Writing is extremely low-bandwidth. If I recall correctly, Shannon did some experimentation and found that per letter, English was no more than a bit and I've seen other estimates that it's less than a bit, per letter. (In comparison, depending on language and encoding, a character can take up to 32 bits to store uncompressed. Even ASCII requires 8 bits/1 byte per character.) And given the difficulty of producing a megabyte of personal information, and the vast space of potential selves...
If we're going to try to preserve ourselves through recorded information, wouldn't it make much more sense to instead spend a few hundred/thousand dollars on lifelogging? If you really do record your waking hours, then preservation of your writings is automatically included - as well as all the other stuff. Plus, this solves the issue of mundane experiences.
EDIT: I've put up some notes at https://www.gwern.net/Differences
Could you say more about said research?
It's basically as above. Traits like IQ offer remarkable predictive power; Big Five on top of IQ allows more prediction, and the second paper's Small 100 seems like it'd add nontrivial data if anyone runs it on a suitable database to establish what each trait does. Much of the remaining variation can be traced to the environment, which obviously doesn't help in establishing that human personalities are extremely rich & complex.
Long-term memory is much smaller than most would guess when talking about 'galaxies of galaxies of neurons', and autobiographical memory is famously malleable and more symbolic than sensory. (Like dreams: they seem lifelike detailed and amazing computational feats, but if you try to actually test the detail, like read a book in your dream, you'll usually fail.) Skills don't involve very much personality, either, since there are so many ways to be a bad amateur and so few to be an expert (consider how few items it takes to make a decent expert system - not millions and billions!) and are measurable anyway. "The mind is flat", one might say. Once you get past the (very difficult) tasks of perceiving and modeling the
Dear Diary,
My doctor told me today that of all the Elven Jedi he has ever treated I have by far the largest penis...
(Words we write down are only very loosely correlated with who we are.)
(Words we write down are only very loosely correlated with who we are.)
No, it says something about you that you made that joke and Morendil made the other. (this is also a response to nickernst's comment about not knowing himself)
Also, using the terms "Elven Jedi" is a pretty clear indicator of you being more affiliated than average with the scifi/fantasy geekdom. Choosing to include "sexual" content such as the size of one's penis also says something - there are lots of people who'd be too prudish to go even that far. Some weak inferences could probably be also drawn from the fact that you used the expression "Dear Diary".
None of this would be enough to justify any firm conclusions by itself, of course. But combined with a large enough number of other weak pieces of evidence...
Alicorn's disavowal is due to fear of something learning about her through a series of weak inferences; however, ironically, her disavowal/vow to avoid revealing data useful for weak inference is itself grist for the weak inference mill. (And hopefully my example inferences show that the weak inferences may not be all that weak.)
That something learning about her might be entirely neutral. On the other hand, it might be an unFriendly AI programmed by Black Panthers who were picked on as kids by nerds and are irritated by Ryan North's longboard stylings.
Unfortunately people who can't afford cryonics are unlikely to have the time or resources to create meticulous records of themselves. When you consider the opportunity cost of creating such records, the actual materials needed, and the cost of preserving them reliably for at least several decades, it isn't obvious that this is much cheaper than cryonics.
There's also the problem that most people don't consider 'make a perfect copy of me' and 'bring me back to life' to be equivalent operations, and the ones who do are almost all Western intellectual types who could easily afford cryonics if they actually wanted to. The world’s poor almost all see their personal identity as tied to their physical body, so this kind of approach would seem pointless to them.
I find most of the public debates on these issues rather myopic, in that they focus on the issue of surveillance by governments as the main problem. What I find to be a much more depressing prospect, however, are the consequences of a low-privacy society that may well come to pass through purely private institutions and transactions.
Even with the most non-intrusive and fair government imaginable, if lots of information about your life is easily available online, it means that a single stupid mistake in life that would earlier have only mild consequences can ruin your reputation forever and render you permanently unemployable and shunned socially. Instead of fading memories and ever more remote records about your past mistakes, they will forever be thrown right into the face of anyone who just types your name into a computer (and not to even mention the future more advanced pattern-matching and cross-referencing search technologies). This of course applies not just to mistakes, but also to any disreputable opinions and interests you might have that happen to be noted online.
Moreover, the social norms may develop to the point where it's expected that you constantly log the details ...
Kaj_Sotala:
But that doesn't make any sense - if almost everyone's going to have some stupid mistakes of theirs caught permanently on file, then all that will happen is that you'll find out you're not the only one who has made stupid mistakes.
There are at least three important problems with this view:
First, this is only one possible equilibrium. Another possibility is a society where everyone is extremely cautious to the point of paranoia, so that very few people ever commit a faux pas of any sort -- and although most people would like things to be more relaxed, they're faced with a problem of collective action. I don't think this is at all unrealistic -- people living under repression quickly develop the instinct to watch their mouth and behavior obsessively.
Second, even under the most optimistic "good" equilibrium, this argument applies only to those behaviors and opinions that are actually widespread. Those whose unconventional opinions and preferences are in a small minority, let alone lone-wolf contrarians, will have to censor themselves 24/7 or suffer very bad consequences.
Third, many things people dare say or do only in private are not dangerous because
I don't have direct access to a large percent of my memories. Many cannot be put into words, and I don't just mean music and imagery. The knot between these memories is utterly complex. In self-reflection, I am dishonest with myself, and I don't feel like it is so. My mother has the idea that poetry is a means of most honestly recording some of the difficult-to-explain thoughts, but I think that the scope, inexpressibility and interconnection of the memories makes this infeasible.
How good a record is good enough? In truth, I don't think we even know enough to get the order of magnitude right. The best I can offer is that you need to record as much as you are willing to.
But this estimate is essential. By deciding to pursue this course of action, you in effect state that you estimate sufficient probability of it being enough to justify the additional effort. You can't say "I don't know" and act on this knowledge at the same time.
This post addresses the subject of the appropriate human data compression format. Though an interesting idea, I think that the proposed method is too low in resolution. You acknowledge the lossiness, but I think it's just going to be too much.
Although the method you advocate is probably the best thing short of cryonics, I doubt there is any satisfactory compression method that can make anyone that's more similar to you than a best friend or a family member who gets stuck with your memories. It's better to have too much data than too much.
Because we share the same evolutionary past as all of our conspecifics, the biology and psychology of our brains is statistically the same. We each have our quirks of genetics and development, but even those are statistically similar among people who share our quirks. Thus with just a few bits of data we can already record most of what makes you who you are.
I'm not confident in this part. Although a large percentage of human biology and psychology are identical, the devil is in the details. From a statistical perspective, aren't humans and chimps practical identical also? Percentage similarity of traits is probably the wrong metric, since sm...
How many hours do you estimate you'll be putting into your autobiography for the resulting record to be "good enough"?
Next question, what is your hourly pay rate?
How good a record is good enough?
No record in English (and I'm using English as a shorthand for any human language) can ever be good enough. English is not a technology for transmitting information.
English is a compression format, and a very lossy and somewhat inaccurate compression format at that. But it has a stupendously high compression rate and compression algorithms with reasonable running speeds on specially adapted hardware (i.e. brains), so for the particular purposes of human communication English is a pretty decent option.
I own a t-shirt with...
A few thoughts:
-This would require both an enormous amount of time spent meticulously documenting your experiences (most of which would be mundane), and incredible writing skill to be able to capture various nuances of emotion. The number of people who are able to satisfy both these conditions may be less than the number who will sign up for cryonics.
-It's not clear to me that there's any consistent way to translate the written expression of a memory (particularly an emotional memory) into a mental state, partially because...
-I'm not sure writing is a fine...
One way to help test the feasibility of this plan is to both write prolificly and undergo cryonic preservation.
On the one hand, I do expect a society after a positive Singularity to be interested in, say, reconstructing Feynman from the evidence he left, and of course the result would be indistinguishable from the original recipe to anyone who knew him or knew his writings, etc. It goes without saying that I expect this to be awesome, and look forward to talking with reconstructed historical figures as if they were the originals.
However, I do suspect that there's a deep structure to an individual human's experience and thinking which might be essential to the cont...
You should probably give a number for the cost of cryo.
As far as I know, $9,000 is the cheapest possibility, which is cheaper than many cars, and there are a lot of those in the world.
Part of the reason why I make available records of e.g. the books I own, the music I listen to and the board games I've played (though this last list is horribly incomplete) is to make it possible for someone to reconstruct me in the future. There's a lot of stuff about me available online, and if you add non-public information like the contents of my hard drive with many years worth of IRC and IM logs, an intelligent enough entity should be able to produce a relatively good reconstruction. A lot would be missing, of course, but it's still better than noth...
You don't want a written diary, you want a highly efficient miniature camera that's always on. And maybe an option to annotate it in real time.
we need to inject the stuff that sets you uniquely apart even from your fellow humans who share your statistically regular quirks: your memories. If the record of your memories is good enough, this should effectively create a person who is so much like you as to be indistinguishable from the original, i.e. restore you.
I put a lot less importance on memory than you do. For instance, if I suffered amnesia and was not conscious of any of my previous experiences, I would still be me. In fact, given the choice between (A) someone who had a completely diffe...
I don't think that writing yourself into the future would work very well, but I've got another idea for a cheap cryonics-substitute: get your brain frozen in plain old ice. By the time we get whole brain emulation, a brain frozen in ice may contain enough information to replicate on a computer, even if it cannot be biologically revived like a cryogenically frozen brain.
Permafrost burial has been explored, but is generally considered an inferior option. If I were going for a cheap cryonics substitute, I'd try plasticization. A lab can do a head for a couple thousand bucks, it preserves enough microstructure for an electron scanning microscope, and there's no worries about staying cool.
I find it deeply weird that nobody has pointed out that the information describing you, written as prose, is not conscious. This is a major drawback. The OP mentioned it, and dared people to take him/her up on it, and nobody did.
I attribute this to a majority of people on LW taking Dennett's position on consciousness, which is basically to try to pretend that it doesn't exist, and that being a materialist means believing that there is no "qualia problem".
I don't follow. The OP didn't claim that just having the written information would be enough. They were saying that the information could be used to build a copy of you. The prose might not be conscious, but the copy would be.
With cryonics at $9000, you have to ask which method is getting you the most utility per unit effort. $9000 equates to about 200-600 hours of work for most reading this, but if the writing takes an hour a day for the rest of your life, that's 10,000+ hours.
Of course the best protection would be to do both.
This post is in dire need of a reference to Hofstadter's I am a Strange Loop.
Also maybe Halperin's The First Immortal which explicitly considers the possibility raised here.
Also maybe Lion Kimbro.
This doesn't seem particularly useful to me. Even if the written copy could be identical to me in every way, I would place a much lower value on the creation of such a copy than on the extension of my current life. You're right that this might be slightly preferable to death, but I certainly wouldn't position it as a real alternative even to cryonics.
We are going to have to rely on simulations of the deceased for the foreseeable future. Individuals who have not left extensive records will be relatively more quality simulations.
Hopefully at some point a sufficiently advanced simulation will exist which can interpolate the remainder of humanity, but even then we are left looking for a reason to do so.
Evidently. :)
This brings up a related point. How do you write your skills into the future? You can't just write "As of 2010 I was an excellent piano player".
But wait - maybe you can. If you're assuming a reconstruction technology which can uncompress verbal descriptions of behaviours into the much more complicated expression of such behaviours in terms of the neural substrate, then quite possibly this technology will also have massive general knowledge about human skills allowing it to uncompress such a statement into its equivalent in neural and muscular organization.
But then, what a temptation! As of 2010 I am not, in fact, able to play the piano, but if this record for the future can also serve as my letter to Santa, why not? It's not as if any of it is readily verifiable. I could say I like the taste of lemon when actually I hate it.
This line of thought isn't to ridicule the idea of writing yourself into the future - just to bring out some consequences the OP may not have thought about.
I've had a similar idea for a while. It involves reconstructing people from the memories of those who knew them, like Kaj_Sotala describes. So, for instance, my maternal grandfather died a few years ago. But if a bunch of us who knew him lived into a future where our memories of him could be scanned and analyzed, a copy of him could be built that we would find indistinguishable from the original as far as those of us who remember him are concerned. I thought it might be useful from a comfort standpoint.
How much is gained by writing about yourself? Aside from personal development aspects, less than video-logging. But likely different information than video-logging. Possibly stuff buried in your childhood. Logging biometrics like heartbeat could be more total benefit than video-logging, but again different information. Having a combination could prove more useful in a synergistic manner than any one information source on its own, because comparing them against each other allows you to infer more information.
The same goes for cryonics. At a guess I'd say pr...
This is an idea Paul Almond suggested a while ago in his article Indirect Mind Uploading.
Also, he has quite a few new AI related articles on his website http://www.paul-almond.com. I haven't read any of them yet, so I can't comment.
I think this is missing out on a lot of other higher bandwidth source of information about us. Part of the problem is a focus on output, as if creating a pale imitation of the process of scanning a brain. But you could also reconstruct much of brain by looking at the things that went into it: DNA, the environment. tack on the records that are already automatically made of its choices, and you isolate a very small part of the potential mind space.
Any person with my DNA, my grades, my bookshelf, my pantry, my bank statements and my web VIEWING history would...
...How good a record is good enough? In truth, I don't think we even know enough to get the order of magnitude right. The best I can offer is that you need to record as much as you are willing to. The more you record, the more there will be to work with, and the less chance there will be of insufficient data. It may turn out that you simply can't record enough to create a good restoration of a person from writing, but this is little different from the risk in cryonics of not being well preserved enough to restore despite best efforts. If you're willing
I'm very curious as to your theory of what happens if you do both. That is, suppose you're cryogenically frozen and then revived, while someone also makes a top-notch copy of you based on the recorded memories you left behind. It seems rather obvious that you can't have double-you, so what happens?
This hypothetical suggests to me that one or both are doomed - and if it's just one, I'd think it's this method you've suggested that wouldn't work. But I really haven't thought too hard on this issue, so I'm curious as to what others think the solution/outcome is.
Cryopreservation works pretty well for embryos, eggs and sperm. Or, if you are feeling optimistic, you could sequence your DNA - and store that. That is not preserving everything - but it should be enough to make some identical twins - which is pretty close for most purposes.
Cryonics appears to be the best hope for continuing a person's existence beyond physical death until other technologies provide better solutions. But despite its best-in-class status, cryonics has several serious downsides.
First and foremost, cryonics is expensive—well beyond a price that even a third of humanity can afford. Economies of scale may eventually bring the cost down, but in the mean time billions of people will die without the benefit of cryonics, and, even when the cost bottoms out, it will likely still be too expensive for people living at subsistence levels. Secondly, many people consider cryonics immoral or at least socially unacceptable, so even those who accept the idea of cryonics and want to pursue taking personal advantage of it are usually socially pressured out of signing up for cryonics. Combined, these two forces reduce the pool of people who will act to sign up for cryonics to be less than even a fraction of a percent of the human population.
Given that cryonics is effectively not an option for almost everyone on the planet, if we're serious about preserving lives into the future then we have to consider other options, especially ones that are morally and socially acceptable to most of humanity. Pushed by my own need to find an alternative to cryonics, I began trying to think of ways I could be restored after physical death.
If I am unable to preserve the physical components that currently make me up, it seems that the next best thing I can do is to record in some way as much of the details of the functioning of those physical components as possible. Since we don't yet have the brain emulation technology that would make cryonics irrelevant for the still living, I need a lower tech way to making a record of myself. And of all the ways I might try to record myself, none seems to better balance robustness, cost, and detail than writing.
Writing myself into the future—now we're on to something.
At first this plan didn't feel like such a winner, though: How can I continue myself just through writing? Even if I write down everything I can about myself—memories, medical history, everything—how can that really be all that's needed to restore me (or even most of me)? But when we begin to break down what writing everything we can about ourselves really gives us, writing ourselves into the future begins to make more sense.
For most of humanity, what makes you who you are is largely the same between all people. Since percentages would make it seem that I have too precise an idea of how much, let's put it like this: up to your eyebrows, all humans (except those with extreme abnormalities) are essentially the same. Because we share the same evolutionary past as all of our conspecifics, the biology and psychology of our brains is statistically the same. We each have our quirks of genetics and development, but even those are statistically similar among people who share our quirks. Thus with just a few bits of data we can already record most of what makes you who you are.
Most people find this idea unsettling when they first encounter it and have an urge to look away or disagree. "How can I, the very unique me, be almost completely the same as everyone else?" Since this is Less Wrong and not a more general forum, though, I'll assume you're still with me at this point. If not, I recommend reading some of the introductory sequences on the site.
So if we begin with a human template, add in a few modifiers for particular genetic and developmental quirks, we get to a sort of blank human that gets us most of the way to restoring you after physical death. To complete the restoration, we need to inject the stuff that sets you uniquely apart even from your fellow humans who share your statistically regular quirks: your memories. If the record of your memories is good enough, this should effectively create a person who is so much like you as to be indistinguishable from the original, i.e. restore you.
But, you may ask, is this restoration of you from writing really still you in the same way that the you restored from cryonics is you? Maybe. To me, it is. Despite what subjective experience feels like, there doesn't seem to be anything in the brain that makes you who you are besides the general process of your brain and its memories. Transferring yourself from your current brain to another brain or a brain emulation via writing doesn't seem that much different from transferring yourself via neuron replacement or some other technique except that writing introduces a lossy compression step, necessitated only by a lack of access to better technology. Writing yourself into the future isn't the best solution, but it does seem to be an effective stopgap to death.
If you're still with me, we have a few nagging questions to answer. Consider this an FAQ for writing yourself into the future.
How good a record is good enough? In truth, I don't think we even know enough to get the order of magnitude right. The best I can offer is that you need to record as much as you are willing to. The more you record, the more there will be to work with, and the less chance there will be of insufficient data. It may turn out that you simply can't record enough to create a good restoration of a person from writing, but this is little different from the risk in cryonics of not being well preserved enough to restore despite best efforts. If you're willing to take the risk that cryonics won't work as well as you hope, you should be willing to accept that writing yourself into the future might not work as well as you hope.
How is writing yourself into the future more socially acceptable than cryonics? Basically, because people already do this all the time, although not with an eye toward their eventually restoration. People regularly keep journals, write blogs, write autobiographies, and pass on stories of their lives, even if only orally. You can write a record of yourself, fully intending for it to be used to restore you at some future time, without ever having to do anything that is morally or socially unacceptable to other people (at least, for people in most societies) other than perhaps specify in your writing of yourself that you want it to be used to restore you after you die.
How is writing yourself into the future more accessible to the poor? If a person is literate and has access to some form of durable writing material, they can write themselves into the future, limited only by their access to durable writing material and reliable storage. Of course, many people are not literate, but the cost of teaching literacy is far lower than the cost of cryonics, and literacy has other benefits beyond writing yourself into the future, so it's an easy sell to increase literacy even to people who are opposed to the idea of life extension.
Will the restoration really be me? Let me address this in another way. You, like everything else, are a part of the universe. Unlike what we believe to be true of most of the stuff in the universe, though, the stuff that makes up what we call you is aware of its existence. As best we can tell, the way that you are aware of your existence is because you have a way of recalling previous events during your existence. If we take away the store and recall of experience, we're left with some stuff that can do essentially everything it could when it had memory, but will not have any concept of existing outside the current moment. Put the store and recall back in, though, and suddenly what we would recognize as self-awareness returns.
Other questions? Post them and I'll try to address them. I have a feeling that there will be some strong disagreement from people who disagree with me about what self-awareness means and how the brain works, and I'll try to explain my position as best I can to them, but I'm also interested in any other questions that people might have since there are likely many issues I haven't even considered yet.