improvement on, y’know, definitely dying.
Cryonics is quite expensive. Success chance has to be non-negligible in order for cyronics to be worth the price.
Depends. What else are you going to spend your money on?
Anything else you like. You can even give it to others while you are alive or after you die.