How To Not Die · 6:18pm Oct 31st, 2013
This post is about the ideas in The Thousand Year Romance Of Clover The Clever. It’s well under 3000 words, so you may as well read it first. Major, major spoilers after the break.
If you’re still here, I assume you’ve finished the story. The core concept is straightforward: what if it were possible to pause the process of death and decay until someone figures out a way to fix the problem? Wouldn’t it be better to explore a strange and wonderful future than to lie back and die?
This is a thing you can actually try.
Cryonics is the process of preserving a legally dead person so that, if and when medical technology reaches the point where it’s possible to reverse the damage, there’s someone left to save. By keeping as much of the brain intact as we can, we’re trying to preserve the information that makes a person a person.
It’s important to understand that this probably won’t work. It’s possible that the freezing process destroys vital information about identity. It’s possible that revivification technology will never be invented. It’s possible that cryonics organizations will fall apart and their patients will thaw. Most people who investigate this think the chance of it actually working is somewhere between 5% and 15%. My own best estimate, taking into account all the things that would have to go right, is closer to 2%.
That sounds low, but if you stop and think, a 2% chance of saving your own life is actually really really good. You wouldn’t balk at spending tens of thousands of dollars for a cancer treatment with a similar chance of saving someone you love. Whereas cryonics coverage (funded mostly through life insurance) costs me under $100 monthly—far less than my health insurance. That’s through Alcor, which is by far the most expensive way to do this. DataPacRat, who commissioned this story, gets coverage through CI and pays about a third of what I do. Either of us would be happy to answer your questions or help you with the process of signing up.
I’ve done my best to tell a story about why it’s good to live, why it’s good to reach the future, why it’s worth making sure the ones you love can be there with you. If you think life is better than the alternative—whether you want to live for fifty more years, or five thousand more, or just one—then take a few minutes to learn more about how to make that more likely.
Yeah, this is the kind of thing I want. It's unfortunate that you have to be well-off to reap the benefits of prolonged life. :/
1469284
'Well-off' can be relative. When I signed up for cryonics, I also happened to stop paying for my basic cable TV package. My overall budget stayed pretty much the same.
1469301
You need to be able to do things like pay... for anything, really. :/ And have extra for extra things like this.
However, knowing it's an option for health care coverage is intriguing. I mean, that puts it within reach, theoretically.
Freeze, petrify, the principle's the same: minimize entropy until someone can cure you.
Interesting way to plug this form of life preservation. One in fifty's pretty good odds.
Cryonics will not save you.
You will find many who will delude themselves and others into believing that there is a means of preserving yourself so that you can be revived long after you're gone. It is a pretty lie, a sweet lie, a lie that they couch in scientific terms. And what could be sweeter than the taste of eternity?
But it is a lie nonetheless.
The truth is that cryonics doesn't work. It is inherently flawed.
"But why is this?" one might ask. "There are living creatures which can indeed survive being frozen, and being revived. Even ones as complicated as frogs! We are not so much more complex than a frog is."
This is deeply misleading, and lies at the heart of why cryonics doesn't work.
The first cold, hard truth is that organisms which "freeze" do not, in fact, freeze. What they do is cheat. They change the chemical composition inside their body to alter the freezing point of water. This allows them to endure sub-zero temperatures (0C, not 0F, though some can indeed survive below 0F) without dying. This is accomplished by their physiology, particularly by increasing the concentration of some molecules in their bodies. These creatures may be at very low temperatures but they are not in fact frozen. This is useless for obvious reasons - while the deterioration and decay of your body will be slow indeed at these temperatures, you are still falling apart, and human bodies are not designed to tolerate such odd chemistry anyway on a very fundamental level.
But there are indeed some animals which do actually freeze... sort of. The problem is that they are not only specially physiologically adapted to do so, but that they STILL cheat - their cells are specially designed so that ice crystals do not form in them, because if ice crystals form inside cells, they will kill them. So again, they keep their cells unfrozen while letting the rest of their body freeze, and have special physiological adaptations which allow this. And, unfortunately, the physiology necessary for much of this is untenable in humans - wood frogs, for instance, cheat by filling themselves up with ridiculously concentrated sugar water - two orders of magnitude more sugar than humans typically possess. Humans have issues with less than one order of magnitude increases in glucose concentration. This is ignoring the fact that, again, these cells are STILL metabolizing extremely slowly, and are designed to deal with extremely low oxygen concentrations.
What about the REAL extremophiles? Some types of single-celled organisms are very difficult to kill, and water bears are infamously invulnerable - surviving such nonsense as the vacuum of space, being boiled, being chilled down to close to absolute zero, pressures greater than are found at the bottom of the Marianas Trench, hilariously high levels of radiation, being almost completely dehydrated... but the problem is, they're not actually invincible, they will die after being exposed to such nasty conditions after on the time scale of minutes to days. Sure, they can survive SOME things for years... but even a decade is far too short, by at least one if not two or three orders of magnitude, and again, like all extremophiles, their bodies are designed to deal with ridiculously stupid stuff like this. Our bodies are not, and even water bears die after a couple weeks of being frozen in liquid nitrogen (LN2). And again, water bears cheat, while the bacteria possess vastly different physiology - humans cannot be dehydrated and rehydrated in the manner that such organisms can be.
The problem, ultimately, is that for proper extreme long-term cryopreservation, one must freeze them at temperatures far below what even water bears can survive for any length of time - we're talking the temperature at which air itself turns to a liquid, in the form of liquid nitrogen, LN2, at the balmy temperature of 77 K - that is to say, 77 degrees above absolute zero.
We can indeed achieve such temperatures. The problem is that we cannot do so in a useful manner for reviving a human being. The trouble is manyfold, but the long, the short, and the middle of it is that humans are not physiologically adapted to be frozen and rethawed. If our cells freeze, they die, as is true of virtually all living cells, but unfortunately the means of preventing our cells from freezing are to either freeze us so quickly that ice crystals don't have time to form - a state known as amorphous ice - or to somehow prevent ice crystals from forming inside cells, which requires us to greatly alter the composition of the fluid inside our cells.
The problem here is that neither works. In the former case, you're dealing with the major barrier of heat transfer. Flash freezing some cells is pretty simple. Flash freezing an entire body, on the other hand, is just outright impossible - it is simply impossible to suck the heat away from a whole human being fast enough to prevent ice from forming. You would need to expose every bit of them - outside and in - to the extremely cold temperature at the same time to prevent ice crystal formation in the places which are furthest away from the cold, and there simply is no way to do this. This isn't a technological problem waiting a solution - it is basic thermodynamics. There is no way around this.
Okay, so if the laws of physics say "No", what about cheating like the frogs do? The problem is that humans are not adapted to such. Our cells are not designed to be dehydrated, which is how many creatures survive being frozen, nor is our body designed to be able to clear intracellular ice which forms, nor is our body designed to support unusual biochemistry in the form of non-aqueous liquids inside our cells. It is indeed possible to preserve a creature in this manner... but the method of preservation will render that creature very dead and cause severe damage long before you freeze them, let alone thawing them out and then changing them back over to normal chemistry.
You may wonder "Why not dehydrate the cells?" Some animals (water bears in particular) do this. However, it requires a physiology and cell structure humans lack. Dehydrating a human to this extent is irreversibly fatal; you will kill their cells.
And all of this is ignoring the fact that anything which does pump all these strange fluids into your body is likely to cause disruption in and of itself, particularly in the brain, which is the most crucial region to preserve (it is what makes you, you, after all) and which is not at all designed to be able to handle such things.
This is just front end stuff. This is completely ignoring the extremely long-term preservation, not to mention the back-end problem that they then have to thaw you back out without damaging your cells, which again has you contending with toxicity and ice crystal formation and all that fun stuff.
Even once you've managed to do all this, your intrepid future scientist now has a cold corpse. Now all they have left to do is the fairly simple procedure of resurrecting a body with no heart beat, brain activity, and organ function. Raising the dead? Piece of cake.
Oh yes, and all of this technology would have to be specially designed for people who tried doing this, who are, naturally, all dead, and the procedure surely would not be inexpensive - given how expensive major surgery is today, and given that this would be far more difficult than putting someone back together after a very bad car crash (something which itself can easily run into hundreds of thousands of dollars), and given that these people can hardly pay you for the decades of research this would all entail, being, you know, dead...
Well, that's a big problem in and of itself, especially when you could spend the time and resources on the already living, and that producing more humans is far cheaper (and a very popular and widely enjoyed activity).
The truth is that the odds of cryonics working is not 15%. It isn't 5%. It isn't 2%. It is 0%. The truth is that it simply will not work. If you want to continue living, your best bet is to spend your life on longevity research, and try to work towards finding a means of extending human lifespan - keeping a human alive is far, far, far easier than ressurrecting the dead.
Cyronics is considered pseudoscience for a reason by most legitimate cryobiologists. The truth is that it simply won't work. Preserving sperm cells and preserving a whole human are completely different ballgames. It is, indeed, magic.
Many people believe magnets are magical because they don't understand them very well. Extremely cold temperatures are the same way, as is most obscure science.
If you want to be preserved like this, you're better off looking for a cockatrice.
EDIT: Note that I am not saying that cryonics is a scam; it (mostly) isn't. That doesn't mean it isn't a waste of money, though.
1470026
> It is 0%.
If you had said 'as close to 0% as makes no significant difference', I would not have disagreed. However, if the thought processes you are using to come to your conclusion actually mean '0%' when you wrote '0%', then I feel it worthwhile to point out that those thought processes don't seem to be applying probability theory quite correctly. There are a number of logical flaws that can be avoided by thinking of probability logarithmically rather than linearly; the unit of that is usually 'decibans', and I've described them, among other places, in the 'CONFIDENCE' section of this link. Put simply - even if everything you say about the difficulties of cryonics is true, that is still /insufficient evidence/ to declare there is a 100% chance that no cryonics procedure, either in use today or that will be developed before a cryonicist's death, will be able to lead to eventual revival. For almost all practical purposes, it's safe to say that it's infeasible to be more than 99.99999999% sure of any real-world proposition.
There's a rather large discussion which could be had about the differences between 95% surety that cryonics won't work, which is my general position; 98%, which is Benman's; 99.9%, which I've seen proposed elsewhere; and all the various levels between there and 99.99999999%. (There are even people who think there's a 90-ish percent chance that cryonics /will/ work.) There is another large discussion about whether cryonics is a worthwhile investment at any given confidence level and price. If you'd like to engage in any such discussion, or a related one, I'd welcome it.
As a starting point, an important question to ask is: in the long term, in general, is science and technology going to continue to progress, to the point of accomplishing some computational feats we would today consider effectively impossible? If so, then once frozen to a temperature low enough that chemistry has stopped, then it matters little whether it takes 30 years or 300 until techniques are developed which can reconstruct a human mind out of the damaged cells produced by any given cryo-preservation technique. If you do not have a solution for, say, the halting problem, then how reliable can an estimate of 100% certainty about the impossibility of such techniques ever being developed at any point in the future be?
Taking a step back: If you choose to look for reasons why something is impossible, you will find them. If you choose to look for reasons why something is possible, you will find them. Performing one of those without at least performing the other can leave you in a state where you know less than you knew before.
There's also a second important consideration - even if cryonics causes ice crystals to shatter cell membranes, the real question is whether we can't develop technology sufficient to repair that damage; if you can do nano-repair, then even that cold damage may possibly be no more than an inconvenience. It becomes a matter of whether it completely wipes the platters, so to speak, or merely damage the hard disk so we can't read it without some extra bit of technology.
1470026
I am familiar with the arguments for why cryonics probably won't work. They're reflected in my estimate.
Agreed. This is my Plan A. I do the obvious things—diet, exercise, avoiding cars—all of which have a higher expected value than cryonics. I'm doing work that I expect will accelerate longevity research (among other things). These strategies are not exclusive.
1470073
It matters a lot! I think Alcor is very likely to last for 30 years, but very unlikely to last for 300.
1470073
What are your odds of surviving being decapitated?
If you say something other than 0% you're just being pedantic.
You can avoid a lot of logical errors by recognizing this, and recognize a lot of lies couched in made up numbers by doing so.
99.99999% is a meaningless number in cases like this, and is a wonderful example of false precision. Saying the odds are 0% is actually far more accurate. Any time you see a number like 99.99999%, you can be sure that either:
1) The experiment has been repeated a truly ridiculous number of times.
2) It is a made up number.
Given the situation, it is quite fair to say the odds are 0% - indeed, there is no statistically significant difference between 0.00001% and 0%, given the likely size of the error bars.
I suspect that any reasonable calculation on the odds of cryonics working is going to include 0% well within its error bars; there probably isn't even a statistically significant difference between 0 and whatever probability they claim.
Saying 0% is wrong is, I think, quite silly and is entirely within reason.
Naively multiplying extremely uncertain numbers is not a very good way to go about assigning probabilities to things, and when your overall number is smaller than your error bars are, that's something of a problem.
As for the question about computation: the answer is that the exponential rate of increase of computation speeds is already declining, and the maximum computational speed is within three orders of magnitude of the present-day maximum speed. People with even a rudimentary understanding of exponential numbers in nature will tell you that all exponential growth in nature is self-limiting, and typically occurs very early on. The overall rate of increase of technology is actually lower today than it was 50 years ago, which surprises a lot of people; cars are not exponentially better than they were in 1970, nor are planes, nor are desks or houses or nuclear power plants or any number of other things. Only computers have really improved exponentially and that is because they are new and because the means of improving them was very linear - further miniaturization. Thus you see computers and things which are directly influenced by computational speeds improving rapidly while most other technology does not do so.
The issue isn't even one of repair, it is knowing what to repair in the first place. If the data is destroyed, you simply cannot recreate it save by chance. Building a brain may or may not be possible in the future, but building a particular brain is only possible if you have a map of it or if you happen to luck across it - extremely long and bad odds. Even 1% loss represents unacceptably extreme loss.
Technology and science are not magic, even if many people who understand them poorly believe them to be so.
1470162
Nanobots in the way you are thinking of simply don't exist; they violate the laws of physics. Nanobots are more like viruses or enzymes than little machines.
More to the point, however, you cannot repair something you don't know the "right way" of going in the first place. Destruction of connections in the brain is death; even if you can hook them up, unless you know where each of them went, you have nothing.
Of course, there is also the question as to whether such a revived, rebuilt being is indeed you in any sort of meaningful way in the first place. If I can simply grow an exact duplicate of a human being, I can grow any number of them, but none of them are truly "You".
1470180
Are you a biomedical engineer, a biologist, or otherwise directly working in a field which does such? Or gathering money in order to fund your own research into such?
1470213
> there is no statistically significant difference between 0.00001% and 0%
This is why I recommended logarithmic probability. To achieve 90% certainty requires 10 decibans of evidence; by LaPlace's Sunrise formula, this can be achieved with, say, 10 successful tests. 99% certainty corresponds to 20 decibans, requiring 100 such tests. 99.9% certainty, 30 decibans, 1,000 tests. 99.99%, 40, 10,000. 99.999%, 50, 100,000. 99.99999999%, 100 decibans, 10,000,000,000 tests. There are only around 10 billion humans alive; at least one of whom can safely be assumed to be in a catatonic state experiencing a universe that has absolutely nothing to do with objective reality. It's essentially impossible to prove that you're /not/ in such a state, meaning the upper limit of certainty is on the order of 99.99999999% - even if you could make enough tests to achieve that level of certainty, which is rather hard to manage in itself.
To achieve 100% certainty - which is equivalent to 0%, just for the negative - is equivalent to an infinite number of decibans, and would require an infinite number of successful tests to achieve. While you may consider this to be 'pedantic', the difference between 'any positive integer' and 'infinity' are rather important, mathematically; and those differences do, in fact, have real-world consequences.
> none of them are truly "You".
From my quotefile:
"But I am not an object. I am not a noun, I am an adjective. I am the way matter behaves when it is organized in a John K Clark-ish way. At the present time only one chunk of matter in the universe behaves that way; someday that could change." -- John K Clark
I've already started working out some of the details involved in having multiple copies of oneself running around. One version of them is at the Orion's Arm site, in the form of the Dividual Interaction Protocols and Dividual Naming Schema.
1470183
Fair enough. The more time before reanimation tech is invented, the lower the odds of any particular corpsicle being preserved until then.
Which is as good a reason as any to promote a maximal level of technological development, by promoting the Enlightenment-derived ideals of science, democracy, and competition-based capitalism, in the face of oligarchical attempts to bend the rules in their own favor... but that's a rather different discussion.
1470260
False. Zero can be safely assumed to do so. Indeed, I am unaware of any such cases. This makes your entire argument disintegrate.
In any case, logarithmic probability gives you exactly the same answers as any other form of probability when performed properly. If you get a different answer, it means you're doing it wrong. It is simply a different means of representing the result. If someone told you something else, they lied or were misinformed, or you misunderstood, or they were deliberately misdirecting you.
It being impossible to prove a negative is utterly irrelevant in any case. It is impossible to disprove we are not all in a perfect simulation, but there are very, very good reasons to believe it is not so, and if it is so, it makes no difference, so the argument is fallacious to begin with.
1470268
Who says democracy and free-market capitalism are optimized for such purposes? There's really no evidence to suggest that it is so.
1470026 As it happens, I've read all the papers I could find on the physiology of wood frogs and their freeze response as part of my (unsuccessful) efforts with Ken Storey to get their genome sequenced. I don't remember much of it now, but...
"Their cells are specially designed" means their cells have a lot of mechanisms for surviving the process. Obviously, we have to enumerate those mechanisms and then engineer them into our own cells.
There are, if I recall, 6 known key parts of the freeze response:
- Production of cryoprotectants, including glucose, but proteins as well, which prevent formation of ice crystals or encourage them to form in shapes other than the usual long stabby ones. See for instance "Evolution of an antifreeze glocoprotein", Nature, 30 Sep 1999, on a protein used by arctic fish.
- Bacteria with ice-nucleating proteins that encourage some ice formation at higher temperatures. This is because rapid freezing (as I think cryonics facilities do) destroys tissue, while slow freezing causes less damage. With bacteria we usually aim to lower temperature by 1 degree celsius per minute to keep them alive.
- Production of proteins to protect from the cryoprotectants.
- Protection from hypoxia (lack of oxygen).
- Protection from dehydration (lack of water).
- General activation of stress-response proteins.
Production of cryoprotectants, and of the proteins needed to protect from hypoxia, and from dehydration, and of those needed to protect from the toxic cryoprotectants, all happens rapidly and at nearly the same time in the wood frog! There's no reason to think we can't re-engineer our cells to do the same. It's a massive gene therapy project well beyond anything that's been achieved, or even contemplated in any journal articles--dozens of genes regulated to activate in a complicated sequence, and then another sequence for thawing. But it's certainly possible.
(It's not possible under the current approach to medical research, which would have us submit a separate grant proposal to study each protein, and then futz around forever studying their kinetics independently without understanding the math needed to engineer a similar regulation system because the NIH won't give grants to mathematicians.)
Also much more sugar than wood frogs typically possess; more than enough to kill them in ordinary circumstances. The production of glucose as a cryoprotectant is part of the freeze response; the production of other proteins to protect cells from glucose is also part of it. I tried to get the NIH interested in this under cover of developing a treatment for diabetes.
Extremely slowly. Possibly slowly enough. I don't think we have any data on how rapidly they are metabolizing. They don't eat all winter.
This is what everyone is trying to do, and it is, I think, wrong. But it may not be necessary. We might not be able to freeze someone for 100 years, but 20 may be possible. I have seen data on metabolic rate in cells at different temperatures. I can't recall the results, but I know I was interested in the question of whether humans could be kept at 0C for a very long time, and my conclusion was that it was not obvious whether 10-20 years would be feasible or infeasible.
If one can thaw, eat, and refreeze yearly, as the frog does, then 1 year is enough. Lower the temperature to just below freezing, as the frog does.
Notice, though, that in cells banks we routinely freeze and thaw bacteria in liquid nitrogen which don't survive freezing at all in the wild, by using cryoprotectants and freezing them slowly, at 1C/minute. Your argument that freezing and thawing a cell not adapted to freezing and thawing is impossible implies this is impossible.
I agree that current work in cryonics is headed down a dead end. They are taking a simple approach and hoping to optimize it without acknowledging how terribly difficult than optimization will be, rather than taking the horrendously complex approach used by nature that is workable but requires decades of upfront research before trying to cryosuspend anyone. I suspect cash flow is the reason.
1470401
You are indeed correct about the physiology of wood frogs (though they do avoid allowing ice crystals to form inside their cells, with the nucleation sites being located outside of them; indeed, it is worth remembering that part of the issue with ice crystals is simple expansion - expanding between cells allows them to push cells around, but expanding inside cells, there's nowhere to go but out. Not that shape isn't important, but it is not the only important factor), but humans lack many of these adaptations. With genetic engineering, it may well be possible to create humans who could be frozen and unfrozen like frogs, but I find it fairly unlikely that a genetically normal human can be frozen and unfrozen so easily - it seems like a difficult process, and frogs are endothermic, which is another important distinction.
The other issue with freezing and unfreezing people repeatedly is that to be eligible for cyropreservation, you must already be legally dead. Even assuming we relaxed this restriction somewhat, I strongly suspect that unhealthy wood frogs are far less likely to survive winter than healthy ones, so the people we would most want to cryopreserve in this way are also the ones who are least likely to survive the process.
Not that there aren't other potential uses for such (interplanetary travel, say), but I don't think saving lives is really a big one. I have to imagine that even if we could do it, it would be rough on your body.
1470026
A relevant passage from "Reptile freeze tolerance: Metabolism and gene expression" (Kenneth Storey, Cryobiology 52: 1-16, 2006):
So metabolism at freezing is going to be much less than 10% of normal resting metabolic rate. I wish I could find a reference on measurements at freezing; I thought I had one. 10% could suffice in some cases.
1470260>>1470356
TD, you're being kind of a jerk here. DPR made perfectly valid points, and his entire argument does not disintegrate because you don't believe people are conscious when in comas or asleep. You are smarter than your answer here is. It is absurd and supremely arrogant for you to claim 99.999999% certainty in this, which is what DPR is rightly objecting to.
I was going to say that you could maintain your position in a way that didn't come across quite so arrogantly, but on reflection, claiming 100% certainty in a way that doesn't come across as arrogant could only be a very skillful deception.
DPR used the phrase 99.999999% to show how absurd what you were saying was, and you then turned it around to say that he was absurd for using a number so close to zero. Well, he was just being accommodating, rather than telling you straight off that it was absurd for you to claim even 99% certainty. And when you say such extreme numbers are meaningless, it only shows you haven't dealt with situations in which such numbers come up, or that you can't think about large numbers, or that you don't believe in math. All DPR is asking is for you to say you believe in math, and you respond with philosophy.
I regularly ran a computer program which required over 99.99999999% accuracy in retrieving bits from a disk drive. I had about three crashes per year due to bits being read inaccurately from the disk, even given that level of accuracy.
1470468
I'm not claiming 99.99999% certainty. Indeed, I specifically criticized such in my post.
I think part of the trouble here is notation. If I say something with 99.99999% confidence, I'm claiming 7 sig figs of accuracy; the proper notation to indicate this confidence level would be P<1*10^(-7) of success (which would represent only a single digit of confidence). With only one sig fig, 1 - 1*10^-7 = 1, not .9999999.
When I say 0%, I am saying "I don't think there is any significant evidence that this is going to work." The null hypothesis is "people who have been frozen solid are irrevocably dead." Cyronics claims that some of them can possibly be revived. I see no evidence for this, thus 0% - the odds of it working have never been proven to diverge from the null hypothesis, so why should I claim they are higher? They are made up numbers which don't really have any basis in reality. To my knowledge, no one has ever been revived, and indeed no one thinks that anyone could presently be revived from this state, so my statement is, I think, quite reasonable. If you took a frozen solid body to a doctor, they would tell you that they were dead and there was nothing they could do.
The argument cyronics advocates are making is that at some point in the future this would not be true, particularly if you were frozen in some specialized manner, but there's no actual evidence of this - it is a purely faith-based claim. Thus, 0%. In reality, the odds are obviously to some extent unknown, but on the other hand I have some good reasons to suspect, from my understanding of biology, that cryonics won't work on a human, and that, in the present day, any investment in such is essentially a very elaborate form of burial, like having your remains sculpted into a throne of bone (my brother's wish), a diamond, or being mummified - it is a hope and a wish rather than a reality that it will be successful.
0% may sound cynical, but from my standpoint, I think it is a reasonable evaluation of the situation, as there is no evidence for it, and a considerable amount of evidence against it. If I was writing a scientific paper, I wouldn't even assign a probability estimate to it, because it would be me making up numbers.
1470260 1470499 One of DPR's responses invoked logarithms because he spelled out the fact that you cannot have enough evidence to claim such confidence--if you spent 8 hours a day observing direct evidence that always indicated that cryonics cannot work, at 2 bits per second (based on some psychology experiments), for 10 years, that would give you only 1.7 million bits of information, and a little over 99.9999% certainty. 0% is therefore not a reasonable evaluation. He's just asking you to admit that you believe in math. The fact that we can't assign a reasonable estimate does not make 0 a reasonable estimate.
1470514
I don't think you're understanding what I'm saying.
I'm not claiming 99.99999% accuracy and I never did. That's just a straw man argument. That's what I was pointing out in my post.
Significant figures are the issue here. 0≤0.1±0.1. Or, not in math speak, zero is not, in fact, less than 0.000001 when your accuracy is limited. That was why I gave the example of 1-1*10^-7 = 1, because, while in "pure math" the answer would be 0.9999999, in science (and thus, in real life) the answer is, in fact, 1, because there is only a single significant digit.
1470537 (1470268)
If you approach real-life problems that way, you will get the wrong answer. For example, I recently developed a gene caller which constructs a Markov model for codon triplets. There are 64^3 = 262,144 possible sequences of three codons. But some sequences are very uncommon in genomes. I wrote one version of the gene-caller in which triplets that did not occur in my training set, and therefore had a probability of less than 0.00000001, were assigned a probability of zero. This was disastrous. It caused my program to miss genes. The reason is that there were many such triplets, and each new genome would have some genes containing one of those triplets. The probability that the sequence was a real gene was computed as the product probabilities produced by each triplet. Finding just one triplet that was assigned a probability of zero on my model cause loss of that gene, regardless of all the hundreds of other probabilities factored into that gene's score.
I didn't have nearly enough accuracy to distinguish between 0.00000001 and 0; in fact, I could probably only claim about three digits of accuracy. Nonetheless, it was critical for me to assign every possible conceivable outcome a non-zero probability. It DID NOT MATTER much at all what probability I assigned such cases; I tested another version where I accidentally assigned them a probability near one. That gave much better results than assigning them probability zero. The reason is that assigning ANYTHING a probability of zero makes that one data point annihilate all other evidence. When you use enough evidence, this inevitably leads to disaster.
1470587
Inevitably is a strong word. There are indeed many things to which you can assign a probability of zero and it will never have any influence at all.
However, more to the point, you are comparing two entirely different things.
In an actual scientific paper, what you say is "below detection limit". In other words, if I'm running some sort of elemental analysis with GCMS or something similar, and the machine fails to detect any, say, uranium, I don't say in my paper "There was no uranium in the sample" but "There was no uranium detected in the sample ([ U ] < 0.5 ppm, P=.99)" or whatever. If I'm giving someone a chart, it will be labelled as [ U ] < 0.5 ppm, indicating that there was less than that amount of uranium inside the sample. This is due to detection limits. If I took a large number of ink samples from Hewlett Packard, I would not find uranium in any of them. That doesn't mean uranium doesn't exist, and not even that there isn't any uranium in any of the ink, but that the amount of uranium in the samples is below my detection limit.
Obviously when you're dealing with real world observations, what you do is say that it is below the detection limit and give your confidence level of said observation.
More to the point, however, you're also talking about something which is behaving deterministically - an expert program - which is also rather different for other reasons, such as the fact that it runs on pure math.
This is completely different from what is going on here, and you know it.
We're not dealing with observations here. We're dealing with an estimation of a future event, based on a history of 0 prior events (because they have narrowed the pool down to the point where they claim there are no priors, because no one has been frozen and then revived with magical future technology). This is purely faith based, as your conclusion is based entirely on the assumptions that you make. Some of the estimates can be based on real world data (for instance, the odds of a nonprofit folding in any particular year can be calculated and extrapolated out) but many of the critical data points have no real data on them or are otherwise purely negative.
If we look at it less charitably, we can say that of the people who have been frozen solid due to exposure in cold climates over the millenia of human existence that there is not a single reputable case of someone being frozen solid, being drug into a warm place, and after they thawed out they were revived. This is not exactly the same thing, but it gives us a bound of probably P < 0.000001, and even with modern technology, no one even bothers to try unthawing said people. Heck, no one bothers trying to unthaw the people who have been frozen by Alcor, because they agree that they would be dead. There is no way to do that presently and be successful.
Thus, they invoke magical future technology, per Arthur C Clarke.
Thus your estimation of the probability of this even being possible is based solely on what you want to be true, with at best a nod to science at and worst is purely hopeful thinking. There is no evidence to suggest that we can actually successfully freeze and thaw a human brain in LN2 and have it remain intact on a level which would allow it to regain function.
Given the extremely large error bars on any of these observations, and that many of the error bars certainly touch 0, and that it only takes one to touch 0 for it to fail, I don't think it is unreasonable to say that P=0.0 is a valid estimate.
1470665 Both your case regarding cryonics, and my case involving codons in genes, are cases where we want to assign a probability to the occurrence of an event that has never occurred. My example shows that if you assign a probability of zero to an event just because you've never seen any evidence that it can ever happen, this will have very bad results.
If you have a hypothesis X, with a thousand different pieces of conditional evidence P(X|Z), and just one P(X|Z) was set to zero, then it doesn't matter what all the other values are (unless they are 1, which is also a mistake for similar reasons).
Titanium Dragon, perhaps a thought exercise might help. Imagine I was standing before a massive pile of beads, some blue, some white (this is mainly to limit the exercise to those two colors); and that I was holding a bag. I take some number of those beads, though you can't see how many are which color, and put them in the bag. You start drawing beads from the bag, one at a time, and checking their color. The first is white; the first ten are white; the first hundred are white; and so on. After how many white beads does the odds of the next bead being blue become zero?
What makes you think that the technology being predicted and/or hoped for breaks the laws of physics?
1470213
I'm working on intelligence amplification and organizational design to improve research in general. Taking into account replaceability and comparative advantage, this seems like my best path.
1470268
I think it's a valuable discussion! I am curious whether you think those particular ideals are the best ways to promote technological development. Your examples are pretty well-accepted. I'd expect the returns of pushing them yet further to be positive, but lower than the returns of pushing, say, effective altruism or transhumanism or something else that isn't (yet) mainstream.
1471321
Thing is, though, assinging non-zero probabilities to events with zero probability is incorrect, as any one probability of 0 means the probability of anything dependent on it is 0, and in any case, you missed my point. They aren't the same thing at all.
One would expect, with a large enough sample size, all codon triplets to be represented because of the nature of the system. There is no particular reason to believe that all technologies which we can imagine are possible; indeed, there are many possible-to-imagine technologies which violate various laws of thermodynamics.
1471336
Because it magically reconstructs information which was not preserved, mostly.
1471336
The correct analogy is this:
There are an unknown number of beads in a heap of unknown size, you have no idea what color (or colors) they might be, and you are completely blind. You have a machine which tells you which color the beads are when you run them under the machine one by one.
If you are wise, after you run the first bead, you will guess the same bead is the same color as the first, as you had no prior information, and you will continue to do so so long as you pull only one color of bead from the pile, and grow increasingly confident in your guess as you do so.
You are making the crucial flaw of assuming there are any other colors of bead in the pile with no prior evidence.
1471973
One thing I've learned from certain aspects of mind-numbing political discussions, is that if a theoretically perfect system gives perfect results (such as 'maximizing scientific progress'), then a middlingly good system which comes one notch closer to such perfection should produce results that are one notch better than any competing system. Taking this principle the other way around, and examining which real-world systems do better than others, you can make educated guesses about what the theoretically perfect system might be, and thus what nudges to existing systems might come closer to that ideal. Over the long term, it seems to be that the most reliable rule-of-thumb here is that political systems which provide the most competitive marketplace of ideas, in as many ways as possible (academic, economic, religious, political, and more) seem to consistently win such comparisons.
If a suggested nudge matches that rule of thumb, then whether it's called transhumanism or a futures prediction market or something else entirely, I'd guess that it would be an improvement, and I would generally support it.
1472277
How much information do you think our brains have in common? If there's enough commonality, a HumanOS if you will, then any information in a brain that's destroyed that's part of that HumanOS can be reconstructed from a source other than the damaged brain. How much of a brain's structure is genetically determined? Again, such information can, at least in theory, be reconstructed, even without any reference to whatever remains of the cryo-preserved brain. Even if those portions and structures of the brain receive 100% damage, and not a single molecule of their original structures remain intact, then, at least in theory, it might not be impossible to re-create them anyway, from scratch.
Unlike some other cryonicists, I've made arrangements in my will for as much information about myself as possible to be placed onto a durable storage medium and stored along with my body. Primarily, this will include having all my digital storage media, including some decades-old floppies and hard-drives, sent to a data-retrieval company. This is, as you might guess by now, specifically to provide a source of data for possible brain reconstruction, even if my brain itself has been significantly damaged.
Even assuming the eventual development of seemingly-near-magical nano-scale technologies, I'm doing what I can to provide as many avenues for potential reconstruction of my neural networks, my self, as possible. Should I come up with any more ideas before I de-animate, I'll try adding them to the mix.
After posting my previous thought experiment, I realized that it actually doesn't matter how many colors are involved, as long as you define a 'successful test' as 'bead being white', and 'failed test' as 'bead being not white', and you keep drawing out white beads.
I would strongly recommend that you read the Wikipedia article on LaPlace's Sunrise Formula, also called the Rule of Succession. (The 'sunrise' name comes from asking the question, "Knowing only how many times the sun has risen, what's the probability it will rise again?") The rule LaPlace came up with for such probability estimates is (Number of successful tests, plus one) divided by (Total number of tests, plus two). That is:
Probability Of Next Success = (PreviousSuccesses + 1) / (TotalTests + 2).
For the example at hand, once you've pulled out zero beads, then you've had zero tests and zero successes, which leads to (0+1)/(0+2) = 1/2 = 50% odds. Which corresponds to intuition - if you don't have very much information at all, you might as well treat success and failure as equally likely. After the first white bead, the estimate for the next bead becomes (1+1)/(1+2) = 2/3 = 66% odds. Which matches your own description - after the first white bead, it's more likely than not that the next bead will be white. After ten trials, we get (10+1)/(10+2) = 11/12, or 91.6% odds that the next bead will be white, which, again, seems about right - after that long a streak of white beads, it's much more likely than not that the next bead will be white. Let's jump ahead: after ten billion white beads, the odds become (1,000,000,000+1)/(1,000,000,000+2) = 10,000,000,001/10,000,000,002 = 99.9999999% odds that the next bead is white.
After any finite number of such tests - even one every Planck time-unit ever since the big bang - there will be a finite number of 9's to the final figure. In other words, in the physical universe, it is infeasible to reach sufficient evidence to achieve 100% certainty of anything. (Or 0%.) You can get sufficiently high levels of evidence that to treat the remaining probability as having any significant meaning would be insane and perverse; but that remainder (usually just called 'epsilon' once it gets small enough) still hangs on, never quite disappearing.
But it begs the question: do you want to wake up in a changed world without the people you once knew and loved?
1473334
I would be even more happy if the people I once knew and loved also decided to sign up (though I know better than to try pushing them any more forcefully than by doing so myself, and being willing to answer any questions). But, however sad it will be if they don't, I feel that waking up in a changed world is extremely likely to be better than never waking up at all. (I'm one of those people who think that the odds of any form of supernatural afterlife, or non-physical soul, existing, are so small that they can be relegated to the minuscule depths of 'epsilon'.) Taking the long view, the year 2000 was generally better all around than 1900, which was better than 1500, which was better than 3000 BC, which was better than 20,000 BC. There doesn't seem to be any significant evidence that this trend is going to stop, even if there are occasional and temporary dips and setbacks, as long as humanity itself manages to keep going. The possibility of X-Risks, such as unfriendly AIs, seems to be the main source of potential unpleasant futures; but simply because the future /might/ be bad doesn't mean people should make their plans as if that was the /only/ possibility, or even the most likely.
Somewhat relatedly - it's entirely possible that I have absolutely no choice in the matter. There's a philosophical proposition, called Everett immortality; it's based on the multiple-world interpretation of quantum physics. Trying to phrase it succinctly, it states that I will never experience a timeline in which I'm dead; and that starting from almost any moment in which I'm alive, there's at least one future timeline in which I continue to be alive, even if it requires an incredible series of coincidences to accomplish. Cryonics opens up a broad swathe of timelines for me to experience that I otherwise would never get to - it may, in fact, be the 'incredible coincidence' which allows my future self to keep on living. ... Not that I'm ever going to actually /rely/ on such an untestable philosophical proposition, if it can at all be helped, of course.
(And I do believe that I've finished dropping my daily quota of LessWrong buzzwords... </injoke>)
1473424
That certainly is a very interesting way of looking at it, definitely something I would be curious of now when I am young and full of curiosity. But I don't believe I would feel the same way when I am old and ready for the sweet embrace of oblivion. No, I don't believe in an afterlife, I think that the complete ending of death is something as infinitely beguiling as it is terrifying, and have my own animistic ideas about what happens to that lump of crazy energy you may or may not call a "soul" after death that don't involve personal experience any more than they involve a true ending so much as a continuation of something better than the life of the individual in themselves.
But though it would be a fascinating intrigue to freeze myself now and wonder in those last living moments of cold what world I will wake up to to live out the rest of my days in, I would be missing the wonder of that which is given to me in the here and the now: Life goes on, and I will race it to the end, not sit by in my little freezer and watch without seeing or sensing through the millenia.
Besides, in the way you measure things, things may be getting "better", but I'm certainly not you and I feel that the shift of time is not something that gets better or worse in summation of all that happens. it is the individual for which life can be better or worse, and I would personally be more suited to a time to which I do not currently belong. That said, I've had a head start on the present and won't gamble on the future.
I plan on living now, not dying later.
1473507
An approach that's at least as good as any I've seen, and better than most. If I may offer an ever-so-slight rephrasing that's closer to my own approach:
I also plan on living now, and on not dying later. If possible.
1472277
Overnight, I've thought of an alternate approach which may help bridge our gap in communication.
Area isn't a fundamental quantity in and of itself; it's a derived measure, the result of a formula and the numbers (lengths) which are inputted into that formula.
Confidence, level of belief, trust, the odds, or whatever you want to call it, is /also/ not a fundamental quantity. It's also derived, resulting from a formula; and the inputs into that formula are the pieces of evidence, and the amounts of evidence. Putting error bars on confidence levels which include 0% confidence is about as meaningful as putting error bars on latitude which include 91 degrees North. The place where the error bars should be put, in order to keep the output sensical, is on the /input/ to the formula rather than its /output/. If you give, for example, an upper and lower bound for the amount of evidence available, then you can easily derive the upper and lower bound for the resulting confidence level - which will not include 0%. (Unless there's an infinite amount of evidence, which we can safely assume isn't going to happen any time soon.)
If this, or any of my other posts on the topic, manage to communicate to you the idea that I'm trying to express... then a corollary of this idea is that a very useful quantity to derive is how big 'epsilon' is. That is, figuring out that you're 99.99% sure of something and have an epsilon of 0.01%, or if you're 99.999 999% sure, and have an epsilon of 0.000 001%, can be an important point. In addition, knowing how much evidence you have, and of what quality, can give you at least an order-of-magnitude estimation of the maximum level of quantity you can have with that amount of evidence - and thus what the minimum epsilon can be, for that amount of evidence. Depending on the amount of evidence available for a proposition, it can make a big difference whether epsilon is 0.1%, or 2%, or 5%, or 20%. But without being able to do at least a rough Feynman estimation of the evidence, it can be hard to pin down even a rough order-of-magnitude of what epsilon might be... and if 'epsilon' corresponds to 'and I just might eventually get revived', knowing whether it's closer to 5% or 0.05% makes a significant difference about where to spend one's life-extension budget.
Generally relatedly, you might find this GoogleDoc worth perusing, as it contains some estimates made by various interested people about how likely they think it is that cryonics can work. Some have done significant research; some less so. Would you be willing to offer your own estimates of a similar sequence of probabilities?
1473555 If meds get good enough in future
1472361
There's no such thing as a "theoretically perfect system", and if you hear people talking about it, you know their opinions are beyond worthless.
There's really no particular reason to believe our prsent system is anywhere even close to optimal, and many good reasons to think otherwise.
Democracy is actually probably very bad for actual efficiency.
Firstly, this is completely the wrong question. This is what is known as misdirection.
The correct question is "How much information do you think our brains DON'T have in common?"
The answer to which is "A ridiculously massive amount". Our brains are not wired the same way throughout on a person-to-person basis, especially not in the part that makes us "us".
Indeed, forming new memories physically changes your brain.
That bodes very poorly for your hopes and dreams.
The gross structure of the brain, and some portions of it, are genetically determined, but the vast majority of what makes us us is, again, NOT genetic. Most human behavior is learned, not instinctive. Knowledge is learned, not instinctive. The list goes on.
There is absolutely no evidence whatsoever for this assertion, and indeed considerable evidence to the contrary.
I will also note that anything which does not involve reanimation of the same matter is just making a duplicate of you, meaning that you are, personally, still dead.
YOu will hear people lie about this, but it is not at all correct. There is indeed some meaning to continuity of existence, as noted in comics like, say, Freefall - making a backup of yourself is not the same as personal immortality.
Even still, though, this won't actaully happen.
I'm afraid this reveals a very basic lack of understanding of probability and statistics on your part. Laplace used the sunrise as an example of something you could calculate the odds of using his method, but if you actually read what he had to say, he noted that using it on the Sun is absurd and incorrect - the number you get is wrong because, quite simply put, the sun rising is not a statistical process and you have other information on its probability. Barring a large energy input into the system which would be readily apprehended, the odds of the sun rising are not calculated via his formula, but is, rather, 1 - it will rise, because it would be a violation of conservation of angular momentum if it didn't. A large energy input could change this but, again, it wouldn't be a statistical process.
The rule of succession is only useful in situations where you lack other information.
In this case, we have other information, so the rule of succession does not apply.
The idea of unfriendly AIs is pure nonsense in the first place. Not that you couldn't build an AI that didn't like people, but the idea that AIs will actually be all that smart (or common) in the first place. An unfriendly AI is no more dangerous than an unfriendly human with a computer, and is actually probably considerably less dangerous because it is far more vulnerable.
There's actually two major problems here:
1) The many-worlds interpretation of quantum mechanics is incorrect.
2) Everett immortality is an incorrect understanding of it anyway, as P=1 for many events.
1474413
You don't understand what fundamental quantities are, and you don't understand that fundamental quantities are interchangable.
It would be trivial to construct a system where area was what determined your fundamental dimension, and you defined all other dimensions relative to area. You could do it with volume instead. There's no reason why "length" is special; you can convert between area, volume, and length in various ways to measure things.
The units we use for measurement do not change what is.
Incorrect. P=0 and P=1 are both valid probabilities, and indeed, there are an enormous quantity of P=0 configurations - indeed, far more P=0 configurations than possible ones. Moreover, 91 degrees north is a reasonable number to get in some cases, depending on what you were calculating - it would most likely represent that whatever you had done had crossed over the north pole and started heading back down the other side of the planet while doing some sort of distance or area calculation. If I'm at 88 degrees north, and I pick a direction based on current north and fly 333 km, given that each degree of latitude is close to 111 km, saying I went to 91 degrees north isn't actually all that silly, even though my actual latitude is 89 north - it means that I have crossed over the North Pole and am actually heading south now.
The statements "Everything that can happen, will happen" and that "everything will happen" are not logically equivalent, because "Everything that can happen" is a non-inclusive subset of "everything".
I've read a number of estimates.
Anyway, as far as all the stuff goes:
Everyone is hopelessly optimistic about the odds of dying somewhere where they can be frozen. The odds of dying in such a circumstance are actually quite high, and are only getting higher over time; only 29% of people die in hospitals these days, and that number is on the decline, not on the rise. Your odds of dying somewhere where you can't be frozen quickly will likely be something close to 3 in 4.
Likewise, dying when your brain has degraded is growing increasingly likely as well as we eliminate other causes of death.
The dying in a hospital which refuses access or dying and your relatives refusing to freeze you are mostly personal odds, but I'd guess for most such people the odds are quite small.
Laws against cyronics are incredibly unlikely, but I think over time brain death and legal death will become more and more the same thing, which makes cryonics increasingly useless. There's no reason not to allow people to freeze people.
Not all of what makes you, you is encoded in the physical state of your brain is 0, as you can have a flatline EEG and recover.
However, whether all of it can be preserved when frozen is a more open question, and assigning even 5% confidence to the idea that it can be (as you did) is generous indeed, though that is mostly because I suspect that the severe damage sustained by freezing your brain will cause irreversible disruption to the connections.
Also, I thought for a moment that "all people die" was a clever number, but I'm not sure whether it is or not. If it means "we develop immortality within our lifetimes", then those are very optimistic numbers indeed (though I think it is possible, I would assign this to south of 1%). Note that this IS an issue, as it would mean that any money you spent on cryonics would be wasted if you became immortal. If it is "all people on the planet die", I think those are very pessimistic numbers and it is very redundant with societal collapse (though honestly, societal collapse itself is just a subset of the cyronics company going out of business). Honestly I would group all of those together, and really, 40% odds of societal collapse are really, really pessimistic. 40% odds of your company going out of business, however, are quite likely - I would actually say the odds are higher than that.
It being impossible to extract all the information seems like a pointless number.
The technology never being developed to extract the information is a big nasty sticking point, and whether it is in fact possible is a big issue. But this is one of those numbers where whatever number you assign is a gigantic guess.
No one being interested in reviving you is entirely reasonable. Likewise expense. And I find it funny that someone said "nanotech", as if nanotech is a magic wand. We have nanotechnology today; it doesn't make things costless.
Reviving people in (real time) simulation for a reasonable cost is, I think, incredibly unlikely, given that at present, it looks like the fastest supercomputer may not be capable of simulating a human brain in real time - we're within three orders of magnitude of the bare minimum, but it looks like the bare minimum may not actually cut it.
Making AIs being outlawed is actually a reasonably likely thing to occur, I think.
The problem with trying to do any analysis like this is that you are guessing on far too many numbers for you to even claim you are doing real analysis, which is why I think it is an exercise in futility. Really the big sticking points are "freezing won't save you" and "you can't be revived", both of which presently have P=1 (that is to say, the odds of failure being 100%), and I don't see any evidence to the contrary. The rest of it is just icing on that, and guessing at many of the numbers is, I think, fairly silly.
1475799
It's entirely true that the rule of succession depends on the evidence being in the form of pass/fail tests, and that more complicated forms of evidence require different mathematical tools, such as Bayes' Theorem. However, this doesn't affect the overall approach that I've been trying to express.
A couple of years ago, I came across a handy little chart which gives some useful rules of thumb for all sorts of discussions. The first criteria seems relevant here. I can envision all sorts of pieces of evidence which could change my mind about the strengths of my beliefs in various aspects of cryonics. However, as long as you continue to state that 0% and 100% are your confidence levels in various propositions based on real-world evidence, instead of being as unapproachable as absolute zero temperatures or the speed of light, I am unlikely to find your estimates of the accuracy of your propositions very persuasive. And, as it seems I've been unable to communicate even the basics of treating probability logarithmically, even after trying four or five different approaches, then it seems unlikely that you'll find my suggestions of how to approach figuring out the order of magnitude of epsilon here to be sensical and/or worth considering.
Put another way, we're not speaking the same language, mathematically. And without even that small amount of common ground, the rest of what we may try to say to each other is likely to be misunderstood, or, at best, be completely ignored.
If you'd like, I can try coming up with further ways to try explaining logarithmic probability, or offer you references to read, either online or in textbook form. Or we could accept that we don't share this particular mental model, and look for alternate ways to express our ideas to each other that leave out mentioning this non-shared area. Or we could each go on to other discussions with other people.
I look forward to any reply you may wish to post; or, if none is forthcoming, then I thank you for the conversation so far and wish you the best.
1477537
The problem is that you're wrong. You can envision all sorts of things which would change your mind... but you won't. You have comforted yourself with the idea that you are open-minded, without actually being open-minded.
Me? It'd be easy to convince me. Demonstrate that you can freeze a brain without causing major damage. Prove that you can thaw frozen neurons and have them work again. Prove that you can freeze and thaw a whole brain and have it work. Demonstrate a means of extracting information from a brain that would allow you to generate a simulation of it. Demonstrate a means of simulating a whole human brain using resources equivalent to an average human being.
If you can't do those things, then why on Earth would I be convinced by your arguments as a rational being? Only an irrational person would be swayed by misdirection or appeals to emotion.
That's reality. Reality is you have to, at the very least, freeze a brain without causing major damage (we're talking 99.9% fidelity, here) as well as demonstrate that it can either be unfrozen or that it is possible to extract said information and simulate it at reasonable cost.
This is how you can tell a pseudo-rationalist (most people on LessWrong and similar sites) from a true rationalist - actual rational behavior.
There's a reason that cryonics is referred to as pseudoscience and woo, because, quite frankly, in its present state, it is. It uses the trappings of science to acquire money from people to fund projects. It isn't that it is fraudulent - it isn't - so much as it is willful self deception. Many cranks believe that they can actually get their ideas to work. That doesn't make them right, though.
I understand logarithmic probability, so you're barking up the wrong tree there. Logarithmic probability won't save you anyway - if different systems give you different results, then you know that one of your systems is wrong, or you are misusing the system. But the point I'm making here isn't about probability, it is that using probalistic calculations in the way you are doing isn't the correct way to arrive at a conclusion.
The problem here is that you're talking about made-up numbers and magical (not hypothetical, actually magical, as they are in some cases non-falsifiable - "Well, with magical future technology they'll be able to compensate for any errors") technologies which don't exist, and how likely X, Y, or Z is to happen on the basis of... what you hope for, ultimately, whereas I'm talking about what knowledge we have and what we can extrapoate out to. The reality is that if you drop a brain into a vat of LN2, you aren't going to end up with something which isn't a mess, cryoprotectants or no - and that alone is a P = 0 barrier. As everything else is contingent on being able to preserve things in this manner, you're in trouble. You have to demonstrate that this is workable; the null hypothesis is that this is not, and indeed it is easy to prove this - take a brain, dunk it in a vat of LN2, and examine it to determine how much damage has been done. Indeed, when you're down at LN2 temperatures, fracturing is inevitable - even if you flash freeze it, the glass itself will be put under tension and deform and crack and craze, and that's ignoring the damage that actual ice crystals will do.
Brain simulation is another P = 0 barrier - depending on the level of fidelity required to create an accurate brain simulation, we're talking about a theoretical future supercomputer, using vast amounts of power, being necessary to simulate a human brain in real time - and that's on the low end, as if you need higher levels of fidelity, it won't be able to do it in real time, or even close to it. We're talking about stuff at the top end of silicon's performance, here, with minimum size transistors. And that's merely simulating a brain, ignoring the difficulty in uploading it. And indeed, IF this is possible, then your best bet is probably not cyronics but simply having an extremely detailed scan done of your brain prior to your death, which is both far cheaper and easier to preserve. Indeed, even a destructive scan would be better.
Taking a brain out of cryo is another P = 0 barrier which is separate from brain simulation. If you can't do this and you can't do simulation, you're doomed. If freezing your brain doesn't preserve the information, you're doomed.
Squaring the circle is not a probablistic activity - it is a P=0 activity. Violating conservation of angular momentum is not a probabilitistic activity - it is a P=0 activity. There is nothing inherently unreasonable about saying something cannot be done, and indeed, there are many things we know cannot be done.
1477556
And one of the things we know cannot be done, in the physical universe, is state things with 100% certainty. :)
Squaring a circle takes place in the realm of mathematics, where any given proposition either follows from a given set of premises, or it doesn't. A four-sided triangle isn't going to be found. But that's abstract, not physical. (But even for the abstract realm, I can always call on Gödel...)
Conservation of angular momentum is a principle based on an extremely large number of observations, and thus inferred to have an extremely high level of confidence that future observations will continue to conform. But, just like drawing a long sequence of white beads out of a bag and making a prediction that future beads will be white, there's still the chance that a future bead will be blue. And, in fact, after looking very closely, conservation of angular momentum is only true with 'perfect precision' up to a certain point, after which we can no longer measure whether or not it continues to be true. You've probably already guessed the name that goes along with this form of uncertainty: Heisenberg. And that's without even getting into the randomness created by vacuum fluctuations creating random particles out of nothingness, which also throw off any attempt to claim that the universe proceeds by perfect Newtonian clockwork that can even theoretically be predicted to 100% certainty.
I can envision all sorts of things which would change my mind about evolution, such as finding rabbits in the pre-Cambrian... but just because I'm not /going/ to find rabbits in the pre-Cambrian, doesn't mean that I'm closed-minded by setting a certain standard for what it would take to change my mind.
1477556
I think it's worth thinking deeply about this paragraph.
Imagine flipping all of the signs. That is, instead of "Demonstrate that you can freeze a brain without causing major damage," read "Demonstrate that you can't freeze a brain without causing major damage." These propositions don't look symmetric- a single working design will suffice as a demonstration that it can be done, but every possible design must be exhibited in a demonstration that it can't be done, or considered in an impossibility proof.
For example, I believe that "the natural numbers can't be put into a one-to-one correspondence with the real numbers." This is an issue on which it would be very difficult to convince me I'm wrong, because if anyone comes to me with an example of a one-to-one correspondence between the naturals and the reals, I know how to construct an example that will blow their correspondence apart. I thus assign probability 0 to the correctness of any 1-to-1 correspondence from the naturals to the reals.
For you to say cryonics can't be done, or that the probability of it ever working is 0, then you need to have an argument at least as strong as Cantor's diagonal argument. I don't think the state of biomedical knowledge is advanced enough for anyone to have that knowledge, and going with the null hypothesis is very much not an argument as strong as Cantor's diagonal argument.
If you instead take the position "cryonics is a very hard challenge which is unlikely to work," then that strikes me as reasonable. Indeed, that's the position Benman and DataPacRat are taking! Benman thinks it's 50 times as likely that modern cryonics attempts will fail than that they'll succeed. With different models of how the various parts of the world that cryonics interacts with, estimates can wildly differ. For example, I think I'm more pessimistic than most people signed up for cryonics on the amount of damage done to the brain immediately after losing access to oxygen, and so assign odds that are roughly a million to one against modern cryonics attempts as they are today working. But if the law changes to allow perfusion of the living, then that objection is significantly reduced, and my odds are then close to Benman's. How likely is it that the law will allow perfusion of the living? It's certainly not impossible- it seems reasonable to expect that it will happen somewhere in the next fifty years. And, I might even argue that by signing up for cryonics now, I might make that more likely by adding more weight to the cryonics lobby.
1477994
Exactly. Which is why the burden is on the person making the claim to prove themselves to be correct. All you have to do is prove me wrong once, and I'm wrong.
This is a very basic principle. Russel's Teapot, the Invisible Pink Unicorn, and the invisible dragon in Sagan's garage all illustrate why. As was once said, "extraordinary claims require extraordinary evidence".
Your argument to the contrary is illogical.
This is an entirely different thing. It is important to remember that math is not science, and in fact, using the term "believe" there is strange; it isn't really something you believe, it is something which is absolutely true, given the standard definitions of natural numbers and real numbers and the general structure of mathematics. It is not even debatable; it is an inherent property of the logical system in question.
Wrong. The burden of proof is on you, the person making the claim that you can bring the dead back to life. The burden of proof is always on the person making the extraordinary claim.
I am holding the null hypothesis - that you cannot.
You are telling me you have a dragon in your garage. It isn't my job to disprove that you have a dragon in your garage - it is your job to prove it to me.
Nothing in science is as strong as Cantor's diagonal argument, because science is not built upon assumptions. Math is. This is why math has proofs and science has theories.
1470401
I'd suggest that it's more a matter of practicality. We don't know how to implement the 'horrendously complex approach' you describe, but do know how to 'take the simple approach. For those cryonicists who do die in the short-term, then the horrendously complex approach simply isn't an option - it's either cryo-preserve them as best we can with the tools we do have, or don't.
That said, shaking out enough cash /to/ work on research for the more complicated approaches is, indeed, an issue. As an interested party, I can make an educated guess that this stems from any cryonics organization's need to keep its finances as absolutely rock-solid as possible for the long term, to avoid a repeat of the Chatsworth incident, in which cash issues led to thawing. Alcor's total assets are on the order of $20M; CI's are on the order of $5M; almost all of which are taken up by equipment, hardware, and other infrastructure. Increasing the dues to fund more research would mean fewer people would remain signed up, which would leave people permanently dead who would otherwise be cryo-preserved. There don't seem to be any easy answers, short of convincing a billionaire to crash-fund cryobiology research... and cryobiologists themselves have become notoriously distasteful of cryonics, seeing it the way most people do, as some sort of weird fringe thing that being associated with would make them seem weird themselves, and unlikely to get as much grant money.
1480705
I don't want to step on cleversuggestion's toes in your conversation, but I would like to point out: science /never/ points out that something is 'correct'; all it does is point out that 'given the evidence, X is more likely than Y, to this particular statistical degree'. The Higgs boson hasn't been proven to exist - it's only been demonstrated to 99.9999-odd percent (five standard deviations) that a Higgs-like particle exists with 125 GeV. /Electrons/ haven't been 'proven' to exist - they've only been demonstrated with more 9's than the Higgs boson. (The rest of my comment on your comment would be an elaboration of this theme, if I were to write further.)
1481838
Sort of.
0<P<1 for confidence in science because all science is probabilistic. That being said, this isn't actually a scientific calculation, and I think most scientists would agree that violating certain physical laws is a P=0 event.
When you make predictions based on science, there are many situations which science won't assign any probability to - anything which violates conservation of matter-energy and conservation of angular momentum, no scientific model is going to give you a result consistent with those. Thus, results which violate those maxims are assigned P=0 when we make predictions.
This is not to say that the laws of physics are inviolable - we know that both Quantum Mechanics and General and Special Relativity are incomplete. But it is perfectly valid to make a prediction that something won't happen, based on present day models. Saying something has P=0 is a perfectly scientific thing to do - all you have to do to falsify it is show that P>0. In fact, the null hypothesis itself is the prediction that nothing will happen.
1481838>>1481872
I think this discussion has passed the point where it's valuable to anyone involved—you've both explained your positions, and you're saying the same things again, with slightly different examples. If you want to continue debating the nature of probability, please do so somewhere else.
1481872
"The thought you cannot think controls you more than thoughts you speak aloud."
As soon as you have declared that this physical law is inviolable, this principle absolute, or this theory proven to 100% - then you have taken that idea out of the realm of science. Any scientific law, even the laws of conservation, are only held to a certain, non-100% degree of confidence, as there always remains at least a small suspicion that this law isn't quite absolute, that there's some tricky detail behind it which includes an exception we haven't noticed yet. It's finding those tricky details, those laws behind what we currently think of as the laws, that allow science to progress at all.
Any scientist who says that violating any scientific law is a "P=0 event", instead of a "P is so close to 0 that we can usually treat it as 0", isn't being a scientist when they make that statement. (Or, at least, not a very good one.)
We've been going back and forth on this item for a while now, about 0% and 100% confidence levels being inapplicable outside of pure math. I'm curious; can you envision any form of argument which could, even potentially, change your mind? Or, looking at it another way, can you describe some level of persuasion which is definitely /insufficient/ to alter your opinion, and which would need to be exceeded to do so?
1481884
Sorry; I didn't see your post until after I finished writing mine. As you wish.
1481838 I think successful cryonic suspension is roughly as difficult as halting aging. So my preference would be for billionaires to fund aging research instead.
Benman, I hope you won't mind a quick response:
1480705
Sure, but I think we disagree on what the claim is. I think the claim is "It may eventually be possible to preserve and revive humans with cryonics," and it looks to me like you think the claim is "It is currently possible to preserve and revive humans with cryonics." You've correctly pointed out that humans put into long-term cryonic suspension have not been revived yet, and that the technologies proposed to do that are speculative at best.
But if you want to strengthen your position from "it isn't currently possible" to "it will never be possible," then you are moving from an ordinary claim towards an extraordinary claim. The null hypothesis is not "X is impossible," but "we don't know if X is possible or not."
In particular, a goal should not be ruled impossible because of the inadequacy of the best known method; saying "humans will never fly!" because humans don't have large enough wings is short-sighted, because there are other ways to fly. Saying the dead won't come back because if you bury or burn someone, the information is lost; well, there are other ways to die. (In particular, the advance of medical technology has already redefined death several times, such as with the advent of CPR, and I consider it reasonable to predict that trend will continue.)
Huh. Not only did I read am excellent short pony Fic today, I also learned a lot about probably and cryonics! Yayyyy! Learning is Magic!
Huh. Not only did I read am excellent short pony Fic today, I also learned a lot about probably and cryonics! Yayyyy! Learning is Magic!
The conversation seems like it got mired down on semantics. I think a fair compromise would be "there is some probability of this happening. One of the possible outcomes is that it won't happen. One of the possible outcomes is that it will happen. I am so confident that it will not happen that I can make decisions as if the chance were 0."
Now clearly some people can take a different stance and with research may discover what they need to display they are right. I think myself, TD, and others would openly accept proper proof. Taking the stance now that it is zero is not a barrier to proof.
What it is a barrier to is being willing to devote resources to it. I could devote some money equivalent to basic cable service monthly to cryonics. One could easily form an argument where this would be a better investment of those funds, due to the overall poor quality of entertainment value derived from basic cable, and the value of lengthened life due to cryonics.
1482153 This is an excellent example of the flaw to the above mindset. There are other alternatives to cryonics which also would be arguably better investments than basic cable television. At some point we have to recognize our limited ability to invest resources and choose which areas to actually put our investments in. The placing of 0% on cryonics (or any other option) in this case is reasonable and appropriate. It is a financial/investment choice not a scientific one.
In scientific/probability terms, the likelihood/confidence level of cryonics is greater than 0%, but too low to warrant investment. Because it will not gain my investment dollars, it is actually equal to 0%. If something were to have an actual probability of 0%, it too would not gain my investment dollars. That is the equivalency. It is up to each investor to decide for themselves what confidence level is needed to rise from that practical 0% into a willingness to invest, and how much.
Now, if in the future the confidence of the project is raised and I am in a position to change my investment, we can revisit the numbers.
The actual problem isn't "if" this will happen, but "when", and specifically, will your preserved body last that long?
The chance that it won't lies in disasters, crises and social revolutions that would lead to destruction (or just decay) of the preservation vault before the technology is developed. And of course the society still believes thawing you is a good idea. (I can easily imagine the society going ultra-capitalist and simply deciding this is not economically viable, or passing through an age of religious zeal where you'll be put "to proper rest".)
Transmetropolitan took an interesting angle on this. Instead of preserving the whole body, you preserve just the head, and wait until time comes when humans can rebuild the body from your DNA, restore all damage caused to your brain by the freezing, and put your brain in the newly cloned body. Obviously this pushes the viable date of thawing another couple centuries away, obviously increasing the chance of 'preservation failure' over that time. Still, the idea is solid, it just hinges on the optimistic assumption that human civilization will continue to develop. As Middle Ages proved, this is not to be taken for granted...
1470026
I doubt that any cryonically-preserved bodies would be simply unfrozen in a way that lets them biologically continue alive as before. The important thing is to preserve the brain patterns, so that a future society could scan the brain patterns into a computer (or into a brand new fleshy brain in a body if one had some particular attachment to them). Whether the frozen cells are still healthy enough to be reanimated is irrelevant. What's important is that the larger structures such as the connections of neurons are still discernible.