Karl Popper is a truly wonderful author whose work is consistently misunderstood. For instance, people mistake the “falsification boundary” as that which separates science and pseudoscience, when in fact old Poppy-pass-the-biscuits was quite explicit that it was the line of demarcation between what can be called “empirical” or “naturalistic” science, and metaphysics.
Excuse a mere SF author with a background in cognitive behavioural therapy from intruding on an erudite blog, but I’ve been studying as best I can quantum mechanics trying to get my head around what appears to be the problem, which given my maths is not up to doing the calculations has taken me some time.
I love the idea of Everett, and Tegmark mathematical universe, but the best explanation of the qm problem is I think the presentation by Ball, and when added to the double-eraser experiment which adds some much needed clarity to QM whoohoo by showing that entanglement is measurement.
So, after hours of listening to the pros and cons, I think that Garret’s argument that there isn’t a multiverse convinces me with how he describes that the universe we live in isn’t real either (as in qm requires non-locality). So a system with multiple probabilities leads to a universe that is one and zero.
What I find intriguing as an SF author is that there are still multiple versions of the universe because the speed of light make everything a relationship between observers and therefore each can have a different perspective on what is in the past and what is in the future.
This is my my very long winded way of saying your meme needs more work. Still, it’s a good meme.
My program’s translation in to 12 Year Old Amerikan English Circa 2011: this is a shiitake mushroom post! Dr. Scott Aaronson is a troll engaged in interconnected network ideology culture warrior keyboard battling!
I keep wondering if I should take the time to persuade you of MWI. Maybe after announcing our fated meeting in advance, AI-Box style, and having you sign an agreement to only reveal the result, and not the argument that persuaded you, until after an undisclosed-to-others waiting period. Not for any important reasons, just because I suspect lots of people would freak out if I did that.
Eliezer Yudkowsky #10
Isnât it a cop-out to believe that, in another universe, Putin is a vegan who spends his time tending to his vegetable garden, when he is not cuddling his pet kittens and puppies?
Eliezer Yudkowsky #10: Sure, I’m happy to have another QM discussion—should I email you the next time I’m in Berkeley?
I do, however, already speak the MWI-language pretty fluently, and even prefer it more often than not (while also being able to translate to and from Bohm-language, Copenhagen-language, etc). I already teach MWI in my undergrad quantum information class, in such a way that according to the poll we give at final exam time, roughly half the students end up as MWI proponents (with the others split among Bohm, Quantum Bayesianism, Penrose-style dynamical collapse theories, agnosticism, and rejection of the whole question as meaningless).
It’s just that, to use your language (as opposed to, say, Popper’s), I still have difficulties with making a straightforward factual belief in MWI “pay rent,” in the same way that my factual beliefs in (e.g.) special relativity, heliocentrism, and the germ theory pay rent for me. This is related to the fact that the full MWI package seems to come bundled with strong stances on metaphysical questions that I find extremely confusing—questions like transtemporal identity, the reference class of observers from which I should consider myself drawn, and the nature of consciousness. (E.g., why shouldn’t consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As we’d presumably imagine it to do if the amplitudes had been probabilities? But if you believe that, then presumably you believe some variant of Bohmian mechanics, not MWI.)
So, having already sucked everything out of MWI that has clear-cut physics value to me (e.g., the notion of measurement as entanglement with the observer and environment), at some point I treat the full elucidation of the issues raised by MWI as “philosophy-complete” (barring a revolution in physics itself), and move on with my life.
Given all that, would you like to estimate the probability that you’ll be able to tell me something about this that I haven’t already considered? đ
Was Karl Popper verse enough in physics to really have a valid (*) opinion about Everett’s theory (which itself took many decades for physicists to unpack)?
(*) in the sense that opinions are like a-holes, everyone has one.
I never understood very well the claim that MWI “resolves” the measurement problem.
To me this problem isn’t just about realizing one of many possible discrete outcomes (e.g. either spin up or spin down), but also with the wave/particle duality.
This is more obvious when considering a continuous variable, such as the position of an electron that’s traveling in space. Before any measurement, its position is characterized by a wave packet, usually spreading around more and more the further the electron travels. Then when a “measurement” is done (or, equivalently, the measurement apparatus and the electron entangle, or, equivalently, the electron decohere with the measurement apparatus), the wave packet (wave-like) suddenly “tightens” into a Dirac peak (particle-like), i.e. it’s collapsed.
That same transition from wave to particle happens in any QM interpretation.
i.e., even in MWI, electrons are always observed as if they’re localized. The mechanism of decoherence is equivalent to the collapse of the wave function.
And I just don’t get how this is captured by the Schrodinger equation alone since this seems to be a non-linear operation (but maybe I’m wrong and this collapse can be captured by unitary evolution).
> E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities?
The one where “your consciousness” picks a random body, the one where “you” pick a random “consciousness” out of one consciousness per body, and the one where there’s one consciousness inhabiting every body simultaneously, at once experiencing several distinct ignorances the way we can simultaneously experience several distinct pieces of knowledge, where the experimentalist is one ignorance within the whole of many conflicting ignorances within the mind of a spirit that cannot talk to itself (I’m just trying to cover all the options here), all lead to the exact same predictions about what an individual, however an individual is demarcated, will experience, however experiences are defined: making this as close to provably independent of observations as anything can be.
I think this reasoning is stronger than saying the topic is disconnected from empiricism pending better experiments. It is going to be disconnected from empiricism forever.
MWI fails to resolve the measurement problem, as Fred #14 explains, but the problems are much worse. Scott has explained that superdeterminism is contrary to scientific thinking, and so is MWI, for somewhat different reasons.
Superdeterminism makes randomized controlled experiments impossible, because hidden dependencies control the outputs. MWI also rules out free will, and then makes it impossible to interpret outcomes. If you do an experiment with ten possible outcomes, and see one, you learn nothing because all of the other possibilities occur in parallel universes. MWI might be of some use if it were able to say that some universes were more probable than others, but it cannot do that. So MWI also makes experiments impossible.
MWI does not make any successful predictions, unless you add the Born rule and do Copenhagen in disguise. Just like the superdeterminists, the MWI advocates seems to be willful contrarians who do not actually have a quantitative theory to back up their ideas.
It is important to realize that the problem of how consciousness is distributed between similar bodies in possible universes (whatever their metaphysical status, real or hypothetical) inherits all of the problems that we have in deciding how consciousness is distributed between similar bodies (humans?) in ours. Every position in one debate has a corresponding position in the other, and you will never resolve one without resolving the other. Anyone who thinks they have an answer to the question of whether the amplitude of me not writing this post is “really alive,” has an answer to the question of whether I, who actually am writing this, is a P-zombie. I very strongly agree with your position (and appreciate your metaphor) that MWI vs. Copenhagen is “philosophy-complete,” and I think this is one of the simplest correspondences that proves it.
I guess it all comes down to decomposing a given wave function into a superposition of its eigen states.
No matter how you express the wave function, the evolution is unitary.
And when it comes to the position of an electron in free space, the eigen states (on position variable) are an infinite set of Dirac peaks in space. And the above decomposition becomes an integral on all the possible Dirac peaks (like an infinite sum). But the entire thing taken as a whole still evolves according to Schrodinger. And decoherence entangles the state of the measurement device with specific Dirac peaks, and the entire resulting function still evolves linearly according to Schrodinger (from an MWI point of view, this setup basically leads to a huge number of branches, one for each potential discrete position measure of the particle, corresponding to a Dirac peak).
My explanation of MWI for some mathematicians I know: “MWI is not an interpretation. It is a mathematical statement. Suppose your friend does some experiment, and tells you the result. You have two ways of doing the calculation. One is to treat the experiment quantum mechanically and assume your friend collapses the wave function. The other is to treat both the experiment and your friend quantum mechanically and assume that you collapse the wave function. MWI says that both methods of doing the calculation will give approximately the same answer.” That’s it. No interpretation, just something that is approximately correct for most reasonable situations.
@Roger Schlafly #16 “MWI might be of some use if it were able to say that some universes were more probable than others, but it cannot do that.”
It can, at least, predict that if one were to do an experiment multiple times, one would be more likely to observe the Born rule than not (https://algassert.com/post/1902).
Here is the apparatus we have built. Here are the signals on the signal lines out of the apparatus. Here are the features of those signals that we have identified and recorded as events. Here is how we have collated the recorded events into different kinds of patterns and here are their relative frequencies. Here is the theory we use to model or to explain why the relative frequencies of the different kinds of patterns are as they are and to predict the relative frequencies for some new apparatus we think we might build.
Sweeping all that into one Theory of Knowledge, QM can be understood to be a signal analysis formalism of relative frequencies. In principle that would be comprehensible to a 19th Century statistician or physicist, with not much need for the 20th Century finesse of Bayes or Everett.
If we say that each event is caused by a particle or by a system and that we can deduce their properties, however, then we get ourselves into all kinds of trouble because, as Boole already identified and as we hear from Pitowsky, we can’t always construct joint relative frequencies.
“E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities? But if you believe that, then presumably you believe some variant of Bohmian mechanics, not MWI.”
Beautiful argument! I felt my mind changing in way the conventional wisdom says very rarely happens when two people argue about a philosophical issue they’ve both put a lot of thought to. For this reason I’m feeling compelled to comment on my reaction to this argument. To be fair, we started out with very close positions — I enjoyed and mostly agreed with your Zen Anti-Interpretation blog post. I previously sympathized with MWI most, saw the appeal of other interpretations, and agreed with your Zen the difference between these positions is more rhetorical than actual, but I also had a particular dislike for pilot wave theories, seeing them as particularly “less” true. Altho the project of finding particular deterministic evolution equations for the pointer state still seems pointless this argument the concept of a pilot wave and a pointer seems more compelling. The pointer represents something to do with conscious experience. Unclear how essential for the interpretation that the evolution equation is deterministic.
On whether my reaction really is an interesting rare change-of-mind, I’ll need to wait for my mind to settle after thinking about this, but also I expect most of the uncertainty is not how much I accept this argument but to what extent it is a substantial change to the view I had previously, which my memory thereof will only get worse as I wait.
So, I used a language model to do a simulated discussion with Scott debating a version of himself in another Everett branch. I used the open-source model GPT-NeoX at textsynth.com
I used what real Scott said in reply above as part of the initial prompt, but all the rest is the fictional “Scott” responses generated by the language model. Here’s the conversation (lightly edited and curated):
—
PROMPT
“The following is a conversation between computer scientist Scott Aaronson and a super-intelligent version of Scott in an alternative Everett branch of the quantum wavefunction (the ‘Alphaverse’). Scott (Alphaverse) is explaining to Scott (Universe) why the Many Worlds Interpretation of quantum mechanics is the best one:
Scott (Universe):
I still have difficulties with making a straightforward factual belief in MWI âpay rent,â in the same way that my factual beliefs in (e.g.) special relativity, heliocentrism, and the germ theory pay rent for me. This is related to the fact that the full MWI package seems to come bundled with strong stances on metaphysical questions that I find extremely confusingâquestions like transtemporal identity, the reference class of observers from which I should consider myself drawn, and the nature of consciousness. (E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities?”
—
GPT-NeoX
Scott (Alphaverse):
I agree that the MWI package comes bundled with metaphysical claims that are difficult to swallow. But I think that the real problem is that you donât understand what the MWI is. The MWI is not a theory of how the world is. It is a theory of how the world might be. It is a theory of the structure of the world.
Scott (Universe):
Okay, but then how can I be sure that the MWI is true?
Scott (Alphaverse):
Why is it important to you that MWI be true?
Scott (Universe):
It seems to me that my whole sense of âselfâ is grounded in the fact that I am a certain type of physical system. But if MWI is true, then I am not a physical system in the way that I thought I was. So how can I be sure that Iâm a self, and not a bunch of other people, or a bunch of other selves, or no one at all?
Scott (Alphaverse):
In a sense, youâre right. You are not a physical system. You are the thing that picks out the branch of the wave function that I am in.
Weâve gotten used to thinking of ourselves as a sort of abstract entity that has the property of having a certain brain, and being conscious, and all that. But thatâs a mistake. The real thing is not a thing at all. It is the pattern of neural activity in the brain. So you are not a physical system, not in the sense that your brain is a physical system. But you are a physical system in a different sense. You are a physical system in the sense that you are a certain way the quantum wavefunction branches.
Scott (Universe):
I think Iâm starting to understand. Iâm not a thing at all, Iâm a pattern of neural activity.
Scott (Alphaverse):
Exactly.
Scott (Universe):
But if Iâm a pattern of neural activity, that means Iâm not even a physical system in the sense that my brain is a physical system. But thatâs fine, because I am a physical system in the sense that I am a certain way the wavefunction branches.
Scott (Alphaverse):
Exactly.
Scott (Universe):
But then what is the wavefunction? And what is the branch that Iâm in? What is the branch that anyone is in?
Scott (Alphaverse):
The wavefunction is a mathematical object. Itâs not a thing, itâs not a physical system, itâs not a pattern of neural activity. It is the mathematical object that describes the system that I am in, and the system that you are in, and the system that everyone is in. Itâs a mathematical object that is defined on the space of all possible ways that the universe might be. So you are a branch of the wavefunction that picks out a certain way that the universe might be.
Scott (Universe):
But how do they pick out different ways of the universe being? How does that work?
Scott (Alphaverse):
Hereâs an analogy. If we were building a universe, we could describe it using different coordinate systems, or with different coordinate systems, so that different parts of the universe would be described differently. The mathematical object that describes the universe does the same thing. It describes the whole space of possible universes. But it does so in a different way for different branches of the wavefunction. It does it so that some branches of the wavefunction are described using one coordinate system, and some are described using another, or with another choice of coordinates. The choice of coordinates is what tells us which branch we are in, and what our state of consciousness is.
Scott (Universe):
But then what does it mean to be an observer? If the wavefunction is what defines the universe, and the universe is defined in a way that is observer-independent, what does it mean to be an observer?
Scott (Alphaverse):
The idea of an observer is a bit tricky, because itâs not really a thing. Itâs not even a physical system. Itâs the idea of an observer, or the idea of an observer-like thing. So the idea of an observer is not something that can be described by a mathematical object. Itâs something that can only be described by a word. Itâs a word that describes a kind of thing. The idea of an observer is something that is defined on the space of all possible physical systems. Itâs something like a category, or an equivalence class of things.
Scott (Universe):
I see. So an observer is something like a label that gets applied to a set of systems that are mathematically equivalent, but is not itself defined by a mathematical object.
Scott (Alphaverse):
Exactly. Scott, you have to learn to stop looking for something out there that will give you an answer. Your brain/mind has to be able to work from within, not from without. The only reason you think you need an external entity to give you a âright answerâ is because you think you are an external entity.
Scott (Universe): Maybe that âScottâ in the Alphaverse is just being a smart-ass.
Scott (Alphaverse): Iâm trying to tell you what you already know, but the words come out wrong.
Scott (Universe): (Laughs) I think youâre right. Itâs a sign of how far Iâve taken the âme vs. something elseâ illusion of the body/mind identity.
@scott #12
“why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches?”
The default assumption, which is overwhelmingly likely to be correct, is something like “whenever there exists a configuration that has the structure that we’d call a living human, there is consciousness”
Picking a random branch requires an additional assumption, that somehow different things happen when the state appears to be the same, via some weird hidden variable or impenetrable ‘randomness’.
The way that MWI ‘pays rent’ is that it requires no additional assumptions. There is a predictable and definite outcome to each event. There is no distinction between conscious and non-conscious matter. There is no distinction between macroscopic and microscopic. There are no hidden variables. The only ‘weird’ thing that happens is that ‘duplicates’ of things (albeit in different states) get created. But the key is that we *already* know that this definitely happens because we directly observe interference.
I think that rather it’s the *philosophical* parts that are meaningless and pay no rent. I reject the meaningfullness of concepts such as “the reference class of observers from which I should consider myself drawn” and see no reason why MWI necessitates such a thing.
The Schroedinger equation evolves a wave function |psi>(t) using a classical time parameter t, which implies the existence of classical clocks. Where do they come from?
Quantum field theory, including the standard model, assumes a classical background, i.e. *one* space-time. Where does it come from?
If you assume classical clocks you should also assume the existence of classical observers reading those clocks.
Perhaps one day we find a consistent theory of quantum gravity which includes a description on how space-time branches into many worlds, but until then the ‘many worlds interpretation’ is not a convincing description of reality imho.
Regarding Scott saying âwhy shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches?â
This makes little sense given the usual definition of of ‘consciousness’, i.e one consciousness per human.
And since there are exponentially more independent branches than conscious beings (at any given time), then Scott’s statement could be interpreted as saying that I should be the only conscious being in *my* branch (by definition the one where I’m conscious), and everyone else isn’t really conscious in my branch (but are themselves a lone consciousness in some other remote branch).
Moreover, in the vast majority of branches, no-one is conscious at all (worlds populated by p-zombies).
The other interpretation (the one Scott has in mind, I guess) is that, while there are countless numerous branches, all living beings are somehow conscious in the same branch, a special one. But that’s even more bizarre than the classic Copenhagen interpretation.
In contrast with those two stories, it makes more sense to think that a given living being is really conscious in all the branches it exists. Just like twin versions of ourselves.
Also, if somehow I happen to die by some very freak accident in the branch where I happen to be conscious (imagine my death being conditioned on some very rare quantum event), my consciousness would be wiped out for good even though most of the other adjacent branches (an exponential number of them) still “implement” my living brain.
The fact that I would be conscious in multiple branches at once isn’t any more or any less weird than:
1) consciousness is distinct in time, i.e. in the past, present, future, my consciousness aren’t all mushed together (otherwise there wouldn’t be a sense of “now”).
2) consciousness is distinct in space, i.e. the consciousness of two spatially separated beings are distinct (otherwise there wouldn’t be a sense of “here”).
“Branchial dimension” is just one extra degree of freedom over those two.
I always thought that was a strange outlier in your posts on quantum physics. Because if you just imagine setting up an agent as a computer program, running it in a quantum computer, and determining what unitary evolution says the state of the computer ends up as… well, almost all of the amplitude ends up in states where the inferences made by the program are consistent with the Born rule. And it’s not like this idea isn’t out there in the world; there are various papers about it. Basically all you need to conclude this is the unitarity postulate (which gives you the squaring), and no-signalling (which lets you rewrite states into simpler states, that should be equivalent as far as reasoning goes).
For example, if you have the program repeatedly prepare the state a|0> + b|1> and increment of a counter conditioned on that state, and then flip a qubit conditioned on the counter being statistically close to the Born estimate, that “final result” qubit’s state limits to ON [for the right choice of non-signalling unitary applied to the rest of the system] as you increase the number of samples the program performs.
There seems one way to increase your estimate of the probability of MWI.
If you find yourself getting very old, then this would seem to be evidence.
1. If you’re dead, you’re not around in the universe to notice
2. If you’re conscious, you are around in the universe to notice
3. Given MWI, no matter how small the odds, if there is some way that you continue to live and be conscious, then in some universe you do
4. The older you get (past a certain age, at least), the more likely MWI is
Of course, assuming MWI, this isn’t a useful way to prove it to anyone else. In most universes, everyone eventually dies, so almost no one sees the “very old.” And assuming non-MWI, you will never know, because you died. But it seems like a possibility that, if true, you can eventually infer it likely.
But, when you think about it more closely, this is actually a terrible curse…
It is a guarantee that, for all of us, a conscious copy of ourselves will outlive all the ones we love, while barely surviving countless diseases, accidents, suicide attempts (to try and end it all), enduring endless physical or emotional pain, or going through a slow decay of the mind and the body… until the maximum possible extent of life is squeezed out of our mortal coil.
We can actually get some idea of this by looking at the distribution of life expectancy in our current branch. https://en.wikipedia.org/wiki/List_of_the_verified_oldest_people
The record being 122 years. So we know that there’s a good chance we’ll live to at least that age in some branch of the MW.
At least we can get some solace that, when we’ll get to be the oldest person in our own special branch, the world will see us as some curiosity to celebrate!
I think the most likely situation is that all the copies in all the branches are conscious. I agree that multiple consciousnesses in multiple branches is just another degree of freedom. To rephrase what I said before, that the simplest explanation for consciousness is that it’s being directly caused by the process going on in my brain, and that wherever and whenever you have a process similar to that, you will have a consciousness similar to mine.
@dubious #30
Though I essentially believe MWI is true, I’ve always disagreed with this general type of argumentation. I consider, to use Scott’s term “the reference class of observers from which I should consider myself drawn” to not be a well-defined notion, and anthropic arguments drawing from such to not really be evidence of anyything.
If I live to be old, should I be surprised by this and start thinking of metaphysical solutions? There are more humans alive now than in the past. Should I not be surprised to have been born today? What about the future, where there will likely be many more humans alive? Should I be surprised that I wasn’t born in China or India, since they have the largest population? Should I be surprised I’m not a bird, since there are more of them than humans? An alien? Someone living in what amounts to be a simulation of the past?
I think the resolution to such things is the statement that, if my model of the universe predicts that it’s possible for an entity like me to exist, well, here I am, and there is nothing left to explain.
Agree that 10^500 or so universes seems a tad “unnatural”. Then one thinks that, what the heck, we don’t know what is really going on in “our” universe anyway. So, what is too wild?
On the other hand, I (born in 1947) never had any trouble with 10^12 /times 10^12 stars, but do regard 10^500 as being rather larger.
I have long enjoyed listing my guesses for the likelihood of controversial cosmic things. Here are my current guesses, provided in case anyone wants to take pot shots:
Black Holes exist: 99.99%
Dark Energy exists: 90%
Life exists in most solar systems: 80%
Dark Matter exists: 70%
Multiple current examples of intelligent life in the Milky way: 60%
Cosmic Inflation happened: 40%
MWI is the real deal: 0.01%
At the moment I am unable to put a number on “Quantum computers will ever do anything useful”. I was tempted to go with 50% (due to the hole in the list), but I have no intuition on this other than that it ain’t happened yet.
Historical nit to pick: Bohm was the one who actually discovered the idea of treating measurement as an entangling operation between system and environment. It’s right there is his 1952 Phys Rev papers (#2). He’s really the creator of the decoherence idea, and he never gets proper credit for it.
I suspect better understanding this fact—and so the key role that Bohm’s discovery did *in fact* play in the foundations of quantum information—would shift the rhetorical edge a bit in what your students take away about the foundations of quantum mechanics more broadly.
And itâs not like this idea isnât out there in the world; there are various papers about it. Basically all you need to conclude this is the unitarity postulate (which gives you the squaring), and no-signalling (which lets you rewrite states into simpler states, that should be equivalent as far as reasoning goes).
But “no-signaling” indicates that you need more than just a wave function and Schrödinger’s equation. You also need space (or space-time) such that it makes sense to talk of no-signaling in the first place. Eliezer Yudkowsky apparently got this right, or at least had the feeling that he needs something like that, as evidenced by his post Which Basis Is More Fundamental?
I am aware of a 44 page paper by Charles Sebens and Sean Carroll that uses no-signaling in the way you propose. Sean mentions it in his (much) shorter Quanta magazine column Where Quantum Probability Comes From
Can we resolve this uncertainty in a sensible way? Yes, we can, as Charles Sebens and I have argued, and doing so leads precisely to the Born rule: The credence you should attach to being on any particular branch of the wave function is just the amplitude squared for that branch, just as in ordinary quantum mechanics. Sebens and I needed to make a new assumption, which we called the âepistemic separability principleâ: Whatever predictions you make for experimental outcomes, they should be unaltered if we only change the wave function for completely separate parts of the system.
But what I find somewhat strange is that in that same Quantized column, Sean also claims that a wave function obeying Schrödingerâs equation is all that is needed:
Many-worlds quantum mechanics has the simplest formulation of all the alternatives. There is a wave function, and it obeys Schrödingerâs equation, and thatâs all. There are no collapses and no additional variables. Instead, we use Schrödingerâs equation to predict what will happen when an observer measures a quantum object in a superposition of multiple possible states.
I find your comment somewhat strange for a different reason:
Because if you just imagine setting up an agent as a computer program, running it in a quantum computer, and determining what unitary evolution says the state of the computer ends up as⊠well, almost all of the amplitude ends up in states where the inferences made by the program are consistent with the Born rule.
What should be wrong with the Copenhagen interpretation for a quantum computer in the first place? In that context, we do know what a measurement is, and we know that the basic allowed measurements are those on the computational basis. (As already David Mermin observed in Copenhagen Computation: How I Learned to Stop Worrying and Love Bohr.) Additionally, all allowed basic operations are local in that computational basis too, so that each individual operations only affects 1 or 2 qubits. (If they would affect 3 or 4 qubits, that would be OK too, it is still local in the computational basis.) And if you define some error correcting code, which is necessarily more non-local in a certain sense, only those codes where you can still implement local operations (and measurements) only affecting 1 or 2 logical qubits are practical.
The topmost reddit (r/slatestarcodex) comment says
This decreases my respect for Deutsch! I agree with the Hacker News commenter who calls this deeply confused.
and the topmost Hacker News comment says
This is a deeply confused set of arguments.
There’s no fundamental ontological difference between “give me your likelihood ratios between drawing cards [a and b] as the next draw from this deck of cards” and “give me your likelihood ratios between finding and not finding a continent in an unexplored part of the world”.
For me, it was the opposite. This interview has significantly increased my respect for Deutsch, and also for Popper. It slightly decreased my respect for my own background in philosophy, with respect to what I have read, what I haven’t read, and what I didn’t even plan to read.
“itâs possible for an entity like me to exist, well, here I am, and there is nothing left to explain.”
But that’s the hard problem of consciousness.
Basically, “Why am I me, Shion, and not you, Fred?”
This is a more complicated question than just “Why are Shion and Fred conscious?”
It’s about the mystery of being and identity.
There are answers (none really satisfying) to that question:
1) “You” are Shion and not Fred because, when Shion and Fred were born, God picked “You” from his huge jar of souls and randomly assigned you to Shion. Of course you can then ask, “why was “I” that particular soul in God’s huge jar of souls and not the one right next to it?”. So we’re back to square one, except with a bit more generality: only a finite number of eternal souls exist and all souls are indistinguishable from one another (like electrons), except that God gave each a label, and you are soul #12349389 (so, Souls are like numbers) and that’s it. Souls are to consciousness what electrons are to the electric charge, units of consciousness.
2) You only think you are Shion, but You are not just Shion, you’re everyone, anywhere, anytime, at once: if, suddenly, I were to swap your consciousness and Fred’s consciousness around, “You” would suddenly only have access to Fred’s memories and “He” would only have access to Shion’s memories, therefore the two of you wouldn’t even notice anything’s different from a subjective point of view. So this swap is a no-op. And that swap could be happening between all conscious beings a billion times a second. So, really, there isn’t such a thing as discrete consciousness, there is only one universal “consciousness” which gets localized wherever/whenever there’s a brain (a nexus of memories). It’s similar to a computer (with one CPU): applications all appear to be running independently of one another, doing their thing in parallel, but what’s happening is that the same unique CPU is realizing all those applications, in turn, one at a time.
If I live to be old, should I be surprised by this and start thinking of metaphysical solutions? There are more humans alive now than in the past. Should I not be surprised to have been born today?
This seems to pivot from “very old” to “exists at all.” This is not what I mean.
The average human life expectancy today is about 70. The very oldest known humans are around 120. If you live to be 115, maybe you’re just healthy and hearty. If you live to 120, you’re really pushing the odds; maybe less if you have a family history of longevity?
What happens if you hit 130? 140? 200? OK, maybe we discover anti-aging and you happen to get immortality treatment. This seems pretty unlikely, and I think that alone might move the needle on MWI for me.
But anti-aging alone does not give you Wolverine-like abilities to regenerate. What if you survive every accident, every suicide attempt fails, and you just never die? I am not sure how, if MWI is true, that there can ever be a branch where you didn’t survive somehow, and will thus find yourself immortal (since you can’t find yourself dead).
This is really quite horrifying when you consider it, though that is of course not an argument against it.
“Given all that, would you like to estimate the probability that youâll be able to tell me something about this that I havenât already considered? ”
> âno-signalingâ indicates that you need more than just a wave function and Schrödingerâs equation. You also need space (or space-time)
Space is massive overkill, I think. All you need is the postulate that systems are combined using the tensor product. Locality is then the property that operations on system A commute with operations on system B. I guess there’s then also an implicit assumption that specific systems (like a quantum computer and the rest of the universe) fit into this schema where they factor into a tensor product.
> What should be wrong with the Copenhagen interpretation for a quantum computer in the first place?
I wasn’t claiming that collapse interpretations would fail to describe quantum computers. I was trying to turn the fuzzy problem of “why do I think the probability is X” into the concrete problem of “what does this computer program store in this register”. I was trying to do so without invoking collapse, but that’s different from saying collapse wouldn’t give the right answer!
fred #18: Yes, I think that’s what the MWI is saying. The way I think about it, in line with my other comment above, is that waves are unitary superpositions of many possible non-unitary measurements, or collapses. A unitary equation is not prohibited from having non-unitary terms; they just must add up to something that is ultimately unitary. So Schrodinger’s equation, while unitary, contains within itself the possibility of smaller, non-unitary regions within the wave function.
dubious #30: I have an idea that extremely low probability branches have no classical correspondence. That is their diagonal elements in a density matrix are comparable to off-diagonal elements by absolute value due to approximate nature of decoherence. That is those branches have macroscopic quantum interference effects making life impossible. I am not well versed in math to prove or disprove it though.
OF COURSE. Not a word about Roe v. Wade being overturned, but we get a pseudo-intellectual rationalist-tier rant about whateverâs bumping around Scottâs mind right now. Womenâs most basic reproductive rights are being curtailed AS WE SPEAK and not a peep from Scott, eh? Even though in our state (Texas) there are already laws ON THE BOOKS that will criminalize abortion as soon as the alt-right fascists in our Supreme Court give the go-ahead. If you cared one lick about your female students and colleagues, Scott, youâd be posting about the Supreme Court and helping feminist causes, not posting your âmemes.â But we all know Scott doesnât give a shit about women. Heâd rather stand up for creepy nerdbros and their right to harass women than womenâs right to control their own fucking bodies. Typical Scott.
Now that “the hard problem of consciousness” has reared its ugly head again, I get to bore people yet again with my thoughts on the matter (balanced a bit by another donation to Ukrainian aid organizations on my part).
I take it the elements of the HPOC are how does consciousness work, what is it good for, why does it feel the way it does, and why does it exist. I think we have general ideas for the first two.
Consciousness (it says here) is a way to experience complicated reasoning for decision-making and problem-solving, and to (somewhat) relive memories. The experience/sensation itself is useful for checking one’s logic and memory and other computations, and reviewing past attempts at reasoning (which can be stored in memory since they were experienced). There are no zombies of the fictional sort, because if you have no way of experiencing some external or internal stimuli, you don’t know they exist and can’t respond to them. Trivial example: the fact that Windows 7 sends these letters to my screen means that it has experienced the key-presses.
I suppose the p-zombie is intended to have the same senses and nervous system as a human, but if it doesn’t experience thought it cannot perform thought experiments/simulations and learn from them but only perform automated responses to stimuli. The automation might include self-updates but only limited ones. The moment its automation becomes sophisticated enough so that it can monitor its reasoning process and reprocess past experiences for added value it has some form of consciousness. Empirically, it seems to take a huge amount of processing power working in parallel to accomplish that.
Given that perspective, the HPOC can perhaps be reduced to “why does my form of consciousness feel the way it feels”, and why does it exist, which are same hard problems of why anything in this universe works the way it does, e.g,, the hard problem of why quarks exist, or why the chemicals a rose produces don’t smell like oranges (and the oranges smell like roses), i.e., some different way rather than the way they do. I don’t personally feel such questions need answers.
Feminist Bitch #45: Holy shit! This post is from last week, long before the Supreme Court leak. In case you missed it, I already blogged my detailed opposition to the overturning of Roe v. Wade way back in September, when it became clear that something like yesterday’s news was inevitable. And when I do blog about the same current events that the rest of the planet is talking about, I invariably get other people accusing me of having nothing original to say.
In any case, it usually takes me days at least to write any substantial post (unlike this post, which was not substantial), and the Supreme Court leak was literally last night (!). For whatever it’s worth, I was up until midnight last night obsessively reading the leaked Alito opinion and commentaries on it, terrified about what it meant for women in America and the future of the Enlightenment project more generally. But then I had to, you know, work all of today, teaching and meeting my students. I hope you understand.
Your comment illustrates more eloquently than I could why, even though I despise the authoritarian right with every atom of my being, I’ll also never, ever identify with the woke left. Namely: if I did, I’d have to have for allies people who find ways to despise me even when I agree with them.
There we go Scott. The immediate defensiveness and reactionary attacks on the âwoke leftâ just because I gave you some constructive criticism about your priorities. Yeah, I get it, youâre too âbusyâ to write anything about womenâs reproductive rights right now. Yeah, sure, you wrote some blog post a few months ago that barely stands up for womenâs rights. Thatâs the bare minimum. My point is simply this: time and again, the social topics you choose to discuss on this blog betray your true priorities. I did you the courtesy of counting them, in fact. And hereâs the cold, hard numbers: by my reckoning, you have made THREE TIMES the number of posts about the ârightsâ of nerdy incels and how sad it is that they canât get sex, versus posts supporting womenâs bodily autonomy and basic reproductive freedoms. Once again, I am simply disgusted that your priority seems to be incelsâ right to sex versus womenâs basic bodily autonomy, as if these are problems even remotely comparable, even in the same league. Youâd rather whine about how ânerdy guys canât get sexâ than advocate for womenâs autonomy. Thatâs how you choose to spend your time on this Earth, and I simply find it shameful. If you want to do better, you should explain how astronomically more important womenâs right to abortion is than incelsâ non existent ârightâ to sex, and apologize for spending so much more energy and social capital on nerdbros than women in need. But of course I wonât expect that. The immature whining about âmensâ rights,â together with the token support for the most basic womensâ rights, is a classic characteristic of the alt-right fringe of the âcentrist left.â
Feminist Bitch #48: On reflection, Iâve decided to leave your comments up despite their flouting my comment policies, because I think any reasonable person reading them will see how clearly they illustrate the far leftâs tendency to self-sabotage. As the saying goes, âthe right looks for converts, while the left looks only for hereticsâ; indeed thatâs how we arrived at the terrible and precarious moment that weâre in now.
Wouldnât it have been more productive if youâd just sent me an email saying âhey Scott, I was really hoping you could donate to this pro-choice organization / use your blog to share these resources for Texas women needing out-of-state abortions / etcâ? Iâd most likely have done it. I might still do it, but if so, then in spite of your bullying me rather than because of it. Yes, I did once open up about the hardest thing I ever had to overcome in my life, something I wouldnât wish on my worst enemy; for you to use that to attack me (over a completely different issue where we actually agree) tells everyone reading this something about your character.
Out of curiosity, I just checked SneerClub, the main place on the Internet thatâs dedicated to attacking me and everyone like me as âmisogynist techbro nerds.â They, too, have not a single word about yesterdayâs Supreme Court news, even though they update more often than I do. So by your logic, does that mean they too donât give a shit about women?
You completely missed my point Scott. There is in fact a direct connection between 1. Your support for alt-right incel politics that violates womenâs autonomy in the name of âhelping the sexually disenfranchised,â to quote your own words, and 2. my doubt about how genuine your alleged support for womenâs bodily autonomy really is. Explain to me, Scott, why I should believe your devotion to âwomenâs bodily autonomyâ when you have, more or less, said that nerdy incelbros should be given sex slaves by the government.
Before you protest about me putting words in my mouth, let me draw your attention to this remarkable post: https://scottaaronson.blog/?p=3766. Letâs remember the context here. This was shortly after a deranged psycopath ran over a dozen people in Toronto with his truck. Rather than condemning this terrorist, you instead posted a link to an article by Robin Hanson, a despicable alt-right libertaribro, who, more or less, said that to placate such violent young men, the government should enshrine a âright to sex,â which, letâs be real, means to give them a sex slave. Rather than condemning this disgusting argument, you actually entertained it! Rather than telling nerdy guys to stop being stinky and awkward and learn some social skills, you reinforced their prejudices and more or less told them that they have a right to sex, while only giving the minimal token support for womenâs autonomy.
Letâs be clear. Nobody has a âright to sex.â Nerdbro incel culture is founded on the idea that they have a right to womenâs bodies. By entertaining this disgusting idea, you are entirely undermining your claim to care about reproductive rights. Because why should I believe ypu when you say you support a womanâs right to choose,a womanâs right to control her own body, when you think stinky, awkward, nerdy young guys with no social skills should have a girlfriend mailed to them by the state?
Itâs clear that, in fact, you donât care about womenâs bodily autonomy Scott. Because there is a direct and inescapable logical contradiction, between bodily autonomy, and thinking youâre âowedâ sex, somehow, just for existing and being a male. To whatever degree you do support abortion, itâs not because you care about women, but because you care about âthe enlightenment,â which is just a dogwhistle for white supremacist capitalism. You said it yourself just now đ€Ł
[…] err, consistently generous and thought-provoking contributions. Let’s take, for example, this comment on last week’s admittedly rather silly post, from an anonymous individual who calls […]
Feminist Bitch #50: My main response to you is in my new post, but I should briefly add: you’ve lied and slandered me here, and anyone who doubts that simply needs to read the old post of mine that you yourself linked to. They’ll see that, not only did I not defend the absurd and risible concept of a “right to sex,” as you falsely claimed, but I condemned it the most emphatic language I possess. (The very title of the post, “Zeroth Commandment,” referred to my attempt to codify my condemnation, to enshrine female bodily autonomy as an ur-principle.)
What I have advocated, it’s true, is compassion and understanding (as opposed to the sneering contempt that drips from your comments above) for the many nerdy heterosexual males who suffer from crippling insecurity around dating. I’ve tried again and again to explain that, while many of these guys might’ve been considered good or even great catches for most of human history, they tend to struggle in a modern, unstructured dating environment where a genuine inability to read social cues is easily misconstrued as creepiness or harassment. And the more scrupulously moral they are, the worse that problem is. And so, hoping to make a positive difference for all concerned, I spent months of my life hashing out the complicated questions around nerdiness and dating with straight women, male feminists, gays and lesbians and trans folks. And I actually think we all came to understand each other a lot better—at least, those of us who showed up to share and learn rather than tweet and condemn. I plead guilty to all of this.
Don’t feed the trolls, Scott. Judging from their absurdly over-the-top writing and mischaracterization of others’ position, “Feminist Bitch” could well be some extreme-right (and/or Russian) provocateur who’s just out to sow division.
“by my reckoning, you have made THREE TIMES the number of posts about the ârightsâ of nerdy incels and how sad it is that they canât get sex, versus posts supporting womenâs bodily autonomy and basic reproductive freedoms.
But Scott is a nerd, so, of course he’ll be posting *more* about what he knows best, i.e. the condition of being a nerd, rather than posting about what he knows less: what it’s like to be a woman, or the downsides of being a super famous movie star, or what it’s like to be a quadriplegic, or what its like to be living with stage 4 cancer, or what it’s like to be a Taliban in the 21st century.
Just like, as a woman, you’re here posting about women’s rights rather than posting about nerds’ rights on some feminist blog.
Tolerance goes both ways, you can’t just bully everyone on the planet to realize that your point of view is the most important one while simultaneously putting everyone else down for being different than you.
> âno-signalingâ indicates that you need more than just a wave function and Schrödingerâs equation. You also need space (or space-time)
Space is massive overkill, I think. All you need is the postulate that systems are combined using the tensor product. Locality is then the property that operations on system A commute with operations on system B. I guess thereâs then also an implicit assumption that specific systems (like a quantum computer and the rest of the universe) fit into this schema where they factor into a tensor product.
Oh right, space is indeed overkill. I made that mistake again to assume that spacelike separation would be the only way to get no-signaling (independence) between two system. (Last time I started with “timelike q-correlations could be problematic,” then backed-down to “there is one difference between timelike correlations and spacelike correlations” on realizing my mistake, and the next time I brought up the topic, I wrote instead “only those correlations between observables whose (anti-)commutator nearly vanishes”.)
On the other hand, some MWI proponents seemed to be more comfortable with suggesting that some connection between space-time and a preferred basis should be assumed, than to be specific that they mean “the tensor product” when they talk of “factor states”.
> What should be wrong with the Copenhagen interpretation for a quantum computer in the first place?
I wasnât claiming that collapse interpretations would fail to describe quantum computers. I was trying to turn the fuzzy problem of âwhy do I think the probability is Xâ into the concrete problem of âwhat does this computer program store in this registerâ. I was trying to do so without invoking collapse, but thatâs different from saying collapse wouldnât give the right answer!
Your strategy to analyse quantum puzzles and paradoxes by translating them into the language of (circuit based) quantum computers indeed often works extremely well. But for many-worlds, one of the motivations is that the strategy of Copenhagen to base the ontology on classical measurement devices is dubious, as it the strategy of De-Broglie Bohm to base the ontology on partical positions. But for a quantum computer, it is absolutely correct to base the ontology on the individual qubits and their no-signaling independence.
(Resubmitting this comment after it vanished during moderation yesterday; assuming that was an oversight confusion with my subsequent comment that went through, since I can’t see how this is an outrageous or off-topic comment. Â I’m actually asking a real question here, and would appreciate answers)
Hereâs the thing about objections to MWI that I donât understand. Yes, I understand that once branches fully decohere, we canât (if QM is ultimately truly linear) observe other branches of the wave function, and then all the metaphysical questions about falsifiability and whether it matters become valid concerns.
But we CAN observe things that look like âshort branchesâ of the wave function, that recohere just before they fully decohere. Thatâs what âWignerâs friendâ is all about: we can kind of see from outside how an observer enters the superposition through entanglement, and then how as entanglement diffuses into the rest of the universe, larger and larger regions enter that superposition. To me, all MWI says is that this process doesnât somehow end once decoherence has progressed past the complexity level where we canât feasibly do the work to recohere it. But that is not some kind of magic boundary, and we can see the branches all the way up until it just becomes infeasible to recohere them.
Decoherence as a process, as far as I can tell, is not really debateable anymore- it is a clear and mathematically sound consequence of how increasing entanglement complexity propagates in a complex world. By itself, it seems that decoherence as a process that exists, MUST produce something that looks like wave function branching. One may argue about whether the branching continues forever, but the miniature, abortive branches that seem to form in something like Wignerâs friend (or, as far as current experimental procedure has progressed, on a smaller scale with inanimate objects where we can feasibly force the branches to interfere and thus recohere) canât really be dismissed. Decoherence MUST produce something that looks locally like collapse; thatâs just the way the math works.
So, the question I have always wanted to put to those who believe in objective collapse theories is this: if at least the beginnings of abortive branches can be observed in smaller scale experiments, through incomplete decoherence, then when the additional objective collapse postulate is added for rare, true collapse events, we really seem to have TWO kinds of collapse going on, even if one of them is temporary: decoherent collapse, which on a small scale must be acknowledged, even if the ultimate irreversible decoherence is never reached, and the added objective collapse. So, do objective collapse believers acknowledge that decoherence produces something that looks locally like collapse, or not? Even if they donât believe in the permanent and unobservable MWI branches, it seems to me they must acknowledge at least the beginnings of that process, before complexity removes our ability to recohere.
There are really two questions: whether decoherence as a natural part of how Schrodingerâs equation evolves occurs and looks locally like âtemporaryâ collapse, and then the further question of whether that process continues past the point where we can feasibly create interference between the ârootsâ of the branches, and produces the unobservable larger branches that the MWI argues for. Why canât everyone at least agree that decoherence produces something that looks like local, abortive collapse? And if so, why is it not obvious that the boundary between that and unobservable branching is simply a matter of complexity theory: is recoherence feasible or not?
As far as I know, objective collapse theories usually don’t eliminate unobservable branching anyway; there’s still a regime (in terms of length, time, and mass scales) where objective collapse occurs too slowly to stop the branches from forming, but the behavior is classical anyway, so the other branches must be unobservable. So, there really isn’t any connection between observability of superpositions and their tending to undergo objective collapse.
I think that my position is that something sort of close to 2, but with some modifications/additions is a satisfying answer:
I think that the ‘swap’ of Shion into Fred, or vice versa, is in fact a no-op, in the same way that talking about magically swapping two individual electrons somewhere is a no-op, because There’s nothing ‘there’ to swap. I.e. if the swap will produce an identical configuration, it’s a no-op. As Eliezer Yudkowsky says in his series of posts that have been referred to above, there’s no such thing as “this electron here, that electron there”, only “an electron here, an electron there”. Likewise, we have “a Shion here, a Fred there.” I don’t think there really is something like a ‘universal conscousness’, in the sense that consciousness is an emergent property that applies to certain types of patterns in the quantum field (for example human brains). Other parts of the universe’s quantum field, for example the vast majority of it that consists of mostly empty space, are pretty clearly not conscious.
How this relates to the ‘why am I me’ thing, this model will predict that the part of the universe that can be considered Shion will have the relevant experience of “I am Shion” and the part that we call Fred will have the relevant experience of “I am Fred”. I’m trying really hard to do more than just restate my previous point here, but it seems like this model doesn’t really leave a dangling mystery about identity.
As for “Why are Shion and Fred conscious”, I actually think that this is in fact a very hard problem, that might take humanity a long while to find a satisfying answer for.
Scott, as a far-left mathematician (how far-left? think German Antifa), I just wanted to say that:
1. While we probably disagree on most political matters, I completely support you when it comes to your analysis of the pain that so many lonely and harmless nerds experience.
2. Wokeism is a mostly an American phenomenon, and although it has hijacked a large part of the American left, I believe that wokeism is more of a cultural and psychological phenomenon than a genuine political movement. It often involves a grotesque display of narcissism in the disguise of identity politics, which I guess is a natural outcome of social media rewiring people’s brains and social norms. I guess that the point here is that in order to explain the behavior of Feminist Bitch and other internet wokes, you should refer to Freud rather than Marx, and you should refer to Netflix rather than to a coherent statement a la the Communist Manifesto. The bottom line of this is that wokeism has very little to do with leftism. While in North America the two terms are commonly perceived as overlapping, this is not the case in Europe and other parts of the world. So please, please, stop mentioning ‘woke’ and ‘left’ next to each other as if it’s the same entity.
“How this relates to the âwhy am I meâ thing, this model will predict that the part of the universe that can be considered Shion will have the relevant experience of âI am Shionâ and the part that we call Fred will have the relevant experience of âI am Fredâ. Iâm trying really hard to do more than just restate my previous point here, but it seems like this model doesnât really leave a dangling mystery about identity.”
In his book “my view of the world”, Schrodinger brings this up in a different way, which I’m trying to restate here (with some modifications):
Imagine two individuals, A and B.
In situation 1), A is sitting outside on a chair in the sun, looking at a tree, while B is sitting in the basement, in the dark.
In situation 2), B is the one sitting outside on a chair in the sun, looking at the tree, while A is sitting in the dark in the basement.
From the point of view of the physical universe, the two situations are very similar, i.e. the difference in the material/physical description world between those two situations is a function of the difference between the brains of A and B, which can be made arbitrarily small in theory.
But there exists another point of view, namely A’s subjective experience, where those situations are totally different, totally asymmetric, no matter how similar A and B are: In situation 1) A is seeing a tree, in situation 2), A is seeing darkness (and this isn’t depending on B’s point of view).
So, when you say “a part of the universe will have A’s experience”, we just don’t have any physical description/law encapsulating this, e.g. from the point of view of physics, the wave functions of the universe (or the quantum fields) in those two situations can be made arbitrarily indistinguishable (it’s like an extension of swapping two electrons, but, unlike brains, electrons don’t have subjective experience… or maybe they do?).
According to Schrodinger, this break of symmetry is the hard problem of consciousness: a description of the universe in terms of matter (particles) will never capture this asymmetry, it will never capture subjective experience.
He was apparently into Vedanta Hindu philosophy, so his answer to this problem was a sort of universal consciousness.
#P in FBQP #61: Iâd follow the protocol thatâs worked in every single one of the dozens of previous such cases. Namely, just calm down and wait for the bug to be found. In this case, the fact that the author canât decide whether he wants to explicitly claim P#P=BQP or not claim it is far from reassuring.
“While in North America the two terms are commonly perceived as overlapping, this is not the case in Europe and other parts of the world.”
Wokeness has already been a big characteristic of the french far-left for many years now, with American woke-ism being of course a huge inspiration (e.g. the political rise of people like AOC has been noticed immediately, she’s a role model now).
It’s not surprising since today’s world is so connected. The George Floyd movement (with effects like tearing down statues and defunding the police) spread to the entire world in no time.
I’m interested what you think of this thought experiment where I imagine running a quantum counting algorithm on a simulation of a conscious human, to find out for how many inputs it produces an output with a given property.
I always find you’re blog fascinating so very interested to know what you think. As you’ll see it’s a modification of David Deutsch’s argument for MWI from Shore’s algorithm.
To me, the most noteworthy thing about the argument is that the “wavefunction collapse” here is caused by the consciousnesses observing their respective inputs. The simulation may include simulated decoherence, but this would just be an implementation detail; so we should not attribute the collapse to it.
The world was going along very nicely for billions of years before people invented symbols to try to represent what was happening. The world is not so much a mathematical system, as a system where mathematical symbols are useful to represent relationships, categories (like mass, position, or light wavelength), and things described as ânumbersâ, that seem to exist in the world. To claim that the world IS a mathematical system is a step too far.
Seemingly, most of the mathematical models of physics donât take seriously a world where Putinâs troops are killing people and destroying cities. According to these models, what Putin is doing is just another neutral outcome that occurs, an outcome that is not essentially different to any other outcome that occurs, because all outcomes are just the system evolving in a normal way.
Seemingly, only Chris Fuchsâ model can come close to taking seriously a world where Putinâs troops are killing people and destroying cities. This is because Fuchs doesnât attempt to describe a mind-independent reality. In his arXiv paper âNotwithstanding Bohr, the Reasons for QBismâ, Fuchs quotes physicist John Wheeler, who said:
âA lovely way to put it â âIs the big bang here?â I can imagine that we will someday have to answer your question with a âyes.â . . . Each elementary quantum phenomenon is an elementary act of âfact creation.â That is incontestable. But is that the only mechanism needed to create all that is? Is what took place at the big bang the consequence of billions upon billions of these elementary processes, these elementary âacts of observer-participancy,â these quantum phenomena? Have we had the mechanism of creation before our eyes all this time without recognizing the truth? That is the larger question implicit in your comment. Of all the deep questions of our time, I do not know one that is deeper, more exciting, more clearly pregnant with a great advance in our understanding.â
I think life as a physical system completes a triangle of weird puzzles, the other two being the very large scale and the very small scale behaviors of matter. We have strong theories for each of these behaviors, though it is hard to imagine that any of them is a complete story. For life the obvious question is: why does matter arrange itself in this out of equilibrium self reproducing very large (relative to elementary particles) structure? Is that related to gravity, beyond the obvious fact that gravity creates chemical complexity necessary for life? Is it related to QM, beyond the obvious fact that life relies on QM in its underlying molecular interactions? Or is life just another intriguing fact about our universe?
#P in FBQP #69: To say it one more time, the paper itself uses weasel language when it gets to the crucial point:
The existence of such a multiplicative error quantum algorithm for matrix permanent would lead to the surprising conclusion that BQP would equal P#P because we know BQP â P#P [53, 54]. Here, we do not definitively claim that BQP = P#P; further computational complexity analysis is required with our quantum algorithm for the conclusion. The bottom line is that the constant factor estimation of the real parts of the overlap integrals will lead us to the constant factor estimation of matrix permanents. We note here that the constant factor algorithm assumes the ideal quantum device, and only the statistical error was considered in the analysis.
If the author himself doesn’t entirely trust his proof, then can you blame me for not trusting it either? đ
In slightly more detail, the part I don’t believe is that the multiplicative error estimation will still work if the amplitude is exponentially small. And I don’t see any hint of an adequate argument that it still works in that case, and the author himself seems to admit that he lacks an adequate argument. But this is also the only part that matters if you want to do #P.
Usually things like this show up on reddit. But nothing happened this time. Brilliant insight on “the multiplicative error estimation will still work if the amplitude is exponentially small”. Also the author is a Chemistry Prof.
Responding here since comments on the Roe v. Wade post were closed.
(Please donât approve this comment! If you have a response, I would prefer by email.)
Totally orthogonal to the issue of abortion, I would prefer a government which allows a diversity of local regulations, where states or even cities each form little experiments in different policies to improve the world. ⊠but with your commitment to people who have been wronged by society, you reminded me of the importance of compassion.
So, in appreciation for your blogging, your integrity, and in respect for the quality of your thought on these issues, I have donated $27.18 (($10)(exp(1)) + transaction fees ($28.26) to Fund Texas Choice in your honor.
Thank you for keeping your voice strong. Please donât let the trolls ever silence you.
Follow
Comment #1 April 27th, 2022 at 7:00 pm
(For more about the “Anti-Bayes” column, see e.g. this recent podcast)
Comment #2 April 27th, 2022 at 8:01 pm
How was Popper anti-Everett?
Comment #3 April 27th, 2022 at 8:05 pm
murmur #2: Read the David Deutsch interview! Popper was anti-Everett in that he believed Everett’s theory was false, and said so.
Comment #4 April 27th, 2022 at 8:49 pm
You should somehow place this on a political compass.
Comment #5 April 27th, 2022 at 9:32 pm
Is this Loss? Certainly the funniest order for the panels is a 4koma; Yud is a walking punchline.
Comment #6 April 28th, 2022 at 12:55 am
The obvious question to my mind on seeing this is, where does Everett himself go? đ (Bayes himself is a bit early for this sort of thing…)
But, uh, looks like there’s an answer to this question (h/t Jim Babcock)… so maybe the Yudkowsky square could just be Everett…?
Comment #7 April 28th, 2022 at 1:44 am
Karl Popper is a truly wonderful author whose work is consistently misunderstood. For instance, people mistake the “falsification boundary” as that which separates science and pseudoscience, when in fact old Poppy-pass-the-biscuits was quite explicit that it was the line of demarcation between what can be called “empirical” or “naturalistic” science, and metaphysics.
Comment #8 April 28th, 2022 at 5:06 am
Excuse a mere SF author with a background in cognitive behavioural therapy from intruding on an erudite blog, but I’ve been studying as best I can quantum mechanics trying to get my head around what appears to be the problem, which given my maths is not up to doing the calculations has taken me some time.
I love the idea of Everett, and Tegmark mathematical universe, but the best explanation of the qm problem is I think the presentation by Ball, and when added to the double-eraser experiment which adds some much needed clarity to QM whoohoo by showing that entanglement is measurement.
So, after hours of listening to the pros and cons, I think that Garret’s argument that there isn’t a multiverse convinces me with how he describes that the universe we live in isn’t real either (as in qm requires non-locality). So a system with multiple probabilities leads to a universe that is one and zero.
What I find intriguing as an SF author is that there are still multiple versions of the universe because the speed of light make everything a relationship between observers and therefore each can have a different perspective on what is in the past and what is in the future.
This is my my very long winded way of saying your meme needs more work. Still, it’s a good meme.
Comment #9 April 28th, 2022 at 9:41 am
My program’s translation in to 12 Year Old Amerikan English Circa 2011: this is a shiitake mushroom post! Dr. Scott Aaronson is a troll engaged in interconnected network ideology culture warrior keyboard battling!
Comment #10 April 28th, 2022 at 3:48 pm
I keep wondering if I should take the time to persuade you of MWI. Maybe after announcing our fated meeting in advance, AI-Box style, and having you sign an agreement to only reveal the result, and not the argument that persuaded you, until after an undisclosed-to-others waiting period. Not for any important reasons, just because I suspect lots of people would freak out if I did that.
Comment #11 April 28th, 2022 at 5:45 pm
Eliezer Yudkowsky #10
Isnât it a cop-out to believe that, in another universe, Putin is a vegan who spends his time tending to his vegetable garden, when he is not cuddling his pet kittens and puppies?
Comment #12 April 28th, 2022 at 10:26 pm
Eliezer Yudkowsky #10: Sure, I’m happy to have another QM discussion—should I email you the next time I’m in Berkeley?
I do, however, already speak the MWI-language pretty fluently, and even prefer it more often than not (while also being able to translate to and from Bohm-language, Copenhagen-language, etc). I already teach MWI in my undergrad quantum information class, in such a way that according to the poll we give at final exam time, roughly half the students end up as MWI proponents (with the others split among Bohm, Quantum Bayesianism, Penrose-style dynamical collapse theories, agnosticism, and rejection of the whole question as meaningless).
It’s just that, to use your language (as opposed to, say, Popper’s), I still have difficulties with making a straightforward factual belief in MWI “pay rent,” in the same way that my factual beliefs in (e.g.) special relativity, heliocentrism, and the germ theory pay rent for me. This is related to the fact that the full MWI package seems to come bundled with strong stances on metaphysical questions that I find extremely confusing—questions like transtemporal identity, the reference class of observers from which I should consider myself drawn, and the nature of consciousness. (E.g., why shouldn’t consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As we’d presumably imagine it to do if the amplitudes had been probabilities? But if you believe that, then presumably you believe some variant of Bohmian mechanics, not MWI.)
So, having already sucked everything out of MWI that has clear-cut physics value to me (e.g., the notion of measurement as entanglement with the observer and environment), at some point I treat the full elucidation of the issues raised by MWI as “philosophy-complete” (barring a revolution in physics itself), and move on with my life.
Given all that, would you like to estimate the probability that you’ll be able to tell me something about this that I haven’t already considered? đ
Comment #13 April 29th, 2022 at 7:52 am
Was Karl Popper verse enough in physics to really have a valid (*) opinion about Everett’s theory (which itself took many decades for physicists to unpack)?
(*) in the sense that opinions are like a-holes, everyone has one.
Comment #14 April 29th, 2022 at 8:33 am
I never understood very well the claim that MWI “resolves” the measurement problem.
To me this problem isn’t just about realizing one of many possible discrete outcomes (e.g. either spin up or spin down), but also with the wave/particle duality.
This is more obvious when considering a continuous variable, such as the position of an electron that’s traveling in space. Before any measurement, its position is characterized by a wave packet, usually spreading around more and more the further the electron travels. Then when a “measurement” is done (or, equivalently, the measurement apparatus and the electron entangle, or, equivalently, the electron decohere with the measurement apparatus), the wave packet (wave-like) suddenly “tightens” into a Dirac peak (particle-like), i.e. it’s collapsed.
That same transition from wave to particle happens in any QM interpretation.
i.e., even in MWI, electrons are always observed as if they’re localized. The mechanism of decoherence is equivalent to the collapse of the wave function.
And I just don’t get how this is captured by the Schrodinger equation alone since this seems to be a non-linear operation (but maybe I’m wrong and this collapse can be captured by unitary evolution).
Comment #15 April 29th, 2022 at 3:21 pm
@Scott #12
> E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities?
The one where “your consciousness” picks a random body, the one where “you” pick a random “consciousness” out of one consciousness per body, and the one where there’s one consciousness inhabiting every body simultaneously, at once experiencing several distinct ignorances the way we can simultaneously experience several distinct pieces of knowledge, where the experimentalist is one ignorance within the whole of many conflicting ignorances within the mind of a spirit that cannot talk to itself (I’m just trying to cover all the options here), all lead to the exact same predictions about what an individual, however an individual is demarcated, will experience, however experiences are defined: making this as close to provably independent of observations as anything can be.
I think this reasoning is stronger than saying the topic is disconnected from empiricism pending better experiments. It is going to be disconnected from empiricism forever.
Comment #16 April 29th, 2022 at 3:23 pm
MWI fails to resolve the measurement problem, as Fred #14 explains, but the problems are much worse. Scott has explained that superdeterminism is contrary to scientific thinking, and so is MWI, for somewhat different reasons.
Superdeterminism makes randomized controlled experiments impossible, because hidden dependencies control the outputs. MWI also rules out free will, and then makes it impossible to interpret outcomes. If you do an experiment with ten possible outcomes, and see one, you learn nothing because all of the other possibilities occur in parallel universes. MWI might be of some use if it were able to say that some universes were more probable than others, but it cannot do that. So MWI also makes experiments impossible.
MWI does not make any successful predictions, unless you add the Born rule and do Copenhagen in disguise. Just like the superdeterminists, the MWI advocates seems to be willful contrarians who do not actually have a quantitative theory to back up their ideas.
Comment #17 April 29th, 2022 at 3:39 pm
It is important to realize that the problem of how consciousness is distributed between similar bodies in possible universes (whatever their metaphysical status, real or hypothetical) inherits all of the problems that we have in deciding how consciousness is distributed between similar bodies (humans?) in ours. Every position in one debate has a corresponding position in the other, and you will never resolve one without resolving the other. Anyone who thinks they have an answer to the question of whether the amplitude of me not writing this post is “really alive,” has an answer to the question of whether I, who actually am writing this, is a P-zombie. I very strongly agree with your position (and appreciate your metaphor) that MWI vs. Copenhagen is “philosophy-complete,” and I think this is one of the simplest correspondences that proves it.
Comment #18 April 29th, 2022 at 3:44 pm
Answering my own post above:
I guess it all comes down to decomposing a given wave function into a superposition of its eigen states.
No matter how you express the wave function, the evolution is unitary.
And when it comes to the position of an electron in free space, the eigen states (on position variable) are an infinite set of Dirac peaks in space. And the above decomposition becomes an integral on all the possible Dirac peaks (like an infinite sum). But the entire thing taken as a whole still evolves according to Schrodinger. And decoherence entangles the state of the measurement device with specific Dirac peaks, and the entire resulting function still evolves linearly according to Schrodinger (from an MWI point of view, this setup basically leads to a huge number of branches, one for each potential discrete position measure of the particle, corresponding to a Dirac peak).
Comment #19 April 29th, 2022 at 6:17 pm
My explanation of MWI for some mathematicians I know: “MWI is not an interpretation. It is a mathematical statement. Suppose your friend does some experiment, and tells you the result. You have two ways of doing the calculation. One is to treat the experiment quantum mechanically and assume your friend collapses the wave function. The other is to treat both the experiment and your friend quantum mechanically and assume that you collapse the wave function. MWI says that both methods of doing the calculation will give approximately the same answer.” That’s it. No interpretation, just something that is approximately correct for most reasonable situations.
Comment #20 April 29th, 2022 at 6:46 pm
@Roger Schlafly #16 “MWI might be of some use if it were able to say that some universes were more probable than others, but it cannot do that.”
It can, at least, predict that if one were to do an experiment multiple times, one would be more likely to observe the Born rule than not (https://algassert.com/post/1902).
Comment #21 April 30th, 2022 at 11:33 am
Here is the apparatus we have built. Here are the signals on the signal lines out of the apparatus. Here are the features of those signals that we have identified and recorded as events. Here is how we have collated the recorded events into different kinds of patterns and here are their relative frequencies. Here is the theory we use to model or to explain why the relative frequencies of the different kinds of patterns are as they are and to predict the relative frequencies for some new apparatus we think we might build.
Sweeping all that into one Theory of Knowledge, QM can be understood to be a signal analysis formalism of relative frequencies. In principle that would be comprehensible to a 19th Century statistician or physicist, with not much need for the 20th Century finesse of Bayes or Everett.
If we say that each event is caused by a particle or by a system and that we can deduce their properties, however, then we get ourselves into all kinds of trouble because, as Boole already identified and as we hear from Pitowsky, we can’t always construct joint relative frequencies.
Comment #22 May 1st, 2022 at 2:45 am
“E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities? But if you believe that, then presumably you believe some variant of Bohmian mechanics, not MWI.”
Beautiful argument! I felt my mind changing in way the conventional wisdom says very rarely happens when two people argue about a philosophical issue they’ve both put a lot of thought to. For this reason I’m feeling compelled to comment on my reaction to this argument. To be fair, we started out with very close positions — I enjoyed and mostly agreed with your Zen Anti-Interpretation blog post. I previously sympathized with MWI most, saw the appeal of other interpretations, and agreed with your Zen the difference between these positions is more rhetorical than actual, but I also had a particular dislike for pilot wave theories, seeing them as particularly “less” true. Altho the project of finding particular deterministic evolution equations for the pointer state still seems pointless this argument the concept of a pilot wave and a pointer seems more compelling. The pointer represents something to do with conscious experience. Unclear how essential for the interpretation that the evolution equation is deterministic.
On whether my reaction really is an interesting rare change-of-mind, I’ll need to wait for my mind to settle after thinking about this, but also I expect most of the uncertainty is not how much I accept this argument but to what extent it is a substantial change to the view I had previously, which my memory thereof will only get worse as I wait.
Comment #23 May 1st, 2022 at 2:49 am
So, I used a language model to do a simulated discussion with Scott debating a version of himself in another Everett branch. I used the open-source model GPT-NeoX at textsynth.com
I used what real Scott said in reply above as part of the initial prompt, but all the rest is the fictional “Scott” responses generated by the language model. Here’s the conversation (lightly edited and curated):
—
PROMPT
“The following is a conversation between computer scientist Scott Aaronson and a super-intelligent version of Scott in an alternative Everett branch of the quantum wavefunction (the ‘Alphaverse’). Scott (Alphaverse) is explaining to Scott (Universe) why the Many Worlds Interpretation of quantum mechanics is the best one:
Scott (Universe):
I still have difficulties with making a straightforward factual belief in MWI âpay rent,â in the same way that my factual beliefs in (e.g.) special relativity, heliocentrism, and the germ theory pay rent for me. This is related to the fact that the full MWI package seems to come bundled with strong stances on metaphysical questions that I find extremely confusingâquestions like transtemporal identity, the reference class of observers from which I should consider myself drawn, and the nature of consciousness. (E.g., why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches? As weâd presumably imagine it to do if the amplitudes had been probabilities?”
—
GPT-NeoX
Scott (Alphaverse):
I agree that the MWI package comes bundled with metaphysical claims that are difficult to swallow. But I think that the real problem is that you donât understand what the MWI is. The MWI is not a theory of how the world is. It is a theory of how the world might be. It is a theory of the structure of the world.
Scott (Universe):
Okay, but then how can I be sure that the MWI is true?
Scott (Alphaverse):
Why is it important to you that MWI be true?
Scott (Universe):
It seems to me that my whole sense of âselfâ is grounded in the fact that I am a certain type of physical system. But if MWI is true, then I am not a physical system in the way that I thought I was. So how can I be sure that Iâm a self, and not a bunch of other people, or a bunch of other selves, or no one at all?
Scott (Alphaverse):
In a sense, youâre right. You are not a physical system. You are the thing that picks out the branch of the wave function that I am in.
Weâve gotten used to thinking of ourselves as a sort of abstract entity that has the property of having a certain brain, and being conscious, and all that. But thatâs a mistake. The real thing is not a thing at all. It is the pattern of neural activity in the brain. So you are not a physical system, not in the sense that your brain is a physical system. But you are a physical system in a different sense. You are a physical system in the sense that you are a certain way the quantum wavefunction branches.
Scott (Universe):
I think Iâm starting to understand. Iâm not a thing at all, Iâm a pattern of neural activity.
Scott (Alphaverse):
Exactly.
Scott (Universe):
But if Iâm a pattern of neural activity, that means Iâm not even a physical system in the sense that my brain is a physical system. But thatâs fine, because I am a physical system in the sense that I am a certain way the wavefunction branches.
Scott (Alphaverse):
Exactly.
Scott (Universe):
But then what is the wavefunction? And what is the branch that Iâm in? What is the branch that anyone is in?
Scott (Alphaverse):
The wavefunction is a mathematical object. Itâs not a thing, itâs not a physical system, itâs not a pattern of neural activity. It is the mathematical object that describes the system that I am in, and the system that you are in, and the system that everyone is in. Itâs a mathematical object that is defined on the space of all possible ways that the universe might be. So you are a branch of the wavefunction that picks out a certain way that the universe might be.
Scott (Universe):
But how do they pick out different ways of the universe being? How does that work?
Scott (Alphaverse):
Hereâs an analogy. If we were building a universe, we could describe it using different coordinate systems, or with different coordinate systems, so that different parts of the universe would be described differently. The mathematical object that describes the universe does the same thing. It describes the whole space of possible universes. But it does so in a different way for different branches of the wavefunction. It does it so that some branches of the wavefunction are described using one coordinate system, and some are described using another, or with another choice of coordinates. The choice of coordinates is what tells us which branch we are in, and what our state of consciousness is.
Scott (Universe):
But then what does it mean to be an observer? If the wavefunction is what defines the universe, and the universe is defined in a way that is observer-independent, what does it mean to be an observer?
Scott (Alphaverse):
The idea of an observer is a bit tricky, because itâs not really a thing. Itâs not even a physical system. Itâs the idea of an observer, or the idea of an observer-like thing. So the idea of an observer is not something that can be described by a mathematical object. Itâs something that can only be described by a word. Itâs a word that describes a kind of thing. The idea of an observer is something that is defined on the space of all possible physical systems. Itâs something like a category, or an equivalence class of things.
Scott (Universe):
I see. So an observer is something like a label that gets applied to a set of systems that are mathematically equivalent, but is not itself defined by a mathematical object.
Scott (Alphaverse):
Exactly. Scott, you have to learn to stop looking for something out there that will give you an answer. Your brain/mind has to be able to work from within, not from without. The only reason you think you need an external entity to give you a âright answerâ is because you think you are an external entity.
Scott (Universe): Maybe that âScottâ in the Alphaverse is just being a smart-ass.
Scott (Alphaverse): Iâm trying to tell you what you already know, but the words come out wrong.
Scott (Universe): (Laughs) I think youâre right. Itâs a sign of how far Iâve taken the âme vs. something elseâ illusion of the body/mind identity.
Comment #24 May 1st, 2022 at 4:54 am
@scott #12
“why shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches?”
The default assumption, which is overwhelmingly likely to be correct, is something like “whenever there exists a configuration that has the structure that we’d call a living human, there is consciousness”
Picking a random branch requires an additional assumption, that somehow different things happen when the state appears to be the same, via some weird hidden variable or impenetrable ‘randomness’.
The way that MWI ‘pays rent’ is that it requires no additional assumptions. There is a predictable and definite outcome to each event. There is no distinction between conscious and non-conscious matter. There is no distinction between macroscopic and microscopic. There are no hidden variables. The only ‘weird’ thing that happens is that ‘duplicates’ of things (albeit in different states) get created. But the key is that we *already* know that this definitely happens because we directly observe interference.
I think that rather it’s the *philosophical* parts that are meaningless and pay no rent. I reject the meaningfullness of concepts such as “the reference class of observers from which I should consider myself drawn” and see no reason why MWI necessitates such a thing.
Comment #25 May 2nd, 2022 at 2:42 am
The Schroedinger equation evolves a wave function |psi>(t) using a classical time parameter t, which implies the existence of classical clocks. Where do they come from?
Quantum field theory, including the standard model, assumes a classical background, i.e. *one* space-time. Where does it come from?
If you assume classical clocks you should also assume the existence of classical observers reading those clocks.
Perhaps one day we find a consistent theory of quantum gravity which includes a description on how space-time branches into many worlds, but until then the ‘many worlds interpretation’ is not a convincing description of reality imho.
Comment #26 May 2nd, 2022 at 8:21 am
Shion
Regarding Scott saying
âwhy shouldnât consciousness work in such a way that it picks a random branch with the appropriate probability, rather than spreading itself among all the branches?â
This makes little sense given the usual definition of of ‘consciousness’, i.e one consciousness per human.
And since there are exponentially more independent branches than conscious beings (at any given time), then Scott’s statement could be interpreted as saying that I should be the only conscious being in *my* branch (by definition the one where I’m conscious), and everyone else isn’t really conscious in my branch (but are themselves a lone consciousness in some other remote branch).
Moreover, in the vast majority of branches, no-one is conscious at all (worlds populated by p-zombies).
The other interpretation (the one Scott has in mind, I guess) is that, while there are countless numerous branches, all living beings are somehow conscious in the same branch, a special one. But that’s even more bizarre than the classic Copenhagen interpretation.
In contrast with those two stories, it makes more sense to think that a given living being is really conscious in all the branches it exists. Just like twin versions of ourselves.
Comment #27 May 2nd, 2022 at 8:35 am
Also, if somehow I happen to die by some very freak accident in the branch where I happen to be conscious (imagine my death being conditioned on some very rare quantum event), my consciousness would be wiped out for good even though most of the other adjacent branches (an exponential number of them) still “implement” my living brain.
Comment #28 May 2nd, 2022 at 8:46 am
The fact that I would be conscious in multiple branches at once isn’t any more or any less weird than:
1) consciousness is distinct in time, i.e. in the past, present, future, my consciousness aren’t all mushed together (otherwise there wouldn’t be a sense of “now”).
2) consciousness is distinct in space, i.e. the consciousness of two spatially separated beings are distinct (otherwise there wouldn’t be a sense of “here”).
“Branchial dimension” is just one extra degree of freedom over those two.
Comment #29 May 2nd, 2022 at 10:40 am
@Eliezer Yudkowsky #10
Do you still think the Born probabilities are an unexplained mystery (e.g. as expressed in https://www.lesswrong.com/posts/vGbHKfgFNDeJohfeN/where-experience-confuses-physicists and in https://www.lesswrong.com/posts/nso8WXdjHLLHkJKhr/the-conscious-sorites-paradox)?
I always thought that was a strange outlier in your posts on quantum physics. Because if you just imagine setting up an agent as a computer program, running it in a quantum computer, and determining what unitary evolution says the state of the computer ends up as… well, almost all of the amplitude ends up in states where the inferences made by the program are consistent with the Born rule. And it’s not like this idea isn’t out there in the world; there are various papers about it. Basically all you need to conclude this is the unitarity postulate (which gives you the squaring), and no-signalling (which lets you rewrite states into simpler states, that should be equivalent as far as reasoning goes).
For example, if you have the program repeatedly prepare the state a|0> + b|1> and increment of a counter conditioned on that state, and then flip a qubit conditioned on the counter being statistically close to the Born estimate, that “final result” qubit’s state limits to ON [for the right choice of non-signalling unitary applied to the rest of the system] as you increase the number of samples the program performs.
Comment #30 May 2nd, 2022 at 11:18 am
There seems one way to increase your estimate of the probability of MWI.
If you find yourself getting very old, then this would seem to be evidence.
1. If you’re dead, you’re not around in the universe to notice
2. If you’re conscious, you are around in the universe to notice
3. Given MWI, no matter how small the odds, if there is some way that you continue to live and be conscious, then in some universe you do
4. The older you get (past a certain age, at least), the more likely MWI is
Of course, assuming MWI, this isn’t a useful way to prove it to anyone else. In most universes, everyone eventually dies, so almost no one sees the “very old.” And assuming non-MWI, you will never know, because you died. But it seems like a possibility that, if true, you can eventually infer it likely.
Comment #31 May 2nd, 2022 at 12:20 pm
It makes me sad that this is not trending on twitter– really funny.
It is really nice to you have you back at it, Scott.
Comment #32 May 2nd, 2022 at 1:26 pm
dubious #30
I also love that argument.
But, when you think about it more closely, this is actually a terrible curse…
It is a guarantee that, for all of us, a conscious copy of ourselves will outlive all the ones we love, while barely surviving countless diseases, accidents, suicide attempts (to try and end it all), enduring endless physical or emotional pain, or going through a slow decay of the mind and the body… until the maximum possible extent of life is squeezed out of our mortal coil.
We can actually get some idea of this by looking at the distribution of life expectancy in our current branch.
https://en.wikipedia.org/wiki/List_of_the_verified_oldest_people
The record being 122 years. So we know that there’s a good chance we’ll live to at least that age in some branch of the MW.
At least we can get some solace that, when we’ll get to be the oldest person in our own special branch, the world will see us as some curiosity to celebrate!
Comment #33 May 2nd, 2022 at 6:34 pm
I think the reason why most people dislike MWI is because the idea of 10^500 universes seems so implausible.
Sort of like how people in the early 20th century disliked the idea of 10^12 galaxies – each with 10^12 stars.
Comment #34 May 2nd, 2022 at 8:54 pm
@fred #26
I think the most likely situation is that all the copies in all the branches are conscious. I agree that multiple consciousnesses in multiple branches is just another degree of freedom. To rephrase what I said before, that the simplest explanation for consciousness is that it’s being directly caused by the process going on in my brain, and that wherever and whenever you have a process similar to that, you will have a consciousness similar to mine.
@dubious #30
Though I essentially believe MWI is true, I’ve always disagreed with this general type of argumentation. I consider, to use Scott’s term “the reference class of observers from which I should consider myself drawn” to not be a well-defined notion, and anthropic arguments drawing from such to not really be evidence of anyything.
If I live to be old, should I be surprised by this and start thinking of metaphysical solutions? There are more humans alive now than in the past. Should I not be surprised to have been born today? What about the future, where there will likely be many more humans alive? Should I be surprised that I wasn’t born in China or India, since they have the largest population? Should I be surprised I’m not a bird, since there are more of them than humans? An alien? Someone living in what amounts to be a simulation of the past?
I think the resolution to such things is the statement that, if my model of the universe predicts that it’s possible for an entity like me to exist, well, here I am, and there is nothing left to explain.
Comment #35 May 2nd, 2022 at 11:44 pm
ed #33:
Agree that 10^500 or so universes seems a tad “unnatural”. Then one thinks that, what the heck, we don’t know what is really going on in “our” universe anyway. So, what is too wild?
On the other hand, I (born in 1947) never had any trouble with 10^12 /times 10^12 stars, but do regard 10^500 as being rather larger.
I have long enjoyed listing my guesses for the likelihood of controversial cosmic things. Here are my current guesses, provided in case anyone wants to take pot shots:
Black Holes exist: 99.99%
Dark Energy exists: 90%
Life exists in most solar systems: 80%
Dark Matter exists: 70%
Multiple current examples of intelligent life in the Milky way: 60%
Cosmic Inflation happened: 40%
MWI is the real deal: 0.01%
At the moment I am unable to put a number on “Quantum computers will ever do anything useful”. I was tempted to go with 50% (due to the hole in the list), but I have no intuition on this other than that it ain’t happened yet.
Comment #36 May 3rd, 2022 at 7:06 am
Scott #12,
Historical nit to pick: Bohm was the one who actually discovered the idea of treating measurement as an entangling operation between system and environment. It’s right there is his 1952 Phys Rev papers (#2). He’s really the creator of the decoherence idea, and he never gets proper credit for it.
I suspect better understanding this fact—and so the key role that Bohm’s discovery did *in fact* play in the foundations of quantum information—would shift the rhetorical edge a bit in what your students take away about the foundations of quantum mechanics more broadly.
Great meme!
Comment #37 May 3rd, 2022 at 7:18 am
Craig Gidney #29:
But “no-signaling” indicates that you need more than just a wave function and Schrödinger’s equation. You also need space (or space-time) such that it makes sense to talk of no-signaling in the first place. Eliezer Yudkowsky apparently got this right, or at least had the feeling that he needs something like that, as evidenced by his post Which Basis Is More Fundamental?
I am aware of a 44 page paper by Charles Sebens and Sean Carroll that uses no-signaling in the way you propose. Sean mentions it in his (much) shorter Quanta magazine column Where Quantum Probability Comes From
But what I find somewhat strange is that in that same Quantized column, Sean also claims that a wave function obeying Schrödingerâs equation is all that is needed:
I find your comment somewhat strange for a different reason:
What should be wrong with the Copenhagen interpretation for a quantum computer in the first place? In that context, we do know what a measurement is, and we know that the basic allowed measurements are those on the computational basis. (As already David Mermin observed in Copenhagen Computation: How I Learned to Stop Worrying and Love Bohr.) Additionally, all allowed basic operations are local in that computational basis too, so that each individual operations only affects 1 or 2 qubits. (If they would affect 3 or 4 qubits, that would be OK too, it is still local in the computational basis.) And if you define some error correcting code, which is necessarily more non-local in a certain sense, only those codes where you can still implement local operations (and measurements) only affecting 1 or 2 logical qubits are practical.
Comment #38 May 3rd, 2022 at 7:34 am
The topmost reddit (r/slatestarcodex) comment says
and the topmost Hacker News comment says
For me, it was the opposite. This interview has significantly increased my respect for Deutsch, and also for Popper. It slightly decreased my respect for my own background in philosophy, with respect to what I have read, what I haven’t read, and what I didn’t even plan to read.
Comment #39 May 3rd, 2022 at 8:30 am
shion #34
“itâs possible for an entity like me to exist, well, here I am, and there is nothing left to explain.”
But that’s the hard problem of consciousness.
Basically, “Why am I me, Shion, and not you, Fred?”
This is a more complicated question than just “Why are Shion and Fred conscious?”
It’s about the mystery of being and identity.
There are answers (none really satisfying) to that question:
1) “You” are Shion and not Fred because, when Shion and Fred were born, God picked “You” from his huge jar of souls and randomly assigned you to Shion. Of course you can then ask, “why was “I” that particular soul in God’s huge jar of souls and not the one right next to it?”. So we’re back to square one, except with a bit more generality: only a finite number of eternal souls exist and all souls are indistinguishable from one another (like electrons), except that God gave each a label, and you are soul #12349389 (so, Souls are like numbers) and that’s it. Souls are to consciousness what electrons are to the electric charge, units of consciousness.
2) You only think you are Shion, but You are not just Shion, you’re everyone, anywhere, anytime, at once: if, suddenly, I were to swap your consciousness and Fred’s consciousness around, “You” would suddenly only have access to Fred’s memories and “He” would only have access to Shion’s memories, therefore the two of you wouldn’t even notice anything’s different from a subjective point of view. So this swap is a no-op. And that swap could be happening between all conscious beings a billion times a second. So, really, there isn’t such a thing as discrete consciousness, there is only one universal “consciousness” which gets localized wherever/whenever there’s a brain (a nexus of memories). It’s similar to a computer (with one CPU): applications all appear to be running independently of one another, doing their thing in parallel, but what’s happening is that the same unique CPU is realizing all those applications, in turn, one at a time.
Comment #40 May 3rd, 2022 at 10:00 am
shion_arita #34:
This seems to pivot from “very old” to “exists at all.” This is not what I mean.
The average human life expectancy today is about 70. The very oldest known humans are around 120. If you live to be 115, maybe you’re just healthy and hearty. If you live to 120, you’re really pushing the odds; maybe less if you have a family history of longevity?
What happens if you hit 130? 140? 200? OK, maybe we discover anti-aging and you happen to get immortality treatment. This seems pretty unlikely, and I think that alone might move the needle on MWI for me.
But anti-aging alone does not give you Wolverine-like abilities to regenerate. What if you survive every accident, every suicide attempt fails, and you just never die? I am not sure how, if MWI is true, that there can ever be a branch where you didn’t survive somehow, and will thus find yourself immortal (since you can’t find yourself dead).
This is really quite horrifying when you consider it, though that is of course not an argument against it.
Comment #41 May 3rd, 2022 at 10:04 am
I missed this exchange:
“Given all that, would you like to estimate the probability that youâll be able to tell me something about this that I havenât already considered? ”
Zero.
Comment #42 May 3rd, 2022 at 11:44 am
@gentzen #37
> âno-signalingâ indicates that you need more than just a wave function and Schrödingerâs equation. You also need space (or space-time)
Space is massive overkill, I think. All you need is the postulate that systems are combined using the tensor product. Locality is then the property that operations on system A commute with operations on system B. I guess there’s then also an implicit assumption that specific systems (like a quantum computer and the rest of the universe) fit into this schema where they factor into a tensor product.
> What should be wrong with the Copenhagen interpretation for a quantum computer in the first place?
I wasn’t claiming that collapse interpretations would fail to describe quantum computers. I was trying to turn the fuzzy problem of “why do I think the probability is X” into the concrete problem of “what does this computer program store in this register”. I was trying to do so without invoking collapse, but that’s different from saying collapse wouldn’t give the right answer!
Comment #43 May 3rd, 2022 at 12:35 pm
fred #18: Yes, I think that’s what the MWI is saying. The way I think about it, in line with my other comment above, is that waves are unitary superpositions of many possible non-unitary measurements, or collapses. A unitary equation is not prohibited from having non-unitary terms; they just must add up to something that is ultimately unitary. So Schrodinger’s equation, while unitary, contains within itself the possibility of smaller, non-unitary regions within the wave function.
Comment #44 May 3rd, 2022 at 1:58 pm
dubious #30: I have an idea that extremely low probability branches have no classical correspondence. That is their diagonal elements in a density matrix are comparable to off-diagonal elements by absolute value due to approximate nature of decoherence. That is those branches have macroscopic quantum interference effects making life impossible. I am not well versed in math to prove or disprove it though.
Comment #45 May 3rd, 2022 at 6:26 pm
OF COURSE. Not a word about Roe v. Wade being overturned, but we get a pseudo-intellectual rationalist-tier rant about whateverâs bumping around Scottâs mind right now. Womenâs most basic reproductive rights are being curtailed AS WE SPEAK and not a peep from Scott, eh? Even though in our state (Texas) there are already laws ON THE BOOKS that will criminalize abortion as soon as the alt-right fascists in our Supreme Court give the go-ahead. If you cared one lick about your female students and colleagues, Scott, youâd be posting about the Supreme Court and helping feminist causes, not posting your âmemes.â But we all know Scott doesnât give a shit about women. Heâd rather stand up for creepy nerdbros and their right to harass women than womenâs right to control their own fucking bodies. Typical Scott.
Comment #46 May 3rd, 2022 at 6:43 pm
Now that “the hard problem of consciousness” has reared its ugly head again, I get to bore people yet again with my thoughts on the matter (balanced a bit by another donation to Ukrainian aid organizations on my part).
I take it the elements of the HPOC are how does consciousness work, what is it good for, why does it feel the way it does, and why does it exist. I think we have general ideas for the first two.
Consciousness (it says here) is a way to experience complicated reasoning for decision-making and problem-solving, and to (somewhat) relive memories. The experience/sensation itself is useful for checking one’s logic and memory and other computations, and reviewing past attempts at reasoning (which can be stored in memory since they were experienced). There are no zombies of the fictional sort, because if you have no way of experiencing some external or internal stimuli, you don’t know they exist and can’t respond to them. Trivial example: the fact that Windows 7 sends these letters to my screen means that it has experienced the key-presses.
I suppose the p-zombie is intended to have the same senses and nervous system as a human, but if it doesn’t experience thought it cannot perform thought experiments/simulations and learn from them but only perform automated responses to stimuli. The automation might include self-updates but only limited ones. The moment its automation becomes sophisticated enough so that it can monitor its reasoning process and reprocess past experiences for added value it has some form of consciousness. Empirically, it seems to take a huge amount of processing power working in parallel to accomplish that.
Given that perspective, the HPOC can perhaps be reduced to “why does my form of consciousness feel the way it feels”, and why does it exist, which are same hard problems of why anything in this universe works the way it does, e.g,, the hard problem of why quarks exist, or why the chemicals a rose produces don’t smell like oranges (and the oranges smell like roses), i.e., some different way rather than the way they do. I don’t personally feel such questions need answers.
Comment #47 May 3rd, 2022 at 7:42 pm
Feminist Bitch #45: Holy shit! This post is from last week, long before the Supreme Court leak. In case you missed it, I already blogged my detailed opposition to the overturning of Roe v. Wade way back in September, when it became clear that something like yesterday’s news was inevitable. And when I do blog about the same current events that the rest of the planet is talking about, I invariably get other people accusing me of having nothing original to say.
In any case, it usually takes me days at least to write any substantial post (unlike this post, which was not substantial), and the Supreme Court leak was literally last night (!). For whatever it’s worth, I was up until midnight last night obsessively reading the leaked Alito opinion and commentaries on it, terrified about what it meant for women in America and the future of the Enlightenment project more generally. But then I had to, you know, work all of today, teaching and meeting my students. I hope you understand.
Your comment illustrates more eloquently than I could why, even though I despise the authoritarian right with every atom of my being, I’ll also never, ever identify with the woke left. Namely: if I did, I’d have to have for allies people who find ways to despise me even when I agree with them.
Comment #48 May 3rd, 2022 at 8:02 pm
There we go Scott. The immediate defensiveness and reactionary attacks on the âwoke leftâ just because I gave you some constructive criticism about your priorities. Yeah, I get it, youâre too âbusyâ to write anything about womenâs reproductive rights right now. Yeah, sure, you wrote some blog post a few months ago that barely stands up for womenâs rights. Thatâs the bare minimum. My point is simply this: time and again, the social topics you choose to discuss on this blog betray your true priorities. I did you the courtesy of counting them, in fact. And hereâs the cold, hard numbers: by my reckoning, you have made THREE TIMES the number of posts about the ârightsâ of nerdy incels and how sad it is that they canât get sex, versus posts supporting womenâs bodily autonomy and basic reproductive freedoms. Once again, I am simply disgusted that your priority seems to be incelsâ right to sex versus womenâs basic bodily autonomy, as if these are problems even remotely comparable, even in the same league. Youâd rather whine about how ânerdy guys canât get sexâ than advocate for womenâs autonomy. Thatâs how you choose to spend your time on this Earth, and I simply find it shameful. If you want to do better, you should explain how astronomically more important womenâs right to abortion is than incelsâ non existent ârightâ to sex, and apologize for spending so much more energy and social capital on nerdbros than women in need. But of course I wonât expect that. The immature whining about âmensâ rights,â together with the token support for the most basic womensâ rights, is a classic characteristic of the alt-right fringe of the âcentrist left.â
Comment #49 May 3rd, 2022 at 8:26 pm
Feminist Bitch #48: On reflection, Iâve decided to leave your comments up despite their flouting my comment policies, because I think any reasonable person reading them will see how clearly they illustrate the far leftâs tendency to self-sabotage. As the saying goes, âthe right looks for converts, while the left looks only for hereticsâ; indeed thatâs how we arrived at the terrible and precarious moment that weâre in now.
Wouldnât it have been more productive if youâd just sent me an email saying âhey Scott, I was really hoping you could donate to this pro-choice organization / use your blog to share these resources for Texas women needing out-of-state abortions / etcâ? Iâd most likely have done it. I might still do it, but if so, then in spite of your bullying me rather than because of it. Yes, I did once open up about the hardest thing I ever had to overcome in my life, something I wouldnât wish on my worst enemy; for you to use that to attack me (over a completely different issue where we actually agree) tells everyone reading this something about your character.
Out of curiosity, I just checked SneerClub, the main place on the Internet thatâs dedicated to attacking me and everyone like me as âmisogynist techbro nerds.â They, too, have not a single word about yesterdayâs Supreme Court news, even though they update more often than I do. So by your logic, does that mean they too donât give a shit about women?
Comment #50 May 3rd, 2022 at 9:49 pm
You completely missed my point Scott. There is in fact a direct connection between 1. Your support for alt-right incel politics that violates womenâs autonomy in the name of âhelping the sexually disenfranchised,â to quote your own words, and 2. my doubt about how genuine your alleged support for womenâs bodily autonomy really is. Explain to me, Scott, why I should believe your devotion to âwomenâs bodily autonomyâ when you have, more or less, said that nerdy incelbros should be given sex slaves by the government.
Before you protest about me putting words in my mouth, let me draw your attention to this remarkable post: https://scottaaronson.blog/?p=3766. Letâs remember the context here. This was shortly after a deranged psycopath ran over a dozen people in Toronto with his truck. Rather than condemning this terrorist, you instead posted a link to an article by Robin Hanson, a despicable alt-right libertaribro, who, more or less, said that to placate such violent young men, the government should enshrine a âright to sex,â which, letâs be real, means to give them a sex slave. Rather than condemning this disgusting argument, you actually entertained it! Rather than telling nerdy guys to stop being stinky and awkward and learn some social skills, you reinforced their prejudices and more or less told them that they have a right to sex, while only giving the minimal token support for womenâs autonomy.
Letâs be clear. Nobody has a âright to sex.â Nerdbro incel culture is founded on the idea that they have a right to womenâs bodies. By entertaining this disgusting idea, you are entirely undermining your claim to care about reproductive rights. Because why should I believe ypu when you say you support a womanâs right to choose,a womanâs right to control her own body, when you think stinky, awkward, nerdy young guys with no social skills should have a girlfriend mailed to them by the state?
Itâs clear that, in fact, you donât care about womenâs bodily autonomy Scott. Because there is a direct and inescapable logical contradiction, between bodily autonomy, and thinking youâre âowedâ sex, somehow, just for existing and being a male. To whatever degree you do support abortion, itâs not because you care about women, but because you care about âthe enlightenment,â which is just a dogwhistle for white supremacist capitalism. You said it yourself just now đ€Ł
Comment #51 May 4th, 2022 at 12:38 am
[…] err, consistently generous and thought-provoking contributions. Let’s take, for example, this comment on last week’s admittedly rather silly post, from an anonymous individual who calls […]
Comment #52 May 4th, 2022 at 12:57 am
Feminist Bitch #50: My main response to you is in my new post, but I should briefly add: you’ve lied and slandered me here, and anyone who doubts that simply needs to read the old post of mine that you yourself linked to. They’ll see that, not only did I not defend the absurd and risible concept of a “right to sex,” as you falsely claimed, but I condemned it the most emphatic language I possess. (The very title of the post, “Zeroth Commandment,” referred to my attempt to codify my condemnation, to enshrine female bodily autonomy as an ur-principle.)
What I have advocated, it’s true, is compassion and understanding (as opposed to the sneering contempt that drips from your comments above) for the many nerdy heterosexual males who suffer from crippling insecurity around dating. I’ve tried again and again to explain that, while many of these guys might’ve been considered good or even great catches for most of human history, they tend to struggle in a modern, unstructured dating environment where a genuine inability to read social cues is easily misconstrued as creepiness or harassment. And the more scrupulously moral they are, the worse that problem is. And so, hoping to make a positive difference for all concerned, I spent months of my life hashing out the complicated questions around nerdiness and dating with straight women, male feminists, gays and lesbians and trans folks. And I actually think we all came to understand each other a lot better—at least, those of us who showed up to share and learn rather than tweet and condemn. I plead guilty to all of this.
Comment #53 May 4th, 2022 at 6:57 am
Don’t feed the trolls, Scott. Judging from their absurdly over-the-top writing and mischaracterization of others’ position, “Feminist Bitch” could well be some extreme-right (and/or Russian) provocateur who’s just out to sow division.
Comment #54 May 4th, 2022 at 8:09 am
feminist bitch
“by my reckoning, you have made THREE TIMES the number of posts about the ârightsâ of nerdy incels and how sad it is that they canât get sex, versus posts supporting womenâs bodily autonomy and basic reproductive freedoms.
But Scott is a nerd, so, of course he’ll be posting *more* about what he knows best, i.e. the condition of being a nerd, rather than posting about what he knows less: what it’s like to be a woman, or the downsides of being a super famous movie star, or what it’s like to be a quadriplegic, or what its like to be living with stage 4 cancer, or what it’s like to be a Taliban in the 21st century.
Just like, as a woman, you’re here posting about women’s rights rather than posting about nerds’ rights on some feminist blog.
Tolerance goes both ways, you can’t just bully everyone on the planet to realize that your point of view is the most important one while simultaneously putting everyone else down for being different than you.
Comment #55 May 4th, 2022 at 8:36 am
Craig Gidney #42:
Oh right, space is indeed overkill. I made that mistake again to assume that spacelike separation would be the only way to get no-signaling (independence) between two system. (Last time I started with “timelike q-correlations could be problematic,” then backed-down to “there is one difference between timelike correlations and spacelike correlations” on realizing my mistake, and the next time I brought up the topic, I wrote instead “only those correlations between observables whose (anti-)commutator nearly vanishes”.)
On the other hand, some MWI proponents seemed to be more comfortable with suggesting that some connection between space-time and a preferred basis should be assumed, than to be specific that they mean “the tensor product” when they talk of “factor states”.
Your strategy to analyse quantum puzzles and paradoxes by translating them into the language of (circuit based) quantum computers indeed often works extremely well. But for many-worlds, one of the motivations is that the strategy of Copenhagen to base the ontology on classical measurement devices is dubious, as it the strategy of De-Broglie Bohm to base the ontology on partical positions. But for a quantum computer, it is absolutely correct to base the ontology on the individual qubits and their no-signaling independence.
In fact, I now had an idea how to make De-Broglie Bohm work in the quantum computer context, using qubits and their classical states as the ontology. (The idea is similar to Why the quantum equilibrium hypothesis? From Bohmian mechanics to a many-worlds theory, so maybe claiming that it still De-Broglie Bohm will be hard to defend.)
Comment #56 May 4th, 2022 at 11:32 am
(Resubmitting this comment after it vanished during moderation yesterday; assuming that was an oversight confusion with my subsequent comment that went through, since I can’t see how this is an outrageous or off-topic comment. Â I’m actually asking a real question here, and would appreciate answers)
Hereâs the thing about objections to MWI that I donât understand. Yes, I understand that once branches fully decohere, we canât (if QM is ultimately truly linear) observe other branches of the wave function, and then all the metaphysical questions about falsifiability and whether it matters become valid concerns.
But we CAN observe things that look like âshort branchesâ of the wave function, that recohere just before they fully decohere. Thatâs what âWignerâs friendâ is all about: we can kind of see from outside how an observer enters the superposition through entanglement, and then how as entanglement diffuses into the rest of the universe, larger and larger regions enter that superposition. To me, all MWI says is that this process doesnât somehow end once decoherence has progressed past the complexity level where we canât feasibly do the work to recohere it. But that is not some kind of magic boundary, and we can see the branches all the way up until it just becomes infeasible to recohere them.
Decoherence as a process, as far as I can tell, is not really debateable anymore- it is a clear and mathematically sound consequence of how increasing entanglement complexity propagates in a complex world. By itself, it seems that decoherence as a process that exists, MUST produce something that looks like wave function branching. One may argue about whether the branching continues forever, but the miniature, abortive branches that seem to form in something like Wignerâs friend (or, as far as current experimental procedure has progressed, on a smaller scale with inanimate objects where we can feasibly force the branches to interfere and thus recohere) canât really be dismissed. Decoherence MUST produce something that looks locally like collapse; thatâs just the way the math works.
So, the question I have always wanted to put to those who believe in objective collapse theories is this: if at least the beginnings of abortive branches can be observed in smaller scale experiments, through incomplete decoherence, then when the additional objective collapse postulate is added for rare, true collapse events, we really seem to have TWO kinds of collapse going on, even if one of them is temporary: decoherent collapse, which on a small scale must be acknowledged, even if the ultimate irreversible decoherence is never reached, and the added objective collapse. So, do objective collapse believers acknowledge that decoherence produces something that looks locally like collapse, or not? Even if they donât believe in the permanent and unobservable MWI branches, it seems to me they must acknowledge at least the beginnings of that process, before complexity removes our ability to recohere.
There are really two questions: whether decoherence as a natural part of how Schrodingerâs equation evolves occurs and looks locally like âtemporaryâ collapse, and then the further question of whether that process continues past the point where we can feasibly create interference between the ârootsâ of the branches, and produces the unobservable larger branches that the MWI argues for. Why canât everyone at least agree that decoherence produces something that looks like local, abortive collapse? And if so, why is it not obvious that the boundary between that and unobservable branching is simply a matter of complexity theory: is recoherence feasible or not?
Comment #57 May 4th, 2022 at 7:43 pm
@David Griswold:
As far as I know, objective collapse theories usually don’t eliminate unobservable branching anyway; there’s still a regime (in terms of length, time, and mass scales) where objective collapse occurs too slowly to stop the branches from forming, but the behavior is classical anyway, so the other branches must be unobservable. So, there really isn’t any connection between observability of superpositions and their tending to undergo objective collapse.
Comment #58 May 5th, 2022 at 12:14 am
@ fred #39:
I think that 1 is obviously wrong (in my opinion)
I think that my position is that something sort of close to 2, but with some modifications/additions is a satisfying answer:
I think that the ‘swap’ of Shion into Fred, or vice versa, is in fact a no-op, in the same way that talking about magically swapping two individual electrons somewhere is a no-op, because There’s nothing ‘there’ to swap. I.e. if the swap will produce an identical configuration, it’s a no-op. As Eliezer Yudkowsky says in his series of posts that have been referred to above, there’s no such thing as “this electron here, that electron there”, only “an electron here, an electron there”. Likewise, we have “a Shion here, a Fred there.” I don’t think there really is something like a ‘universal conscousness’, in the sense that consciousness is an emergent property that applies to certain types of patterns in the quantum field (for example human brains). Other parts of the universe’s quantum field, for example the vast majority of it that consists of mostly empty space, are pretty clearly not conscious.
How this relates to the ‘why am I me’ thing, this model will predict that the part of the universe that can be considered Shion will have the relevant experience of “I am Shion” and the part that we call Fred will have the relevant experience of “I am Fred”. I’m trying really hard to do more than just restate my previous point here, but it seems like this model doesn’t really leave a dangling mystery about identity.
As for “Why are Shion and Fred conscious”, I actually think that this is in fact a very hard problem, that might take humanity a long while to find a satisfying answer for.
Comment #59 May 5th, 2022 at 2:30 am
Scott, as a far-left mathematician (how far-left? think German Antifa), I just wanted to say that:
1. While we probably disagree on most political matters, I completely support you when it comes to your analysis of the pain that so many lonely and harmless nerds experience.
2. Wokeism is a mostly an American phenomenon, and although it has hijacked a large part of the American left, I believe that wokeism is more of a cultural and psychological phenomenon than a genuine political movement. It often involves a grotesque display of narcissism in the disguise of identity politics, which I guess is a natural outcome of social media rewiring people’s brains and social norms. I guess that the point here is that in order to explain the behavior of Feminist Bitch and other internet wokes, you should refer to Freud rather than Marx, and you should refer to Netflix rather than to a coherent statement a la the Communist Manifesto. The bottom line of this is that wokeism has very little to do with leftism. While in North America the two terms are commonly perceived as overlapping, this is not the case in Europe and other parts of the world. So please, please, stop mentioning ‘woke’ and ‘left’ next to each other as if it’s the same entity.
Comment #60 May 5th, 2022 at 10:48 am
shion #58
“How this relates to the âwhy am I meâ thing, this model will predict that the part of the universe that can be considered Shion will have the relevant experience of âI am Shionâ and the part that we call Fred will have the relevant experience of âI am Fredâ. Iâm trying really hard to do more than just restate my previous point here, but it seems like this model doesnât really leave a dangling mystery about identity.”
In his book “my view of the world”, Schrodinger brings this up in a different way, which I’m trying to restate here (with some modifications):
Imagine two individuals, A and B.
In situation 1), A is sitting outside on a chair in the sun, looking at a tree, while B is sitting in the basement, in the dark.
In situation 2), B is the one sitting outside on a chair in the sun, looking at the tree, while A is sitting in the dark in the basement.
From the point of view of the physical universe, the two situations are very similar, i.e. the difference in the material/physical description world between those two situations is a function of the difference between the brains of A and B, which can be made arbitrarily small in theory.
But there exists another point of view, namely A’s subjective experience, where those situations are totally different, totally asymmetric, no matter how similar A and B are: In situation 1) A is seeing a tree, in situation 2), A is seeing darkness (and this isn’t depending on B’s point of view).
So, when you say “a part of the universe will have A’s experience”, we just don’t have any physical description/law encapsulating this, e.g. from the point of view of physics, the wave functions of the universe (or the quantum fields) in those two situations can be made arbitrarily indistinguishable (it’s like an extension of swapping two electrons, but, unlike brains, electrons don’t have subjective experience… or maybe they do?).
According to Schrodinger, this break of symmetry is the hard problem of consciousness: a description of the universe in terms of matter (particles) will never capture this asymmetry, it will never capture subjective experience.
He was apparently into Vedanta Hindu philosophy, so his answer to this problem was a sort of universal consciousness.
Comment #61 May 5th, 2022 at 10:01 pm
https://arxiv.org/abs/2205.01328 shows BQP contains P^#P.
Comment #62 May 5th, 2022 at 10:07 pm
#P in FBQP #61: Iâd follow the protocol thatâs worked in every single one of the dozens of previous such cases. Namely, just calm down and wait for the bug to be found. In this case, the fact that the author canât decide whether he wants to explicitly claim P#P=BQP or not claim it is far from reassuring.
Comment #63 May 6th, 2022 at 8:46 am
Far-Left mathematician
“While in North America the two terms are commonly perceived as overlapping, this is not the case in Europe and other parts of the world.”
Wokeness has already been a big characteristic of the french far-left for many years now, with American woke-ism being of course a huge inspiration (e.g. the political rise of people like AOC has been noticed immediately, she’s a role model now).
It’s not surprising since today’s world is so connected. The George Floyd movement (with effects like tearing down statues and defunding the police) spread to the entire world in no time.
Comment #64 May 6th, 2022 at 9:47 am
The author has many publications in the area.. https://arxiv.org/search/quant-ph?searchtype=author&query=Huh%2C+J and https://scholar.google.com/citations?user=P9AW-UYAAAAJ&hl=en&oi=ao.
Comment #65 May 6th, 2022 at 10:04 am
Scott #62
“Namely, just calm down and wait for the bug to be found.”
Note that if everyone adopts this procedure, no bug will ever be found!
So it’s best to keep this procedure to yourself đ
Comment #66 May 8th, 2022 at 12:33 pm
Scott #12
I’m interested what you think of this thought experiment where I imagine running a quantum counting algorithm on a simulation of a conscious human, to find out for how many inputs it produces an output with a given property.
https://www.lesswrong.com/posts/ZtWHfjjR8KvaGsLsB/demonstrating-mwi-by-interfering-human-simulations
I always find you’re blog fascinating so very interested to know what you think. As you’ll see it’s a modification of David Deutsch’s argument for MWI from Shore’s algorithm.
Comment #67 May 9th, 2022 at 3:42 pm
To me, the most noteworthy thing about the argument is that the “wavefunction collapse” here is caused by the consciousnesses observing their respective inputs. The simulation may include simulated decoherence, but this would just be an implementation detail; so we should not attribute the collapse to it.
Comment #68 May 11th, 2022 at 6:45 pm
The world was going along very nicely for billions of years before people invented symbols to try to represent what was happening. The world is not so much a mathematical system, as a system where mathematical symbols are useful to represent relationships, categories (like mass, position, or light wavelength), and things described as ânumbersâ, that seem to exist in the world. To claim that the world IS a mathematical system is a step too far.
Seemingly, most of the mathematical models of physics donât take seriously a world where Putinâs troops are killing people and destroying cities. According to these models, what Putin is doing is just another neutral outcome that occurs, an outcome that is not essentially different to any other outcome that occurs, because all outcomes are just the system evolving in a normal way.
Seemingly, only Chris Fuchsâ model can come close to taking seriously a world where Putinâs troops are killing people and destroying cities. This is because Fuchs doesnât attempt to describe a mind-independent reality. In his arXiv paper âNotwithstanding Bohr, the Reasons for QBismâ, Fuchs quotes physicist John Wheeler, who said:
âA lovely way to put it â âIs the big bang here?â I can imagine that we will someday have to answer your question with a âyes.â . . . Each elementary quantum phenomenon is an elementary act of âfact creation.â That is incontestable. But is that the only mechanism needed to create all that is? Is what took place at the big bang the consequence of billions upon billions of these elementary processes, these elementary âacts of observer-participancy,â these quantum phenomena? Have we had the mechanism of creation before our eyes all this time without recognizing the truth? That is the larger question implicit in your comment. Of all the deep questions of our time, I do not know one that is deeper, more exciting, more clearly pregnant with a great advance in our understanding.â
Comment #69 May 12th, 2022 at 3:31 pm
Scott: Did any verdict emerge on the paper https://arxiv.org/abs/2205.01328? People online are thinking you might know the truth. https://cstheory.stackexchange.com/questions/51426/new-quantum-algorithm-for-approximating-permanent.
It can’t be true correct? Is it really possible? Did any flaws show up?
Comment #70 May 12th, 2022 at 5:58 pm
I think life as a physical system completes a triangle of weird puzzles, the other two being the very large scale and the very small scale behaviors of matter. We have strong theories for each of these behaviors, though it is hard to imagine that any of them is a complete story. For life the obvious question is: why does matter arrange itself in this out of equilibrium self reproducing very large (relative to elementary particles) structure? Is that related to gravity, beyond the obvious fact that gravity creates chemical complexity necessary for life? Is it related to QM, beyond the obvious fact that life relies on QM in its underlying molecular interactions? Or is life just another intriguing fact about our universe?
Comment #71 May 13th, 2022 at 11:22 am
#P in FBQP #69: To say it one more time, the paper itself uses weasel language when it gets to the crucial point:
The existence of such a multiplicative error quantum algorithm for matrix permanent would lead to the surprising conclusion that BQP would equal P#P because we know BQP â P#P [53, 54]. Here, we do not definitively claim that BQP = P#P; further computational complexity analysis is required with our quantum algorithm for the conclusion. The bottom line is that the constant factor estimation of the real parts of the overlap integrals will lead us to the constant factor estimation of matrix permanents. We note here that the constant factor algorithm assumes the ideal quantum device, and only the statistical error was considered in the analysis.
If the author himself doesn’t entirely trust his proof, then can you blame me for not trusting it either? đ
In slightly more detail, the part I don’t believe is that the multiplicative error estimation will still work if the amplitude is exponentially small. And I don’t see any hint of an adequate argument that it still works in that case, and the author himself seems to admit that he lacks an adequate argument. But this is also the only part that matters if you want to do #P.
Comment #72 May 13th, 2022 at 8:30 pm
Usually things like this show up on reddit. But nothing happened this time. Brilliant insight on “the multiplicative error estimation will still work if the amplitude is exponentially small”. Also the author is a Chemistry Prof.
Comment #73 May 17th, 2022 at 1:26 am
Dear Scott,
Responding here since comments on the Roe v. Wade post were closed.
(Please donât approve this comment! If you have a response, I would prefer by email.)
Totally orthogonal to the issue of abortion, I would prefer a government which allows a diversity of local regulations, where states or even cities each form little experiments in different policies to improve the world. ⊠but with your commitment to people who have been wronged by society, you reminded me of the importance of compassion.
So, in appreciation for your blogging, your integrity, and in respect for the quality of your thought on these issues, I have donated $27.18 (($10)(exp(1)) + transaction fees ($28.26) to Fund Texas Choice in your honor.
Thank you for keeping your voice strong. Please donât let the trolls ever silence you.