= 27d567bc2d48d9f40e9ccc20ea370b15)
David_Gerard comments on PSA: Very important policy change at Cryonics Institute - Less Wrong Discussion
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (98)
There isn't so much of a "standard point of view" because the people in the area just really don't take cryonics seriously at all.
Here is a comment from one of the previous threads on the topic, with ensuing discussion.
Not even a superintelligence can restore an ice sculpture from a glass of water.
Yes.
It may be helpful to outline what exactly - in terms of information - makes an ice sculpture irrecoverable or recoverable. It is the fact that distinct ice sculptures will result in precisely identical glass of water. Even if you look at the individual water molecules in the glass and try to retrace their motion backwards, due to the introduction of unknowns (interaction of those molecules with the molecules in the actual glass, then in the air, etc etc), they map to every possible ice sculpture.
The ice sculpture is irrecoverable because the final state corresponds to many possible initial states.
Likewise, massive changes in the solvent - which occur in either cryoprotected or non-cryoprotected cryonics - will force bistable molecules and molecular complexes to transition into a third state, losing their state information. This is because changes in the solvent affect intermolecular forces between parts of a protein (making proteins denature, i.e. unfold or re-fold into a different shape), and between different proteins.
Cryonics as it is can not be seen as science fictional stasis field with cracking and distortion that can in principle be undone someday. It involves massive, many-to-one chemical changes.
It is clear that if the cryonics involved cooking your head in a pot and then freezing it - or even letting the head remain at room temperature for a few hours - the chances would seem fairly minuscule to you, due to extensive many to one chemical changes that would occur during cooking. Likewise, the chances of cryonics - without any cooking - seem fairly minuscule to me due to extensive many to one chemical changes that result from either the introduction of the "cryoprotectants" (at concentrations which denatures some proteins) or due to the concentration of all solutes including salt in the inter crystal boundaries (which also denatures proteins). This is all quite far outside the range of any "robustness" against normal environmental conditions, too - I do not expect memories to be any more delicate than rest of the changeable chemical state (By the way, more chemically 'robust' storage would also require more energy for writing memories).
Now, of course, given the unknowns, we can't tell for sure that cryonics does not work. But we can have no reasonable expectation for cryonics to work better than, say, doing good deeds in the hope that it raises chances at resurrection through some sort of look-into-the-past technology utilizing unknown laws of physics, or resurrection possibilities in simulated worlds, or the like - all the other things that no one can prove impossible.
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.
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.
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).
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.
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).
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.
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.