This essay was submitted to Open Philanthropy's Cause Exploration Prizes contest.
If you're seeing this in summer 2022, we'll be posting many submissions in a short period. If you want to stop seeing them so often, apply a filter for the appropriate tag!
Nothing OpenPhil or in fact anything anyone has ever done has truly "saved" anyone's life. Regardless of what we do, everyone still dies eventually. And almost every death happens at a time and in a manner that is not of the dying person's choosing. Truly saving someone's life would mean causing them to be in a state where they can live as long as they want, free of coercion. The brutal fact is that so far, we've only ever managed to delay the inevitable.
There is a vast scientific case to be made for preservation of people via aldehyde fixation and vitrification. Our current dominant theory of neuroscience, the synaptic basis of memory, says that preservation works. Even our less popular fringe theories of neuroscience say that preservation works. Preservation of human memory engrams via aldehyde fixation is a direct logical consequence of our current theories of neuroscience and information theory. Suggesting that preservation via chemical fixation does not work is logically equivalent to proposing a new memory system with exotic, biologically implausible elements and with no supporting evidence. Either preservation, done correctly, is sufficient to capture the essence of a person, or else the most basic foundations of the fields of neuroscience and biochemistry are catastrophically wrong, to the point of being literally fraudulent.
We can and should implement a broad program of engram preservation today, and it's possible for OpenPhilanthropy to make major strides in this area. A true option of preservation, made broadly accessible, will likely have profound secondary effects for society as a whole. If people feel that they truly have a place in the future, they are more likely to plan for that future. If people have a practical solution for dealing with their own death, they will be able to think intelligently about their own death without as many psychological blocks. And if people can think about their own death more coherently, then they are likely to also think about other scary future possibilities in a more sober manner.
Preservation can be implemented cheaply and become the standard way we deal with death as a responsible society, and there is not much more work to be done to achieve a robust preservation program in as soon as five years. The limiting factors are funding, establishing common knowledge among disparate scientific fields, and political will.
We've never actually saved anyone's life.
The point of philanthropy, and in fact the point of most human endeavors, is to enable human flourishing. We enable human flourishing in many ways, such as by increasing our capabilities through technology, or by decreasing coercion whether natural or man-made.
And the most central thing to human flourishing, and indeed the flourishing of any organism, is life itself. As long as you are alive, you have a chance to improve your condition and help others. You can move freely in the world, and learn, and potentially find happiness. But as soon as you are dead, your voice is silenced forever, and any chance of growth or improvement is permanently terminated.
Unfortunately, no intervention invented thus far has actually saved anyone's life! At the end of the day, everyone still dies.
Truly saving someone's life would involve transforming them from a state where they are definitely going to die, whether now or in 50 years, to a state where they can live as long as they like in good health. And unfortunately, none of the interventions we've thus far developed are up to the task.
It is better to think of the interventions available to us as simply delaying the inevitable. Cure cancer, for example, and you add about 5 QUALYs to everyone's life, and then everyone still dies from other problems like cardiovascular disease. It's a simple fact that almost everyone dies at a time and manner that they would not choose for themselves, Unwilling death is a grave injustice, perhaps the ultimate injustice of our age. Even the people who make use of end-of-life protocols to choose their time of death are not making a truly free, uncoerced choice, as the suffering associated with aging and terminal diseases is a natural form of coercion that prevents true consent. Take away the terminal disease and health problems, and they would likely not choose to end their life.
These are grim facts, but it is important to address them head on, because only by acknowledging the fact that we still don't know how to truly save anyone's life can we hope to soberly consider our options and work towards a true solution to the most profound injustice today.
Taking stock of our situation
It's important to establish a few bits of common ground, a reasonable agreement on facts about the world, to discuss our options for improving our situation. The following points are the baseline on which to build the case that we can start saving people’s lives—for good—today.
1) There is no magical thing, unconnected from the physical world, which makes a person who they are. In this sense, materialism is broadly correct. Humans are extremely complicated, finite structures that obey the laws of physics. A person's memories, skills, personality, and so on are physically encoded in their body and especially their brain.
Of course I do not deny that we humans are embedded in a greater family / society / ecosystem, and that the context in which we are embedded can dramatically change our attitudes and capabilities. Without an English-speaking society around me, or without my family, I am in some sense a different person. Nevertheless, you can draw a bright line around the physical border of a person's body, specifying a certain collection of matter, and that matter will contain the colloquially-understood "essence" of the person, though that essence may need the right social context to be fully realized.
2) It is possible in principle to perform vast feats of engineering, especially if they've already been accomplished by nature. Specifically, I believe that it is possible for us to rise to the challenge and conquer the immensely complex task of "mind uploading", which involves scanning an entire human body and simulating its internal causality in enough detail that the uploaded version will behave in all situations in the same manner that the pre-uploaded's causality would dictate.
I do not for a moment want to undersell the vast complexity of mind uploading. The brain has a quadrillion synapses and it's currently unknown what level of simulation detail is enough to make a reasonably faithful version of a person. Nevertheless, the technological components of mind uploading are clearly possible from an engineering perspective. Scanning a brain at nanoscale resolution, a likely prerequisite for mind uploading, would be possible within 10 years for a budget of around $1,000 billion, and there's every reason to believe that that cost will continue to exponentially decrease, as it has for the last 40 years. Simulating a single neuron at a level of detail sufficient to recapitulate neuronal dynamics and learning is clearly possible in principle, because the neuron itself is built out of biomolecules which are themselves simulatable. Doing the same for 100 billion neurons is a matter of (monstrous) scale. The ultimate size/power requirements of a computer adequate for simulating a human, if we really know what we're doing, should require no more than the size of a human body and no more energy than 2,000 Calories a day,, since each person is living proof that evolution has already achieved this engineering feat. If nature has already met the engineering challenge, then there is no reason in principle that we can't eventually match or exceed what nature has already done.
3.) Preservation of a person is physically possible using today's technology.
There is a vast scientific case to be made that preservation of a person is possible via aldehyde fixation and subsequent vitrification, which will not fit in this proposal. I've worked over the last several years to refine this technology, including winning the Large Mammal Brain Preservation Prize in 2016 and more recently running a human brain bank in the Pacific Northwest. The basic summary is:
Information Theory tells us what it means to preserve a complicated information-containing artifact using a preservation method, and that is that information-theoretically distinguishable initial states need to map onto distinct final states. Some examples:
Example: preserving books.
Say you want to preserve books. There’s more than one way to define a unique book. For example, you could count two different printings of the same title as the same book, if your criteria for a book was based on the same sequence of letters. Let’s look at a more stringent definition: a book is different from another book if their ink patterns are perceptually distinct according to a typical human observer paying careful attention. That is, a book with even a slight, smaller-than-a-letter difference that's noticeable to the human eye—an ink smudge, a missing period—would be considered different than the same title without that ink smudge. Now, imagine pouring epoxy onto the book in order to preserve it. The epoxy works its way into every page, soaking the book through-and-through. And the epoxy glues the ink firmly to each page and the pages themselves to each other, before fully solidifying the book into a solid block of plastic. Does this method of preservation preserve books according to our criteria? Yes, because the epoxy does not move ink molecules enough to result in any confusion about which ink-pattern must have generated the resulting plastic block. Note that actually being able to open the book is not a requirement for a successful preservation; the resulting plastic book only needs be physically distinct from plastic blocks created from other ink patterns. Note also that we don't need to be able to actually read the book; we don’t need to have good theories about the nature of literature; we don’t need to know how to write the book ourselves; we don’t need to be able to critique this book particularly or literature in general. In order to be confident that we're preserving the book, we just have to understand how ink, paper, and epoxy works and have a reasonable understanding of which differences are perceptible to the human eye.
Example: preserving immune memory.
Consider immune memory, the body's knowledge of self and other, a memory which is continuously updated throughout life in response to infection, vaccines, etc. Is aldehyde fixation sufficient to preserve immune memory? If we consider what meaningfully distinguishable immune memory states might be, a sensibly conservative starting point would be the capability of the immune system, taken as a whole, to generate different antibodies at different rates in response to stimuli such as infections. If two immune systems are able to make different antibodies, then they are different. Consider my own entire complement of antibodies and antibody-generation capabilities. This is some vast but ultimately finite set of distinct proteins. Now consider an exact copy of me, except that a single b-cell is modified so that it produces an antibody that differs from the previously produced antibody by a single amino acid. We can say that this subtle difference counts as the two immune systems being meaningfully distinct. Can aldehyde fixation preserve this subtle immune memory difference in an information-theoretic sense? Yes, because aldehyde fixation is capable of preserving every cell, and essentially all proteins, DNA, RNA, and other large biomolecules in an entire organism. Although the difference between these two example immune systems exists in only a single cell, the difference is actually quite large from a biomolecular perspective. That single amino acid difference is encoded in that cell in multiple places: in the DNA in the cell's nucleus, in the RNA that is transcribed to construct antibodies, and in the final antibodies themselves. There are thousands of different molecules that meaningfully encode the difference between the two immune systems in that single cell, and aldehyde fixation is able to preserve essentially all of them. Therefore, aldehyde fixation is up to the task of preserving immune memory. Note that we don't need a complete theory of immune memory—we don’t need to understand exactly how immune cells specialize to make different antibodies, for example—to be extremely confident that immune memory as we've defined it is preserved by aldehyde fixation in an information theoretic sense.
Example: preserving memory engrams
Similar to the above two examples, we must first define what it means for a complex information-containing system, in this case a person, to have "meaningfully different" engrams from another person. And then we must consider whether our preservation method will truly produce two different artifacts when starting with two similar, but different, beginning points. As a conservative starting point, we say that two people have "different" engrams if there is some stable difference in their behavior that persists over at least 24 hours. For example, consider two versions of me which are entirely identical, except that when I go to a party with a host I’ve never met, the host introduces herself to one version of me as Alice, and to the other version as Brittany. Twenty-four hours later, both versions of me still remember the name and will respond differently when asked who the host was.
What does neuroscience say would have to happen in these near-identical nervous systems to generate a stable behavioral change that lasts more than 24 hours? While we don't have a complete theory of neuroscience, we know the low-level biochemistry of engram formation reasonably well and can definitively answer this question. Neuroscience says that hundreds to thousands of synapses would have to be different between the two nervous systems to reliably result in a different behavior. If those specific synapses were disrupted, then the behavioral difference would vanish. Now, what level of physical difference is aldehyde fixation able to preserve? Aldehyde fixation, competently performed, will not only preserve every single cell and synapse in the two nervous systems, but will also preserve essentially all proteins within each synapse. This means that aldehyde fixation is able to preserve such subtle differences between nervous systems that even two nervous systems that differed in protein content at a single synapse could in principle still be differentiated after preservation. Note that, like the two previous examples, we DO NOT need a complete theory of neuroscience to say with confidence that memory engrams can be preserved in an information theoretic sense. We just need to know enough about how the biochemistry of engram formation works to understand the magnitude of the changes that must happen in a brain to meaningfully encode a memory, and then compare that to the preservation capabilities of aldehyde fixation.
Getting comfortable with these concepts
I've previously talked with OpenPhil about engram preservation and was told that it's an interesting idea, but that OpenPhil doesn't have the specialists to properly evaluate it. I suggest that for something as important as engram preservation that it's worth it to actually evaluate it. I call on OpenPhil to consult with experts in information theory, biochemistry, neuroscience, and clinicians who perform deep hypothermic circulatory arrest or otherwise expose the nervous system to extreme states. Get these experts in a room together, and have a focused conversation about the concept of preservation. I've done this before and already know how it will go. When you do this exercise, you find something very interesting: The neuroscientist will begrudgingly admit that in order to have a long-term difference in behavior you must have synaptic-level changes in the nervous system. The biochemist will acknowledge that aldehydes like glutaraldehyde will retain essentially all nanostructure and proteins during fixation. The information theorist will admit that this is a sufficient condition for preservation. If anyone brings up objections along the lines of the dynamic activity of the brain being important, the clinician will be able to respond with clinical data from deep hypothermic circulatory arrest, ischemia, and other extreme brain conditions that conclusively show that the long-term memory is not stored in the dynamic activity of the brain. All of the knowledge to clearly derive that preservation is possible is there in the basic, 101-level facts in these fields. Almost all of the objections that each expert will have are simply ignorance of basic facts in fields not their own. The knowledge is out there, it's just not common knowledge, because these fields aren't good at talking to one another.
At the very least, I hope that OpenPhil will take this idea seriously enough to actually properly evaluate it, because it's too important to continue to ignore.
Making preservation happen
I've already demonstrated essentially perfect preservation in rabbits and pigs, in a laboratory euthanasia setting. To make preservation a legitimate option for real people it must be made into a robust medical technique, able to deal with the many exceptional conditions that occur with each person at the end of their life.
Like introducing any new medical technique, it's a complex task, but it's by no means insurmountable. With enough funding and attention, it will be relatively straightforward to complete the additional research to determine the constraints under which preservation can be safely achieved in realistic settings.
We can start a clinical trial for preservation methods, studying the quality that can be achieved as well as the effect on the end-of-life experience. We can start extremely high quality brain banks that will further demonstrate quality and greatly accelerate research. We can fund high quality reviews of the literature to help establish shared knowledge. We can build partnerships with hospitals to offer opt-in emergency preservation for when their risky surgeries fail with no further recourse. We can create educational materials to help the general public understand the new option that is available to them. And looking back, we can know that we preserved the living memory of our generation and potentially saved the first lives for real.
Preservation can absolutely become a mainstream end-of-life option in a short number of years. I will work to make it happen whether or not OpenPhil is able to help. But I think the timeline could be accelerated with substantial funding and additional political will.
2. Ostasiewicz, Paweł, et al. “Proteome, phosphoproteome, and N-glycoproteome are quantitatively preserved in formalin-fixed paraffin-embedded tissue and analyzable by high-resolution mass spectrometry.” Journal of proteome research (2010).
3. Hayashi-Takagi, Akiko, et al. “Labeling and optical erasure of synaptic memory traces in the motor cortex.” Nature (2015)
4. Percy, Andrew, et al. “Deep hypothermic circulatory arrest in patients with high cognitive needs: full preservation of cognitive abilities.” The Annals of thoracic surgery (2009)
5. McIntyre, Robert L., and Gregory M. Fahy. "Aldehyde-stabilized cryopreservation." Cryobiology 71.3 (2015): 448-458.
Coming here to say this is an interesting cause area that I've never seen before. In general I'm liking the fact that new causes are being suggested for EA that aren't standard.
Sharmake, this is an encouraging comment, thank you. Unfortunately I don't think the EA leaders share your view. They are fundamentally conservative, not willing to fund things that make them look weird. There's nothing weird about talking about x-risks. It makes you seem noble and foresighted. But it is definitely weird to talk about living on after biological death. It makes a lot of people with alternative beliefs uncomfortable. Sadly, this means EA ignores that small minority who would today advocate that everyone in a better future society should have the option of brain preservation at the end of their lives, whether or not they choose to exercise that option.
Said in another way, almost all the folks and all the leaders that I've met in EA seem to position their views on BP in relation to what the majority of people think of death, rather than considering the future of death in an accelerating and information-protective universe. The world is poorer for it, as 150,000 unique, irreplaceable minds and memory sets continue to die daily. I think it is an easy and evidence-based prediction that an increasing fraction of those minds will eventually demand the option to preserve themselves at biological death. I hope some in this community wake up to this knowledge and become advocates for this cause, rather than choosing the psychologically and politically safer option to continue to ignore it or deny its value.
It seems useful to at least explore beyond the simplistic life=good, death=bad paradigm which is so common to the human experience.
First, any claims any of us may make about death are based on exactly no information from that realm, should such a realm exist.
Certainly everyone is entitled to their own preferred theory about death, and it's not my intention to argue with any of the theories, given that I too have no idea what death is.
Further, I would agree that maybe it's unkind to even mention that we have no idea what we're talking about when it comes to death, given the strong human need to have some kind of answer.
Personally, while I'm a very lucky person who has enjoyed 70 years of life, it's not clear to me that I would even wish to go on being Phil Tanny indefinitely.
One of the things that can happen as one ages is that you see patterns in the human experience that repeat over and over and over. When we're young these patterns can be fascinating, because they're new. When you get old sometimes these patterns seem more like a TV show that was really good the first couple of times you watched it, but after 70 viewings the show can become more than a bit tiresome. One of my aunts turns 100 today. I can't imagine how tired of the show she must be....
I quite liked this post, but just a minor quibble. Engram preservation still does not directly save lives, it gives us an indefinite amount of time, which is hopefully enough to develop the technology to actually save them.
You could say that it's impossible to save a life since there's always a small chance of untimely death, but let's say we consider a life "saved" when the chance of death in unwanted conditions is below some threshold, like 10%.
I would say widespread engram preservation reduces the chance of untimely death from ~100% (assuming no longevity advances in the near future) to the probability of x-risks. Depending on the threshold, you might have to deal with x-risks to consider these lives "saved".
Pablo, I submit you haven't thought carefully enough about the nature of the postbiological future. Once humanity has the capacity to preserve and emulate minds, those minds are as impervious to x-risks as is the entire network of backups. Once minds are stored redundantly and both on and off Earth, x-risks themselves become negligible. There is something deeply accelerative and protective of advanced complexity in our universe that is typically ignored by the x-risk community. It doesn't serve their fundraising and political purposes to see it, and it is truly weird vs. biology's dependence on planets, suns, etc.. Yet it is apparently (most likely, the default model) how evolutionary development works, on all Earthlikes in our universe. We just need the courage to see and learn from it.
Would you say that anesthesia doesn't "directly" extend life? After all, it only makes it possible to do certain surgeries, and it's really the surgery that is "directly" extending the life. And yet "the hospital" extends lives through it's interventions, one of which is anesthesia and without which the hospital would not function or be able to do surgeries effectively.
This is just the standard problem of assigning credit when multiple causes are involved. I'd propose the same sorts of tests we use in other cases, such as considering whether in the absence of preservation it would still be possible to save someone's life with future technology. The conclusion I draw is that preservation technology saves lives in a similar way to how anesthesia extends lives by enabling better surgeries. So it's perfectly sensible to talk about preservation directly saving lives even though it's not the only technology required to do so -- after all, if the life does get saved eventually, then preservation would deserve a hefty amount of the credit. Just as anestheologists deserve a hefty amount of credit whenever a surgery is performed successfully, and anesthesologists can be said to be directly extending people's lives directly through their work, as a critical pillar of a surgical team.
Dealing with x-risks in a satisfactory way and inventing uploading technology are also necessary to save someone's life, and will deserve substantial credit if lives are truly saved. And preservation is a substantial and irreplacable part of the constellation of truly life saving technologies for people alive today.
Great post. I fully agree that this seems to be a worthwhile area of funding. Although it was written too soon to be included in the Open Phil prize, I wrote a post on a similar topic here: https://forum.effectivealtruism.org/posts/sRXQbZpCLDnBLXHAH/brain-preservation-to-prevent-involuntary-death-a-possible
I wonder if the EA community feels they have already spent too many "weirdness points" on other areas -- mainly AGI x-risk alignment research -- and don't want to distribute them elsewhere. Evidence for this would be that other new cause areas that get criticized as "sci-fi" or people use the absurdity heuristic to discount would be selected against; evidence against it would be the opposite.
It's also possible that the EA community doesn't think it's a very good idea for technical reasons, although in that case, you would at least expect to see arguments against it or research funded into whether it could work.