And fear not lest Existence closing your
Account, should lose, or know the type no more;
The Eternal Saki from the Bowl has pour'd
Millions of Bubbles like us, and will pour.

When You and I behind the Veil are past,
Oh, but the long long while the World shall last,
Which of our Coming and Departure heeds
As much as Ocean of a pebble-cast.

    -- Omar Khayyam, Rubaiyat



The typical argument for cryonics says that if we can preserve brain data, one day we may be able to recreate a functioning brain and bring the dead back to life.

The typical argument against cryonics says that even if we could do that, the recreation wouldn't be "you". It would be someone who thinks and acts exactly like you.

The typical response to the typical argument against cryonics says that identity isn't in specific atoms, so it's probably in algorithms, and the recreation would have the same mental algorithms as you and so be you. The gap in consciousness of however many centuries is no more significant than the gap in consciousness between going to bed at night and waking up in the morning, or the gap between going into a coma and coming out of one.

We can call this a "consequentialist" view of identity, because it's a lot like the consequentialist views of morality. Whether a person is "me" isn't a function of how we got to that person, but only of where that person is right now: that is, how similar that person's thoughts and actions are to my own. It doesn't matter if we got to him by having me go to sleep and wake up as him, or got to him by having aliens disassemble my brain and then simulate it on a cellular automaton. If he thinks like me, he's me.

A corollary of the consequentialist view of identity says that if someone wants to create fifty perfect copies of me, all fifty will "be me" in whatever sense that means something.


An argument against cryonics I have never heard, but which must exist somewhere, says that even the best human technology is imperfect, and likely a few atoms here and there - or even a few entire neurons - will end up out of place. Therefore, the recreation will not be you, but someone very very similar to you.

And the response to this argument is "Who cares?" If by "me" you mean Yvain as of 10:20 PM 4th April 2012, then even Yvain as of 10:30 is going to have some serious differences at the atomic scale. Since I don't consider myself a different person every ten minutes, I shouldn't consider myself a different person if the resurrection-machine misplaces a few cells here or there.

But this is a slippery slope. If my recreation is exactly like me except for one neuron, is he the same person? Signs point to yes. What about five neurons? Five million? Or on a functional level, what if he blinked at exactly one point where I would not have done so? What if he prefers a different flavor of ice cream? What if he has exactly the same memories as I do, except for the outcome of one first-grade spelling bee I haven't thought about in years anyway? What if he is a Hindu fundamentalist?

If we're going to take a consequentialist view of identity, then my continued ability to identify with myself even if I naturally switch ice cream preferences suggests I should identify with a botched resurrection who also switches ice cream preferences. The only solution here that really makes sense is to view identity in shades of gray instead of black-and-white. An exact clone is more me than a clone with different ice cream preferences, who is more me than a clone who is a Hindu fundamentalist, who is more me than LeBron James is.


There are various theories lumped together under the title "big world".

The simplest is the theory that the universe (or multiverse) is Very Very Big. Although the universe is probably only 15 billion years old, which means the visible universe is only 30 billion light years in size, inflation allows the entire universe to get around the speed of light restriction; it could be very large or possibly infinite. I don't have the numbers available, but I remember a back of the envelope calculation being posted on Less Wrong once about exactly how big the universe would have to be to contain repeating patches of about the size of the Earth. That is, just as the first ten digits of pi, 3141592653, must repeat somewhere else in pi because pi is infinite and patternless, and just as I would believe this with high probability even if pi were not infinite but just very very large, so the arrangement of atoms that make up Earth would recur in an infinite or very very large universe. This arrangement would obviously include you, exactly as you are now. A much larger class of Earth-sized patches would include slightly different versions of you like the one with different ice cream preferences. This would also work, as Omar Khayyam mentioned in the quote at the top, if the universe were to last forever or a very very long time.

The second type of "big world" is the one posited by the Many Worlds theory of quantum mechanics, in which each quantum event causes the Universe to split into several branches. Because quantum events determine larger-level events, and because each branch continues branching, some these branches could be similar to our universe but with observable macro-scale differences. For example, there might be a branch in which you are the President of the United States, or the Pope, or died as an infant. Although this sounds like a silly popular science version of the principle, I don't think it's unfair or incorrect.

The third type of "big world" is modal realism: the belief that all possible worlds exist, maybe in proportion to their simplicity (whatever that means). We notice the existence of our own world only for indexical reasons: that is, just as there are many countries, but when I look around me I only see my own; so there are many possibilities, but when I look around me I only see my own. If this is true, it is not only possible but certain that there is a world where I am Pope and so on.

There are other types of "big worlds" that I won't get into here, but if any type at all is correct, then there should be very many copies of me or people very much like me running around.


Cryonicists say that if you freeze your brain, you may experience "waking up" a few centuries later when someone uses the brain to create a perfect copy of you.

But whether or not you freeze your brain, a Big World is creating perfect copies of you all the time. The consequentialist view of identity says that your causal connection with these copies is unnecessary for them to be you. So why should a copy of you created by a far-future cryonicist with access to your brain be better able to "resurrect" you than a copy of you that comes to exist for some other reason?

For example, suppose I choose not to sign up for cryonics, have a sudden heart attack, and die in my sleep. Somewhere in a Big World, there is someone exactly like me except that they didn't have the heart attack and they wake up healthy the next morning.

The cryonicists believe that having a healthy copy of you come into existence after you die is sufficient for you to "wake up" as that copy. So why wouldn't I "wake up" as the healthy, heart-attack-free version of me in the universe next door?

Or: suppose that a Friendly AI fills a human-sized three-dimensional grid with atoms, using a quantum dice to determine which atom occupies each "pixel" in the grid. This splits the universe into as many branches as there are possible permutations of the grid (presumably a lot) and in one of those branches, the AI's experiment creates a perfect copy of me at the moment of my death, except healthy. If creating a perfect copy of me causes my "resurrection", then that AI has just resurrected me as surely as cryonics would have.

The only downside I can see here is that I have less measure (meaning I exist in a lower proportion of worlds) than if I had signed up for cryonics directly. This might be a problem if I think that my existence benefits others - but I don't think I should be concerned for my own sake. Right now I don't go to bed at night weeping that my father only met my mother through a series of unlikely events and so most universes probably don't contain me; I'm not sure why I should do so after having been resurrected in the far future.


What if the speculative theories involved in Big Worlds all turn out to be false? All hope is still not lost.

Above I wrote:

An exact clone is more me than a clone with different ice cream preferences, who is more me than a clone who is a Hindu fundamentalist, who is more me than LeBron James is.

I used LeBron James because from what I know about him, he's quite different from me. But what if I had used someone else? One thing I learned upon discovering Less Wrong is that I had previously underestimated just how many people out there are *really similar to me*, even down to weird interests, personality quirks, and sense of humor. So let's take the person living in 2050 who is most similar to me now. I can think of several people on this site alone who would make a pretty impressive lower bound on how similar the most similar person to me would have to be.

In what way is this person waking up on the morning of January 1 2050 equivalent to me being sort of resurrected? What if this person is more similar to Yvain(2012) than Yvain(1995) is? What if I signed up for cryonics, died tomorrow, and was resurrected in 2050 by a process about as lossy as the difference between me and this person?


Personal identity remains confusing. But some of the assumptions cryonicists make are, in certain situations, sufficient to guarantee personal survival after death without cryonics.

New Comment
140 comments, sorted by Click to highlight new comments since:
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

The only downside I can see here is that I have less measure (meaning I exist in a lower proportion of worlds) than if I had signed up for cryonics directly. This might be a problem if I think that my existence benefits others - but I don't think I should be concerned for my own sake.

Yvain is quantum suicidal? I didn't expect that! People whose preferences 'add up to normal' do (implicitly) care about measure.

Quantum suicide seems like a good idea to me if we know that the assumptions behind it (both quantum and identity-related) are true, if we're purely selfish (eg don't care about the bereaved left behind), and if we don't assume our actions are sufficiently correlated with those of others to make everyone try quantum suicide and end up all alone in our own personal Everett branch.

However, I might have the same "It's a good idea, but I am going to refuse to do this for reasons of personal sanity" reaction as I have with Pascal's Mugging.

Quantum suicide seems like a good idea to me if we know that the assumptions behind it (both quantum and identity-related) are true, if we're purely selfish (eg don't care about the bereaved left behind), and if we don't assume our actions are sufficiently correlated with those of others to make everyone try quantum suicide and end up all alone in our own personal Everett branch.

Fortunately, if you combine the second and third potential problems you end up with a solution that eliminates both of them. Then you just have the engineering problem involved in building a bigger death box.

However, I might have the same "It's a good idea, but I am going to refuse to do this for reasons of personal sanity" reaction as I have with Pascal's Mugging.

I hope so. Your position is entirely consistent - I cannot fault it on objective grounds and what you say in your post does directly imply what you confirm in your comment. That said, the preferences you declare here are vastly different to those that I consider 'normal' and so there remains the sneaking suspicion that you are wrong about what you want. That is, that you incorrectly extrapolate your volition.

On the other hand the ... (read more)

(Not sure where to put this:) Yvain's position doesn't seem sane to me, and not just for reasons of preference; attempting to commit suicide will just push most of your experienced moments backwards to regions where you've never heard of quantum suicide or where for whatever reason you thought it was a stupid idea. Anticipating ending up in a world with basically no measure just doesn't make sense: you're literally making yourself counterfactual. If you decided to carve up experience space into bigger chunks of continuity then this problem goes away, but most people agree that (as Katja put it) "anthropics makes sense with shorter people". Suicide only makes sense if you want to shift your experience backwards in time or into other branches, not in order to have extremely improbable experiences. I mean, that's why those branches are extremely improbable: there's no way you can experience them, quantum suicide or no.

Here is fine. This doesn't seem to be a problem that comes from being quantum suicidal but rather from an entirely different kind of anthropic based suicidal insanity. That is, I would not predict that experience as evaluated by Yvain's model of caring would be perceived this way. It certainly could be but that would be an additional insanity to the one that makes quantum roulette desirable. (No offense to Yvain and his Quantum Suicidal ilk by referring to this as 'insanity'. I mean only 'drastically different preferences preferences to my own in an agent similar enough to me that such comparison is meaningful'. In fact, if you're going to limit your optimisation to tiny amounts of measure then go ahead and exterminate humanity to maximise paperclips for all I care!) To expand somewhat: Quantum suiciding at (subjective) time t results in you at time t-1 having more measure than you at time t+1 but under default quantum-suicidal preferences these are in no way in competition. Relative measure between past and future selves isn't any particular issue. There are just various different subjective experiences at t-1, t and t+1, a desire to have each of them as positive-on-average as can be but no particular inclination to trim measure in one part of a timeline to increase it in another. For example, I wouldn't expect Yvain to (consider it rational to) commit conventional-and-complete suicide whenever it seemed like all his peak experiences are in the past and all that remained in life is to make the most of the remaining dregs.
You're not saying that if I perform QS I should literally anticipate that my next experience will be from the past, are you? (AFAICT, if QS is not allowed, I should just anticipate whatever I would anticipate if I was going to die everywhere, that is, going to lose all of my measure.)
(Not Will, but I think I mostly agree with him on this point) There is no such thing as an uniquely specified "next experience". There are going to be instances of you that remember being you and consider themselves the same person as you, but there is no meaningful sense in which exactly one of them is right. Granted, all instances of you that remember a particular moment will be in the future of that moment, but it seems silly to only care about the experiences of that subset of instances of you and completely neglect the experiences of instances that only share your memories up to an earlier point. If you weight the experiences more sensibly then in the case of a rigorously executed quantum suicide the bulk of the weight will be in instances that diverged before the decision to commit quantum suicide. There will be no chain of memory leading from the QS to those instances, but why should that matter?
2Scott Alexander
So, if Omega was willing to put a copy of you in an Everett branch that didn't already have one, how much money would you be willing to bid for this service? If Omega was going to charge $100, and the offer remained open for as many Everett branches as you wanted, how many $100s would you give Omega?
I'm not used to evaluating the worth of Everett branches by count. But for the purpose of this question may I assume you mean "another Everett branch of equal measure to this one, as of the time I click 'comment'"? As for an answer... um... I'm not sure, a fair bit? Working out my preferences - quantitatively - in situations so far outside the usual realm of operation is tricky. After I gave him everything I had I would get a new job that more closer matched my potential for financial gain. Two extra considerations: * Even aside from a terminal preference for measure maximization I would consider buying more measure purely for the purpose of giving me acausal bargaining power. (I'm even less sure about qualitatively evaluating the usefulness of acausal bargaining power.) * Buying more equal-measure branches is different to trying to preserve measure in the one we are in. While I think I have preferences such that I would buy a new one I'm not sure if the default behavior of humans would be to do so.
But would you take a partial amnesia pill before you go to sleep (which acts while you're not conscious and does or does not dissolve based on a quantum event)? As a partial quantum suicide. What about total amnesia? How much of your state has to be lost until its a suicide where QM will save you? Why won't QM save you from quantum-hangover where you do or don't get a terrible headache and IQ impairment of 20 points?
Just because. Really, that's all there is to it. It so happens that Yvain's (alleged) subjectively objective preferences are such that given the existence of a non-hangover branch the existence of a hangover branch is less desirable than if the hangover was substituted for death in that branch. This is not entirely unrelated to the natural outcome of 'average utilitarianism'. Under that value system the most simple solution is to go around killing the unhappiest person in the world until such time as the remaining people would be made on average less happy by the removal of the lowest happiness people. Sure, this is totally crazy. But if you have crazy preferences you should do crazy things.
Well, that was kind of rhetorical question; I want to make him be less into quantum-suicide by appeal to how (my guess) he's not expecting MWI to save him from experiencing quantum-hangover. edit: to explain, I myself believe death to be a real value from epsilon to 1 rather than a binary value, the epsilon being the shared memories remaining in other people, other people nearby who think like you, etc. and the 1 being the state where you are never forgetting anything at all, with living being somewhere close to 1, and amnesias somewhere between 0 and that. Having binary valued anything seem to result in really screwed up behaviours. (by other people nearby i mean within distances that are much smaller than your information content; the other people at distances that have as many bits in their relative coordinate as you have in your brain, it seems to me those really should be counted as substantially different even though I can't quite pin point why)
It was a question that got a straight answer that is in accord with Yvain's position. To the extent that such rhetorical questions get answers that do not require the rhetoric victim to change their mind, the implied argument can be considered refuted. In practical social usage rhetorical questions can, indeed, often be used as a way to make arguments that an opponent is not allowed to respond to. Here on lesswrong we are free to reject that convention and so not only are literal answers to rhetorical questions acceptable, arguments that are hidden behind rhetorical questions should be considered at least as subject to criticism as arguments made openly.
Well, it is the case that humans often strive to have consistent goal systems (perhaps minimizing their descriptive complexity?), so while he can just say 'because I defined my goal system to be this' he is also likely to think and try to come up with some kind of general principle that does not have weird discontinuities at some point when the amnesia is too much like death for his taste. edit: i think we are talking about different issues; i'm not making a point about his utility function, i'm making a point that he is expecting to become 'him 20 points dumber and with a terrible headache', who's a rather different person, rather than to become someone in a galaxy far far away who doesn't have the hangover and is thus more similar to him before the hangover.
Yvain does present a consistent goal system. It is one that may appear either crazy or morally abhorrent to us but all indications are that it is entirely consistent. If you were attempting to demonstrate to Yvain an inconsistency in his value system that requires arbitrary complexity to circumvent then you failed.
I think you're misunderstanding me, see edit. The point I am making is not so much about his values, but about his expectations of subjective experience.
Yvain's expectations of subjective experience actually seem sane to me. Only his values (and so expected decisionmaking) are weird.
Well, my argument is that you can propose a battery of possible partial quantum suicide set ups involving a machine that partially destroys you (e.g. you are anaesthetised and undergo lobotomy with varying extent of cutting, or something of this sort such as administration of a sublethal dose of neurotoxin). At some point, there's so little of you left that you're as good as dead; at some other point, there's so much of you left that you don't really expect to be quantum-saved. Either he has some strange continuous function inbetween, that I am very curious about, or he has a discontinuity, which is weird. (and I am guessing a discontinuity but i'd be interested to hear about function)

Isn't dissolving the concept of personal identity relatively straightforward?

We know that we've evolved to protect ourselves and further our own interests, because organisms who didn't have that as a goal didn't fare very well. So in this case at least, personal identity is merely a desire to make sure that "this" organism survives.

Naturally, the problem is defining in "this organism". One says, "this" organism is something defined by physical continuity. Another says, "this" organism is something defined by the degree of similarity to some prototype of this organism.

One says, sound is acoustic vibrations. Another says, sound is the sensation of hearing...

There's no "real" answer to the question "what is personal identity", any more than there is a "real" answer to the question "what is sound". You may pick any definition you prefer. Of course, truly dissolving "personal identity" isn't as easy as dissolving "sound", because we are essentially hard-wired to anticipate that there is such a thing as personal identity, and to have urges for protecting it. We may realize on an intellectu... (read more)

So if I want to improve the world, it makes sense for me to care about "my own" ... well-being - even though future instances of "me" are actually distinct systems ... because A) I care about the well-being of minds in general, and B) they share at least part of my goals, and are thus more likely to carry them out.

I think it's clear that there is also terminal value in caring about the well-being of "me". As with most other human psychological drives, it acts as a sloppily optimized algorithm of some instrumental value, but while its purpose could be achieved more efficiently by other means, the particular way it happens to be implemented contributes an aspect of human values that is important in itself, in a way that's unrelated to the evolutionary purpose that gave rise to the psychological drive, or to instrumental value of its present implementation.

(Relevant posts: Evolutionary Psychology, Thou Art Godshatter, In Praise of Boredom.)

It's not clear to me. To get us to behave selfishly, evolution could have instilled false aliefs to the effect that other people's mental processes aren't as real as ours, in which case we may want to just disregard those. Even if there's no such issue, there's not necessarily any simple one-to-one mapping from urges to components of reflected preference, especially when the urges seem to involve concepts like "me" that are hard to extend beyond a low-tech human context. (If I recall correctly, on previous occasions when you've made this argument, you were thinking of "me" in terms of similarity in person-space, which is not as hard to make sense out of as the threads of experience being discussed in this thread.)
Fair enough. I don't personally endorse it as a terminal value, but it's everyone's own decision whether to endorse it or not.

I don't personally endorse it as a terminal value, but it's everyone's own decision whether to endorse it or not.

I don't believe it is, at least it's relatively easy to decide incorrectly, so the fact of having (provisionally) decided doesn't answer the question of what the correct decision is. "It's everyone's own decision" or "everyone is entitled to their own beliefs" sounds like very bad epistemology.

I cited what seems to me like a strong theoretical argument for antipredicting terminal indifference to personal well-being. Your current conclusion being contrary to what this argument endorses doesn't seem to address the argument itself.

I thought that your previous comment was simply saying that 1) in deciding whether or not we should value the survival of a "me", the evolutionary background of this value is irrelevant 2) the reason why people value the survival of a "me" is unrelated to the instrumental benefits of the goal I agree with those claims, but don't see them as being contrary to my decision not to personally endorse such a value. You seem to be saying that the question of whether or not a "me" should be valued is in some sense an epistemological question, while I see it as a choice of personal terminal values. The choice of terminal values is unaffected by epistemological considerations, otherwise they wouldn't be terminal values.
Wait - what? Are you partly defining terminal values via their being unaffected by epistemic considerations? This makes me want to ask a lot of questions for which I would otherwise take answers for granted. Like: are there any terminal values? Can a person choose terminal values? Do choices express values that were antecedent to the choice? Can a person have "knowledge" or some closely related goal as a personal terminal value?
Nay, I don't think it is. I don't take issue with anything in particular you said in this comment, but it doesn't feel like a classic, non-greedy Reduction of the style used to reduce free will into cognitive algorithms or causality into math. The sense in which you can create another entity arbitrarily like yourself and say, "I identify with this creature based on so-and-so definition" and then have different experiences than the golem no matter how like you it is is the confused concept that I do not think has been dissolved; I am not sure if it a non-fake dissolving of it has ever even started. (Example: Susan Blackmore's recent "She Won't Be Me". This is clearly a fake reduction; you don't get to escape the difficulties of personal identity confusion by saying a new self pops up every few minutes/seconds/plancktimes. Your comment is less obviously wrong but still sidesteps the confusion instead of Solving it.) Hell, it's not just a Confusing Problem. I'd say it's a good candidate for The Most Confusing Problem. Edit (one of many little ones): I made this comment pretty poorly, but I hope the point both makes sense and got through relatively intact. Mitchell Porter's comment is also really good until the penultimate paragraph.
I tried responding to this example, but I find the whole example so foreign and confused in some sense that I don't even know how to make enough sense of it to offer a critique or an explanation. Why wouldn't you expect there to exist an entity with different experiences than the golem, and which remembers having identified with the golem? You're not killing it, after all.
I don't consider them something I should necessarily endorse. I consider them something that I do actually endorse because all things considered I want to. (Although given that the endorsement of such emotional considerations thereby makes them parts of my goals I suppose we could declare technically and tautologically that any emotions that go above and beyond what that which is incorporated into my goals should not be endorsed, depending on the definition of a few of the terms.)
Interesting. I was thinking about same regarding the early stages of AGI - it is difficult to define the 'me' precisely, and its unclear why would one have a really precise definition of 'me' in an early AGI. It's good enough if the life is 'me' to AI but the Jupiter isn't 'me' , that's a negligible loss in utility from the life not being kosher food for the AI.

This, along with the simulation argument, is why I'm not too emotionally stressed out with feelings of impending doom that seem to afflict some people familiar with SIAI's ideas. My subjective anticipation is mollified by the thought that I'll probably either never experience dying or wake up to find that I've been in an ancestral simulation, which leaves the part of me that wants to prevent all the empty galaxies from going to waste to work in peace. :)

Also, we can extend the argument a bit for those worried about "measure". First, an FAI might be able to recreate people from historical clues (writings, recordings, other's memories of them, etc.). But suppose that's not possible. An FAI could still create a very large number of historically plausible people, and assuming FAIs in other Everett branches do the same, the fact that I probably won't be recreated in this branch will be compensated for by the fact that I'll be recreated in other branches where I currently don't exist, thus preserving or even increasing my overall measure.

My subjective anticipation is mollified by the thought that I’ll probably either never experience dying or wake up to find that I’ve been in an ancestral simulation, which leaves the part of me that wants to prevent all the empty galaxies from going to waste to work in peace. :)

Update: Recent events have made me think that the fraction of advanced civilizations in the multiverse that are sane may be quite low. (It looks like our civilization will probably build a superintelligence while suffering from serious epistemic pathologies, and this may be be typical for civilizations throughout the multiverse.) So now I'm pretty worried about "waking up" in some kind of dystopia (powered or controlled by a superintelligence with twisted beliefs or values), either in my own future lightcone or in another universe.

Actually, I probably shouldn't have been so optimistic even before the recent events...

I updated downward somewhat on the sanity of our civilization, but not to an extremely low value or from a high value. That update justifies only a partial update on the sanity of the average human civilization (maybe the problem is specific to our history and culture), which justifies only a partial update on the sanity of the average civilization (maybe the problem is specific to humans), which justifies only a partial update on the sanity of outcomes (maybe achieving high sanity is really easy or hard). So all things considered (aside from your second paragraph) it doesn't seem like it justifies, say, doubling the amount of worry about these things.

I agree recent events don't justify a huge update by themselves if one started with a reasonable prior. It's more that I somehow failed to consider the possibility of that scenario, the recent events made me consider it, and that's why it triggered a big update for me.

1junk heap homotopy
Now I'm curious. Does studying history make you update in a similar way? I feel that these times are not especially insane compared to the rest of history, though the scale of the problems might be bigger.

Now I’m curious. Does studying history make you update in a similar way?

History is not one of my main interests, but I would guess yes, which is why I said "Actually, I probably shouldn’t have been so optimistic even before the recent events..."

I feel that these times are not especially insane compared to the rest of history, though the scale of the problems might be bigger.

Agreed. I think I was under the impression that western civilization managed to fix a lot of the especially bad epistemic pathologies in a somewhat stable way, and was unpleasantly surprised when that turned out not to be the case.

prevent all the empty galaxies from going to waste

(Off-topic: Is this a decision theoretic thing or an epistemic thing? That is, do you really think the stars are actually out there to pluck in a substantial fraction of possible worlds, or are you just focusing on the worlds where they are there to pluck because it seems like we can't do nearly as much if the stars aren't real? Because I think I've come up with some good arguments against the latter and was planning on writing a post about it; but if you think the former is the case then I'd like to know what your arguments are, because I haven't seen any really convincing ones. (Katja suggested that the opposite hypothesis—that superintelligences have already eaten the stars and are just misleading us, or we are in a simulation where the stars aren't real—isn't a "simple" hypothesis, but I don't quite see why that would be.) What's nice about postulating that the stars are just an illusion is that it means there probably isn't actually a great filter, and we aren't left with huge anthropic confusions about why we're apparently so special.)

do you really think the stars are actually out there to pluck in a substantial fraction of possible worlds

Assuming most worlds start out lifeless like ours, they must have lots of resources for "plucking" until somebody actually plucks them... I guess I'm not sure what you're asking, or what is motivating the question. Maybe if you explain your own ideas a bit more? It sounds like you're saying that we may not want to try to pluck the stars that are apparently out there. If so, what should we be trying to do instead?

I guess I didn't clearly state the relevant hypothesis. The hypothesis is that the stars aren't real, they're just an illusion or a backdrop put their by superintelligences so we don't see what's actually going on. This would explain the great filter paradox (Fermi's paradox) and would imply that if we build an AI then that doesn't necessarily mean it'll get to eat all the stars. If the stars are out there, we should pluck them—but are they out there? They're like a stack of twenties on the ground, and it seems plausible they've already been plucked without our knowing. Maybe my previous comment will make more sense now. I'm wondering if your reasons for focusing on eating all the galaxies is because you think the galaxies actually haven't already been eaten, or if it's because even if it's probable that they've actually already been eaten and our images of them are an illusion, most of the utility we can get is still concentrated in worlds where the galaxies haven't already been eaten, so we should focus on those worlds. (This is sort of orthogonal to the simulation argument because it doesn't necessitate that our metaphysical ideas about how simulations work make sense; the mechanism for the illusion works by purely physical means.)

The hypothesis is that the stars aren't real, they're just an illusion or a backdrop put their by superintelligences so we don't see what's actually going on.

If that's the case, then I'd like to break out by building our own superintelligence to find and exploit whatever weaknesses might exist in the SIs that are boxing us in, or failing that, negotiate with them for a share of the universe. (Presumably they want something from us, or else why are they doing this?) Does that answer your question?

BTW, I'm interested in the "good arguments" that you mentioned earlier. Can you give a preview of them here?

If the SI wanted to fool us, why would it make us see something (a lifeless universe) that would make us infer that we're being fooled by an SI?
It would seem it is trying to fool just the unenlightened masses. But the chosen few who see the Truth shall transcend all that...
Will, as per amit's point, how do you anticipate your decision to tell us about the superintelligent fake stars hypothesis the decision of the superintelligences to create (or otherwise cause to exist) human life on earth with the illusion of living in a free universe? All things considered (and assuming that hypothesis as a premise) I think you might have just unmade us a little bit. How diabolical!
I agree with your rationale, i.e. assuming they're actually around then the superintelligences clearly aren't trying that hard to be quiet, and are instead trying to stay on some sort of edge of influence or detectability. Remember, it's only atheists who don't suspect supernatural influence; a substantial fraction of humans already suspects weird shit is going on. Not "the chosen few" so much as "the chosen multitude". Joke's only on the atheists. Presumably if they wanted us to entirely discount the possibility that they were around then it would be easy for them to influence memetic evolution such that supernatural hypotheses were even less popular, e.g. by subtly keeping the U.S. from entering WWII and thus letting the Soviet Union capture all of Europe, and so on and so forth in that vein. (I'm not a superintelligence, surely they could come up with better memetic engineering strategies than I can.) Nitpick: they don't have to have chosen to create or cause us to exist as such, just left us alone. The latter is more likely because of game theoretic asymmetry ("do no harm"). Not sure if you intended that to be in your scope.
The intended scope was inclusive - I didn't want to go overboard with making the caveat 'or' chain mention everything. The difference in actions and inactions when it comes to superintelligences that are controlling everything around us including giving us an entire fake universe to look at become rather meaningless.
Let's hear them.
Forthcoming in the next year or so, in my treatise on theology & decision theory & cosmology & moral philosophy, which for better or worse I am going to write in English and not in Latin. But anyway, I'll say now that my arguments don't work for utilitarians, at least for preference utilitarianism the way it's normally considered. Even if the argument's valid it still probably doesn't sway people who are using e.g. the parliamentary meta-moral system and who give much weight to utilitarian intuitions.
Oh thank all the gods of counterfactual violence.
This feels very close to my moral intuitions, but I cognitively believe very strongly that things can be bad even if there are no observers left to regret that they happened. That follows logically from the moral intuition that people suffering where you can't see is still bad. that a natural moral intuition? I don't think so. That could explain the dissonance.

If anthropics makes any sense, then cryonics in a Big World is still controlling what you should mostly expect to experience. The anthropic escape valves for a version of me who experiences death and remembers not being signed up for cryonics range from Boltzmann brains to lunatics to completely random ancestor simulations, and I think I value experiencing the mainline expected cryonics outcome more highly than I do experiencing these.

And even if anthropics doesn't make sense, then as a matter of decision theory I value actually being around in a substantial fraction of the futures of the current world-state.

If all copies count as you, then that includes Boltzmann brains who die in the vacuum a second after their formation and copies of you who awaken inside a personally dedicated hell. And this is supposed to provide hope?

There is clearly a sense in which you do not experience what your copies experience. The instance of you who dies in a car crash on the way to your wedding never experiences the wedding itself; that is experienced by the second instance, created from a backup a few weeks later.

Any extension of identity beyond the "current instance" level is therefore an act of imagination or chosen affiliation. Identifying with your copies and almost-copies scattered throughout the multiverse, identifying with your descendants, and identifying with all beings who ever live, all have this in common - "you", defined in the broad sense supplied by your expansive concept of identity, will experience things that "you", defined in the narrow but practical sense of your local instance, will never experience.

Since it is a contradiction to say that you will experience things that you will never experience, it is desirable to perceive very clearly that these e... (read more)

It doesn't seem too much more distressing to believe that there are copies of me being tortured right now, than to believe that there are currently people in North Korea being tortured right now, or other similarly unpleasant facts everyone agrees to be true.

There's a distinction between intuitive identity - my ability to get really upset about the idea that me-ten-minutes-from-now will be tortured - and philosophical identity - an ability to worry slightly about the idea that a copy of me in another universe is getting tortured. This difference isn't just instrumentally based on the fact that it's easier for me to save me-ten-minutes-from-now than me-in-another-universe; even if I were offered some opportunity to help me-in-another-universe, I would feel obligated to do so only on grounds of charity, not on grounds of selfishness. I'd ground that as some mental program that intuitively makes me care about me-ten-minutes-from-now which is much stronger than whatever rational kinship I can muster with me-in-another-universe. This mental program seems pretty good at dealing with minor breaks in continuity like sleep or coma.

The problem is, once death comes into the picture, the menta... (read more)

Once you get resurrected, wont the mental program continue carrying business as usual and so wont the "me-ten-minutes-from-now" keep being there?
I don't understand why this is downvoted
"that is experienced by the second instance, created from a backup a few weeks later." What scenario is this referring to?

I like this post! Two comments barely related to the post:

  1. I would be interested in those calculations about how big the universe would have to be to have repeating Earths if anyone recalls where they saw them.

  2. A meta-LessWrongian comment: A great part of the value I get out of LessWrong is that there's always someone out there writing a post or comment tying together various thoughts and musings I've had into coherent essays in a way I don't have the mental discipline to do. So this is a big thanks to all of you writing interesting stuff!

And a comment more directly related to the post:

I can only assume that some other type of intelligence would be bemused by our confusion surrounding these issues in much the same way I'm often bemused by people making hopelessly confused arguments about religion or evolution or whatever.

Other Intelligence: "Of course continuity isn't important! There's no difference between waking up after a coma, being unfrozen, or your clone in a Big World! Sheesh, you humans are messed up in the head!"

(Or maybe OI would argue that there is a difference.)


I would be interested in those calculations about how big the universe would have to be to have repeating Earths if anyone recalls where they saw them.

If the universe is spatially infinite, then, on average, we should expect that no more than 10^10^29 meters away is an exact duplicate of you. If you're looking for an exact duplicate of a Hubble volume - an object the size of our observable universe - then you should still on average only need to look 10^10^115 lightyears. (These are numbers based on a highly conservative counting of "physically possible" states, e.g. packing the whole Hubble volume with potential protons at maximum density given by the Pauli Exclusion principle, and then allowing each proton to be present or absent.)

Google search query: "duplicate earth lightyears away". Estimated time to search and go through first page: 30 seconds.

This is simply the distance where the information content of the position is on same order that the information content of the brain for which the second instance is being 'found'. Essentially, infinite universe allows to encode brains into spatial positions.
Dangit! Pointing to a Google search is normally my modus operandi.
Or maybe: "All of those, including changing your preferences in icecream, constitute enough change to be considered death. It is theoretically possible to keep a human alive in a controlled environment but this has never been the case in the history of your species."
I wouldn't say a change in ice cream preference constitutes a death. People's taste buds change as they get older. Or if you do say getting older is a younger person's death, then it probably isn't that strict definition of death that you're concerned about, as you don't mourn people getting older.

If you don't care about measure then why try to solve friendly AI? In a VERY BIG world some AI's will turn out to be friendly.

5Scott Alexander
I'm uncertain of many-worlds, Big World theory, and the reductionist view of identity as not linked to a particular body. If any of those fail, then if this body dies I'm just plain dead, and same for everyone else. And there are many ways that an unfriendly AI - or no AI at all - could screw things up besides killing everyone instantly.
I understand and agree. I am curious. What are you more confident of, MWI and reductionism or unfriendly AI according to SI? As far as I am aware, the Born probabilities are about the only big uncertainty when it comes to MWI. When it comes to reductionism, the only real puzzle that I can think of is how one could possible build a conscious machine in the Game of Life. It would be really weird to look at the board and say, this 10000x10000 section of the board is consciousness, but the 10x10 section to the left isn't, nor is this 10k x 10k section of noise. Ain't there many more uncertainties when it comes to the urgency and possibility of risks from superhuman intelligence? ETA My question was triggered by your statement that you are uncertain of many-worlds and reductionism. It seems that people more often proclaim their uncertainty about relatively straightforward ideas like reductionism or cryonics than risks from AI. Which seems weird. Skeptic: I am uncertain if MWI is the correct interpretation and if cryonics is going to work. Advocate: Yes, there are some valid objections. Especially MWI is a very complicated subject that a layman is seldom able to judge. But I think there is more that speaks in favor of those ideas than against. vs. Skeptic: I am uncertain if AI could undergo explosive recursive self-improvement and take over the universe. Advocate: The uncertainty isn't justified. It is straightforward that it will happen soon. The idea is based on years of disjunctive lines of reasoning.
How many people fit the second category to that level of certainty? I consider myself an advocate, yet I would fall under the skeptic label given those phrasings.

It should be mentioned that when considering things like Cryonics in the Big World, you can't just treat all the other "you" instances as making independent decisions, they'll be thinking similarly enough to you that whatever conclusion you reach, this is what most "you" instances will end up doing. (unless you randomize, and assuming 'most' even means anything)

Seriously, I'd expect people to at least mention the superrational view when dealing with clones of themselves in decide-or-die coordination games.

I have a feeling that I'm missing something and that this is going to be downvoted, but I still have to ask. In the event that a big universe exists, there are numerous people almost exactly like me going about their business. My problem is that what i would call my consciousness doesn't seem to experience their actions. This would seem to me like there is some factor in my existence that is not present in theirs. If a perfect clone was sitting next to me, I wouldn't be able to see my computer through his eyes. I would continue to see it through mine. This chain of experience is the thing I care most to preserve. I have interest in the continued existence of people like me, but for separate reasons.

I know the idea of an "inner listener" is false, but the sensation of such a thing and a continuous stream of experience do exist. I am emotionally tied to those perceptions. I don't know how enthusiastically I can look forward to the future if I won't be able to experience it any more than I can the nearest parallel universe.

This chain of experience is the thing I care most to preserve.

Okay, think of it this way.

You go to sleep tonight, your "chain of experience" is briefly broken. You wake up tomorrow morning, chain of experience is back, you're happy.

But what makes you say "chain of experience is back"? Only that a human being wakes up, notices it has the memories of being pleeppleep, and says "Hey, my chain of experience is back! Good!"

Suppose Omega killed you in your sleep, then created a perfect clone of you. The perfect clone would wake up, notice it has the memories of being pleeppleep, and say "Hey, my chain of experience is back! Good!" Then it would continue living your life.

Right now you have zero evidence that Omega hasn't actually done this to you every single night of your life. So the idea of a "chain of experience", except as another word for your memories, is pretty tenuous.

And if I told you today that Omega had really been doing this to you your whole life, then you would be really scared before going to sleep tonight, but eventually you'd have to do it. And then the next day, your clone would still be pretty scared before going to... (read more)

I suppose you're right. If A clone of me appeared next to me he would also think it curious that he could not experience exactly the same sensations as the identical person next to him. Still, I don't think that a "me" from another world coming into existence and living through an apparently identical series of events to me life would count as a resurrection, as the events he lives through are not the same events that I live. In order for it to be a resurrection, I would have to return with the specific set of memories I have from this section of causality. It would be more like reincarnation than anything. Also I have evidence that I have not been intentionally deleted every night because I have zero evidence that this has occurred.
What if there is a perfect clone of you not sitting next to you, but sitting in another part of the universe that is atom-by-atom identical to this one throughout the span of a 46 billion light-year diameter sphere (the size of our observable universe) whose history has been identical to ours since the big bang. Every thought you have had over the course of your entire life this clone has had too. And what's more, they've had them for the exact same reasons. The chain of events leading up to each of their thoughts is identical to the chain for yours. Suppose in fact that there are billions such clones (or even infinitely many) spread (unimaginable far apart) all across our giant universe. A moment from now due to quantum events the copies' experiences will diverge. Some will experience one sensation, and some another, so at that point only a fraction of them continue to have histories identical to yours. But can you say which one of them you will be? If there is no way to distinguish one copy from another, then perhaps it makes sense to say that they are all the same thing -- so that all of the clones with identical histories to you are together 'you' right now, but as their experiences diverge going forward, they split into different possible future versions of you. What do you think?
That's all true, but I think the fact that I can't see more than one of their points of view is enough to distinguish between them. The only difference I can think of that pertains to a reductionist universe is location. I know the "inner listener" to be an illusion, but still can't shake the fact that if I died in this universe, but not in another one, then I wouldn't experience "waking up". Its possible that I'm being irrational and misinterpreting the way the brain works, but I think it is clearly observable that I don't feel that I'm experiencing more than one universe, and this feeling is the thing that I care most to preserve.
I think you're missing the part where "their points of view" are exactly the same. What would it mean to see more than one of them when they're exactly the same. Are you picturing them lined up next to each other in your field of view so you can count them? Similarly there is no "I just definitely died" feeling that we know of. (How would we know?) You shouldn't picture "dying and then waking up in another universe." You should picture "I experience passing out knowing I may die, but that there is a least one of me that probably doesn't. So when I wake up it will turn out that I was one of them." Does this make more sense? I think the barrier to intuition is in just how indistinguishable," indistinguishable" is. You can be a billion exact copies and you'll never notice, because they're exact.
I meant that the "me" in a different universe is different from me in this one. The distance between universes is not trivial. I might never notice the difference between a million "me"s and a billion, but the overall number of "me"s is significant. If multiple versions of myself live side by side, and one dies, then that one does not really continue living, unless it i replaced. Does that make sense? Its not very easy to word ideas regarding this topic.
I suppose you mean they have different positions. But if indistinguishable particles in quantum mechanics can freely switch places with each other whenever, and which is which has no meaning, then what argument do you have that the universe can even keep different versions of you apart itself? Not very formal, but I'm trying to convey the idea that certain facts that seem important have no actual meaning in the ontology of quantum physics.

First, I wish people stopped using this untestable Big World nonsense in their arguments.

For things to "add up to normality", your decisions should not be affected by a particular interpretation of QM, and so you should arrive to the same ones by sticking to the orthodox interpretation (or any other). If your argument fails without invoking a version of the MWI, it is not a sound argument, period. Similarly, your belief that in a galaxy far far away you are a three-eyed Pope named LeBron should not affect your decision of whether to sign-up for cryonics here and now.

Your other point is eminently worthwhile: since the cryonic resurrection is manifestly not exact, how much deviation from the original are you prepared to allow while still considering the resurrected object to be you for practical purposes? The answer is not objective in any way. For some losing a single memory is enough to say No, for others retaining even a small fraction of memories is enough for a Yes. The acceptable range on emotions, volitions and physical makeup can also vary widely.

Big World theories are derived from existing theories of mainstream physics, which are not generally considered untestable.
You are using the world "derived" rather loosely. There are untested but suggestive speculations in the string theory (which in itself is not what one would call a sound physical model), and most models of the universe being significantly larger than some 10^15 light years in whichever direction are also speculative. Certainly there are no indication of a size large enough to duplicate our corner of it in any kind of detail.
Hmm. I was under the impression that Big World theories would be relatively accepted. At least Bostrom and Tegmark seem to argue as if they were: Self-Locating Belief in Big Worlds: Cosmology’s Missing Link to Observation (Bostrom 2002): Parallel Universes (Tegmark 2003): Though those papers are from 2002 and 2003 - have the theories in question been disproven since then? If so, I'd be curious to read about it.
It is accepted that the Universe is likely much bigger than what is visible. There are no indications that it is infinite or even large enough to ensure the Big Worlds-type recurrence. My point is that your decision of whether to sign up for cryonics now should not depend on whether the universe is 10^10 (not big enough for recurrence) or 10^10^10 times larger than what we can presently see.

Something oddly relevant: quantum insomnia:

Of course that guy not being me or you, it is obvious that he is not a quantum insomniac, but there's a rather creepy thought: Today may be the day that you woke up for the last time in your life, and you'll remain awake forever, just because there will always be branch where you are awake, and you after a night's sleep is more different from you falling asleep, than you one second later, still awake. Good night. On second thought, I probably shouldn't joke like this here but fo... (read more)

Looks like my idea wasn't as original as I thought.

How many bytes in human memory? is a very brief article providing estimates of just that. Evidence from human learning experiments suggests that, after using a very good data compression algorithm, human long term declarative memory holds only a few hundred megabytes.

How much of that information is common knowledge, such as knowledge of the English language, memories from media such as books or television, or knowledge of local buildings and streets, is unclear.

Additional information specific to an individual could be gained from email, internet posts, and... (read more)

The simplest is the theory that the universe (or multiverse) is Very Very Big.

Do you mean this in an Occamian way? I suspect not, but I think you should make it clearer.

Anyway, this is a subject I've actually thought about a lot.

A lot of people (including Derek Parfit himself) think continuity is absolutely essential for personal identity / selfhood, but I don't. I've had such terribly marked psychological changes over the years that I cannot even conceive of the answer to the question, "Am I the same person as Grognor from 2009?" being yes in... (read more)

Yes. Reading LessWrong has nearly convinced me that I don't exist ...
You don't.
I dunno... I feel being essentially same as me from 18 years ago. I have terabytes of that person's memories, at any rate, and no-one else does until you go so far from here as to start encoding those memories into the spacetime coordinates. The personality changes may very well be down to some change to levels of hormones; few hundred bits worth of change, some terabytes worth of extra memories, and learned skills. That should make me the closest match, much closer than anything else. There's continuity to data inside your computer; you write an essay, you copy it around, you edit it, the file system may move file around to close gaps, you may store it in google docs which will store it in really odd ways and move it around outside your knowledge, but there is a causal chain.
If your personhood definition requires there to be only one of you, I think it already fails. You have not made a case that a causal chain is necessary or even a little bit relevant. I don't think it is. If an algorithm appears twice at distances so disparate they could not possibly be causally related, they're still the same algorithm. Your comment is one of those philosophical confusions that, unbeknownst to its author (that'd be you), puts its conclusion in its premise.
The full quote was, "and no-one else does until you go so far from here as to start encoding those memories into the spacetime coordinates." . Don't quote out of context. It is a physical fact of reality that within ridiculously huge volume, I am the only one who remembers what I thought about on a train ride in the year 2000 or so on the way back home from work, err, i mean school. It's only when you get extremely far, where the memory of this and other thoughts can be encoded into the coordinate, that you begin to see instances that hold this memory. (subject to there being a zillion of mes comparatively 'nearby' if the MWI is true, of course) edit: and okay, i was thinking about electrical aircraft propulsion, using electric arc to accelerate the air. I can remember in vivid details what I imagined then because I committed it to memory (the train was packed and I had to stand, so I couldn't write it down). I have terabytes of such stuff in the head, and while someone on earth can have approximately similar memories, the edit distance is huge.

You know, Stross tacitly considered an interesting form of resurrection in Accelerando--a hypothetical post-singularity (non-Friendly) AI computes a minimum message length version of You based off any surviving records of what you've done or said (plus the baseline prior for how humans work) and instantiates the result.

I'm having real trouble proving that's not more-or-less me, and what's more, that such a resurrection would feel any different from the inside looking back over its memories of my life.


I still have yet to see anyone adequately address how I am supposed to relate in any way to the magical copy of me a universe away.

If they have a shitty day, I feel nothing. If they have a good day, I feel nothing. If they die, big whoop. When I die, I will not be magically waking up in their body and vice versa. I will be dead.

Can you relate to the person who will remember having been you in twenty years? If so, how do you do that?
The same way an instance of a class is still that instance even if I change some of it's private variables. Now, the other instances of the same class are still quite different, and are not the same, even though they are similar.
Well, as long as we're switching to metaphor, the way I as an instance of class C relate to a separate instance of C is through our shared inheritance of C's methods, defaults, and other attributes. Is that enough relationship for me to care? Well, it depends. I mean, I don't necessarily care about my self in twenty years either... even if I know he's going to have a shitty day in 2032, or a good day, etc., I feel nothing. Sometimes it's enough to believe that I will feel something. Sometimes it isn't. In much the same way, I don't necessarily care about other people, but sometimes I do.
Sure. I would argue that me in 20 years is the same instance as the current Sly, but older, while the parallel universe Slys are just other instances of the Sly Class. Hence why I can easily differentiate between the two.
Of course you can differentiate between them. I can differentiate between me-in-five-years and me-in-twenty-years, as well. There exist differences between these things. I initially thought you were asking how one could identify with a different-but-similar person, given the absence of feedback ("If they have a shitty day, I feel nothing. If they have a good day, I feel nothing."). With respect to that, it seems to me that my ability to identify with myself-in-the-future despite the lack of feedback suggests that the lack of feedback isn't a showstopper, and more generally that what I identify with is more a function of my capacity for empathy than it is of any "me"-ness in the world. I'm no longer sure I understood you correctly in the first place, though.
And for that matter, I don't really have that much feedback from me in 20 seconds from now, or me 20 seconds ago either. My current remembering self has instantaneous inclinations, some of which are predicated on memories or anticipations, but at no point am I ever really a smearing of multiple time slices of myself. (I am probably a smearing of different quantum branches of myself, though. Until those selves decohere and I incrementally discover which branch that "I" have been on "all along"). For example, what is the difference between what we would commonly call "me", and an entity whose conscious experience is the Heaviside function with argument equal to the entire description of my brain state right as I type this question mark? That version of me just started existing, but luckily had molecules and quarks in all the right places to feel and remember everything as if he's been alive for 26 years and ate tofu for dinner.
Well, in practical terms, the anticipations matter. I expect decisions I make now to affect the state of me-in-twenty-seconds; if I jump off a tall building, for example, I expect me-in-twenty-seconds to be a smear on the pavement, so if I value me-in-twenty-seconds not being a smear on the pavement, that inclines me not to jump off a tall building. But no such relationship exists between me and me-twenty-seconds-ago; I don't expect decisions I make to affect the state of that entity. You are of course correct, though, that my anticipations are facts about my minds and not facts about reality-other-than-my-mind, which might have all kinds of properties that make my anticipations (and recollections, and current perceptions and beliefs) simply false.
Yeah, it was your last paragraph that I was meaning. In a given moment, I don't value something about me 20 seconds from now. I value the current experience of thoughts that involve simulations about an idealization of me extrapolated in time. The thing I am valuing are immediate thoughts though. Much like altruistic values being rooted in your own immediate value of anticipations. My meat computer will act on observations to induce an anticipation of X in my brain. If I want to anticipate Y then I should do Z to bring about the actions that lead to the anticipation of Y. Once I am at the precise instant that I'm experiencing Y, I'm not valuing Y because my mind is valuing anticipations post-Y. It is very interesting given things like closed-timelike curves and so on that there is no anticipation of past selves. It would be great to see a write up of why the perceived flow of entropy causes me to only future-value an anticipation like being proud of my former actions or accomplishments. I'm sure that the right level of articulation is evolutionary biology. I don't see how having visceral cognitive anticipations of the past could be adaptive. But it's still interesting. And even more interesting to think that there is some most-like-me entity within the subspace of entities that do have past-looking anticipations, probably wondering why people don't have future-looking anticipations right now (in a Big Universe, anyway)
The point I was trying to convey was that there is a huge difference between if I die vs a copy elsewhere of me dying. Hence me using the programming metaphor. If I am an instance, it matters a hell of a lot if it is this particular instance that gets scrapped vs some other instance of Sly. My argument is that ageing is more like modifying the variables, whereas the Big Universe copy of sly is a separate instance. Therefore it makes a lot of sense why I do not consider the copy of Sly to be me. I do not equate the two as other people here really want too. I also reject the idea that ageing identity loss is comparable to death identity loss, this seems to be completely wishful thinking.
You started out by asking how you were supposed to relate in any way to a copy of you . What I'm gathering from our subsequent discussion is that this was a rhetorical question; what you actually meant to express was that you don't relate in any way to such a copy, and you don't feel obligated to. I accept that you don't, and I agree that you aren't obligated to. Identity in the sense we're discussing is a fact about the mind, not a fact about the world. If you choose to identify solely with your present self and future selves in the same body, and treat everything else that could conceivably exist as not-you, that's fine. It's a perfectly reasonable choice, and I've no doubt that you can come up with lots of perfectly valid arguments to support making that choice. The fact that other people make different choices about their identity than you do doesn't mean that either of you is wrong about what your identity "really" is, or that either of you is ignoring reality in favor of "wishful thinking". There are consequences to those choices, of course: if I choose to identify with me-now but not with me-in-ten-years, for example, I will tend to make decisions such that ten years from now I am worse off. If I choose to identify with me-in-this-body but not copies of me, I will tend to make decisions such that copies of me are worse off. (Obviously, this doesn't actually create consequences in cases where none of my decisions can affect copies of me that exist.) Etc.
Yes, it was rhetorical. I see now, I was operating on the fact about the world level.

Is the amount of bits necessary to discriminate one functional human brain among all permutations of matter of the same volume greater or smaller than the amount of bits necessary to discriminate a version of yourself among all permutations of functional human brains? My intuition is that once you've defined the first, there isn't much left needed, comparatively, to define the latter.

Corollary, cryonics doesn't need to preserve a lot of information, if any, you can patch it up with, among other things, info from what a generic human brain is, or better wh... (read more)

suppose that a Friendly AI fills a human-sized three-dimensional grid with atoms, using a quantum dice to determine which atom occupies each "pixel" in the grid. This splits the universe into as many branches as there are possible permutations of the grid (presumably a lot)

How is that a Friendly AI?

Once again, Nick Bostrom's Quantity of experience: brain-duplication and degrees of consciousness comes to the rescue. Cryonics greatly expands the proportion of "you-like" algorithms as opposed to "!you-like" algorithms, for the same reason that quantum russian roulette greatly shrinks that proportion.

Pfsch, silly. I couldn't wake up as someone from another part of the universe - he's already busy waking up as himself :P

Several of your examples are also equivalent to quantum suicide situations. Similar comments about measure apply, except in this case we have a process (cryonics) that actually can restore measure (therefore, we can ignore all differences from our intuitive idea of resurrection).

See Tegmark's "Multiverse Hierarchy": Also there has been good work done showing that a spatial Big Universe is a realization of the quantum multiverse in 3 spatial dimensions.


Pretty ok article.

It has been years since I've last thought about personal identity. The last time it seemed a pretty reasonable and obvious conclusion to slightly less value the "me" stored in my body and value humans who where similar to me a bit more.

There seemed to be little point in (ceteris paribus) me being willing to spending more to save "my" life compared to saving the life a of a random average human A and not also expending at least something extra on person B who is say half way in "meness" between said random ... (read more)

Every man's death diminishes you, because you are involved in all mankind?
Well yes to a very small extent. However when I get to very small amounts of me I don't give away any extra effort, much for the same reason animals responding to the pay offs of kin selection don't: (p. 290, The Selfish Gene by Richard Dawkins)

I'm signed up for cryonics, and I believe you're leaving out two crucial motivations for cryonics. These are not particularly smart reasons. You don't logic your way into them. But they are emotionally strong reasons, and emotions are the primary motivators of humans.

The first is that it reduces my fear of death significantly. Perhaps unreasonably so? But the fear of death had been a problem for me for a long time, and now not as much.

The second is that I want OTHERS to be frozen for MY sake. I want to see my parents and brothers and friends again in this ... (read more)

As I've said elsewhere, I mostly think that any notion of preserving personal identity (from moment to moment, let alone after my heart stops beating) depends on a willingness to acknowledge some threshold of similarity and/or continuity such that I'm willing to consider anything above that threshold to share that identity.

Defining that threshold such that entities in other universes share my identity means there's lots of me out there, sure. And defining that threshold such that entities elsewhere on Earth right now share my identity means there's lots of me right here.

It's not clear to me why any of that matters.

This is an interesting argument that I have been giving a lot of thought to lately. Are all randomly occuring me-like entities out there in the big universe just as much me as the ones I anticipate in the future?

Well, there's one difference: causality. I'm not so sure a me with no causal relationship to my current self is something I can justifiably consider a future me. When I step in a teleporter that vaporizes my molecules and reconstructs me, I still have a very well-defined causal relationship with that future self -- despite the unorthodox path by wh... (read more)


Whenever I think about anthropics I'm always worried that I'm only experiencing thinking about anthropics because I'm thinking about anthropics, as if anthropics itself is the domain of some unfathomably powerful god who can change relative measures at whim and who thinks it's really fun to fuck with philosophers.

[This comment is no longer endorsed by its author]Reply

There may be another alternative to cryonics that doesn't require a Big World - indirect mind uploading (scroll down to "Using ‘indirect mind uploading’ to avoid staying dead"). The idea is that if you record every second of your life (e.g. on video), using this information a future AI might be able to converge on a specific brain configuration that is close to your original brain. Since only certain brains would say, write, do or think (although we currently can't record thoughts) the things you recorded yourself saying, writing and doing, depen... (read more)

Right now I don't go to bed at night weeping that my father only met my mother through a series of unlikely events and so most universes probably don't contain me; I'm not sure why I should do so after having been resurrected in the far future.

Because you understand that you can't change it:

For nothing is more certain, than that despair has almost the same effect upon us with enjoyment, and that we are no sooner acquainted with the impossibility of satisfying any desire, than the desire itself vanishes.

I don't think this claim is true in all cases, ... (read more)

4Scott Alexander
As I said below, if Omega offered me a financial deal to increase my measure - $100 per alternate Everett branch to make my father and mother meet and conceive me - this wouldn't seem like a remotely good use of my money.

Parallel copies of me are not me. Dying in X% of Everett branches has X% of disutility of dying.

Gradual change feels OK. Less gradual change feels less OK (unless I percieve it as an improvement). Going to sleep at night and knowing that in the morning I will feel a bit differently already makes me nervous. But it's preferable to dying. (But if I could avoid it without bad consequences, I would.) Small changes are good, because more of the future selves will be more similar to my current self.

How exactly does one measure the similarity or the change? Some ... (read more)

Let us suppose you are forced to play a quantum roulette (assume the payoffs are along the lines of those described here). The next day, someone asked you whether you were glad that you were forced to play quantum roulette. Do you answer: * NO! I just 50% died! Those @%#%@ assholes! * Yes! I just got $300k for free! I ask because while from the perspective of evaluating expected payoffs in the future your two assertions are compatible but from the perspective of evaluating outcomes that have happened they are not.
I guess my answer would be like: "It's too bad they 50% killed me... but now, I'm not going to cry for my parallel dead bodies (they are a sunk cost) and I'll enjoy the money." So I would be both happy that I am in the winning branch, but also aware of the cost, so I would not be retrospectively happy about being forced to play. Does this make sense? I would be both happy and unhappy about two different aspects of the situation. The part that makes me worry about the death of non-me's is that they were killed by something that was a threat for me too. (Something like when terrorists capture 16 prisoners and kill 15 of them and you are the one they release, and then somehow illogically they also pay you a lot of money. They did not kill you, and you even profited from the action, but that was not a personal decision on their side, just a random choice. So in some sense, they wanted to kill you too, and almost succeeded.)
Thanks, the illustration with the terrorists nailed the meaning down.
The answer is that I just had a 50% chance of dying. Assholes. This should be pretty obvious considering the pile of corpses the quantum murderers are leaving behind as they repeat their game.
That actually isn't an answer to the clarification I asked of Viliam. If you (are sane and so) consider quantum roulette undesirable then naturally you consider the folks who forced it upon you to be assholes. Yet you are the (measure of the) guy that won the lottery, didn't die and got the $300k. If Viliam considers parallel copies not-me then after the coin is tossed he doesn't care (in the direct personal sense) about the other 'not-me' guy who lost and got killed. Mind you the language around this subject is ambiguous. It could be that Viliam's expression was intended to place parrallel-everett-branch selves into a qualitatively different class to other forms of parallel selves.
I see what you are saying now. Thanks for clarifying.
Glad to hear it. I hope my inclusion of parenthetical 'sanity' claims conveyed that I essentially agree with what you were saying too.
Yea, that helped.
Curiously, for all the enormous scope of the higher level Tegmark multiverses this isn't necessarily the case. The evaluation of utility is determined by an extrapolation from you. If you are the kind of person that does not have an unbounded utility function it is entirely possible that "heaven" does not exist even Tegmark's level IV ultimate ensemble. It would require going beyond that, to universes that aren't mathematically possible.

Well, those copies, they are very very far away, and thus are more different from you than something nearby; it feels you are using the intuitions for things nearby, whose relative position is only a minuscule fraction of their total information content, for the things whose information content is entirely their position information. In principle we can recreate this conversation by iterating every value of every letter on a very powerful computer, but unless there's process for selecting this conversation out of the sea of nonsense, that won't constitute a backup.

Yes, to the degree that you accept the existence of a Big World, together with the usual assumptions about personal identity, you should expect never to die.

Even if there is no Big World, however, no one will ever experience dying anyway. Your total lifespan will be limited, but you will never notice it come to an end. So you might as well think of that limited span as a projection of an infinite lifespan onto an open finite interval. So again, one way or another you should expect never to die.



[This comment is no longer endorsed by its author]Reply

I'm curious how we distinguish copies of ourselves and near-copies of ourselves from ourselves. I mean, this intuition runs strongly through personal identity discussions: in this post you are identifying possible candidates for "clone of me", not "me". If it was the latter, we'd just go looking for ourselves in the Big World, which is a much simpler problem: "I'm me, and there's 863 of me here and there, some brighter than others." No concerns for them needing to be closer to some canonical source.

But we keep dragging in this... (read more)

Well, for starters, would you distinguish between a "more me" function and a "similar to me" function? Because there's been a lot of cogsci research done on what underlies human similarity judgments, so that would be a fine place to start, I suppose.

Why do you care if you continue to exist?

Also, if you don't care about your measure, than why is the Big Universe even necessary? You already know you have a measure of at least how long you've been alive. You can't cease to have ever existed.

1Scott Alexander
The main thrust of this post is that assumptions sufficient to support cryonics are also sufficient to survive without cryonics. The "why do you care if you continue to exist?" question is just part of the background assumption - people sign up for cryonics because they care about continuing to exist. All I'm saying is that if you believe that, you can get the same results by not signing up for cryonics, maybe.

This is a restatement of quantum immortality, right?

[This comment is no longer endorsed by its author]Reply
0Scott Alexander
I'm not sure why you crossed that out; it more or less is. The only new part is that I've never heard anyone discuss the relevance to cryonics before.

But this is a slippery slope. If my recreation is exactly like me except for one neuron, is he the same person? Signs point to yes. What about five neurons? Five million? Or on a functional level, what if he blinked at exactly one point where I would not have done so? What if he prefers a different flavor of ice cream? What if he has exactly the same memories as I do, except for the outcome of one first-grade spelling bee I haven't thought about in years anyway? What if he is a Hindu fundamentalist?

These questions apply equally to the person who wakes u... (read more)

Since I inherently desire to struggle for life regardless of whether or not my efforts will have any effect this argument does not alter my motivations or the decisions I will make. I'm perfectly fine with struggling for a lost cause if the process of struggling is either valuable or inevitable.

In a Big World, the process of struggling is all that we have, and success doesn't matter so much.