It's also interesting to consider other Boltzmann "objects" as well.

In theory, just as it can be argued that it'd be more thermodynamically probable that a single brain would occur than an entire universe - what about a single quantum computer running a simulation of such a universe?

But why even necessitate the physical at all?

Could there simply be a Boltzmann simulation - a collection of abstract information and rules representing a multi-dimensional physical universe? In particular, if it contained techniques internally similar to procedural generation, a great deal of abstracted information could be represented by a much more compact collection of algorithms and rules.

Of course, some might even take the concept further, reducing from a Boltzmann simulation to simply a Boltzmann architect - a purely information based intelligence capable of creating and sustaining such a simulation without the need of any physicality at all, and spawning off child threads to extend that property of "consciousness."

Such a foundational basis for existence (in either the ontological spontaneous simulation or intelligent designer cases) might both be argued to be more likely to occur than the massive physical universe itself.

This is all really awesome from my perspective, but I kiiinda got lost about here:

>Of course, some might even take the concept further, reducing from a Boltzmann simulation to simply a Boltzmann architect - a purely information based intelligence capable of creating and sustaining such a simulation without the need of any physicality at all, and spawning off child threads to extend that property of "consciousness."

I guess the concept of 'pure information' or 'abstracted information' is where I start to get confused. You'd still need that instantiated on some physical medium. A bunch of hydrogen atoms (or whatever) would have to jostle into place to represent the needed information, right? Unless this is a variant of a Kalam cosmological argument that takes minds to be different from, and simpler than, physical things? Or is this non-physical 'stuff', and if so, does that non-physical stuff get subjected to the random jostling the same way as physical stuff?

>and sustaining such a simulation without the need of any physicality at all, and spawning off child threads to extend that property of "consciousness."

And where are we here? I'm not clear what physical medium the intelligence/simulation/pure information is on, or if there's no physical medium, I'm not clear on what way the intelligence/simulation/pure information is popping into existence or what it means to spawn child threads. Maybe for your purposes physicalism is entirely assumed away at this point?

Then there's the 'foundational basis for existence' which I think departs from the terms of the original thought experiment, which presupposes that a certain kind of universe already got there, and that Boltzmann Objects are already busy poofing into and out of existence within it.

Though I take your more general points that (1) there's a whooole taxonomy of Boltzmann "objects" that can be imagined and (2) among those objects, some of them will be relatively simple and make for interesting considerations about what's most likely to poof into existence.

The book Permutation City explores a concept like this. I really recommend it.

>Your decisions only matter in the case that you're an actual human

Why do you say that decisions matter if you're a human? At least, why would decisions matter more as a human than if you're a Boltzmann brain?

I'm not using "matter" in a very deep way here. As a human, your decisions matter to yourself in the sense that you're going to experience the consequences of them. If you choose to eat an inedible object, then you're going to have a bad time. But if a Boltzmann brain chooses to eat an inedible object, it does not matter because the Boltzmann brain can't execute that decision in reality and is going to stop existing in a moment.

in the sense that we could be "brains in boxes" like in the matrix, one could argue that this reality we experience could happen to a boltzmann brain as well. similar to a dream, in which the dream reality is a result of feeding all the senses with data from within.

An important distinction is that a boltzman brain only "works" for an instant before the harsh environment of space destroys it. There's simply no time for a whole dream matrix reality to play out.

Another Boltzmann brain can pop up, which contains a short term memory of being previous Boltzmann brain. Sensory inputs aren't constrained though. Most probable sequence of such brains should feel like noise all over your senses, I think.

Another Boltzmann brain can pop up with the result of whatever decision the previous brain could have made, the actual decision it made is irrelevant.

Why? The time passed for the Boltzmann Brain won’t necessarily be tied to its real frame of existence i think, it may simulate any timeframe imaginable, depending on its starting conditions.

In fact the simulation of any timeframe would be constrained by physical laws.

For example for a Boltzmann Brain with a similar chemical composition as ours the speed at which chemical reactions occur would limit the speed of experience.

For a Boltzmann Brain that simulates ours with a very different physical substrate the speed of light would still be a limit.

>I find this question really fun: "Am I a Boltzmann brain, or a real human?". If there exists both a human and a Boltzmann brain with your current mental state, then it doesn't make sense to try to answer this question with a single answer.

While probably not entirely in the spirit of your question, I would emphasize this part of the wiki article:

>Like any brain in such circumstances, it would almost immediately stop functioning and begin to deteriorate.

I think if you pause long enough to consider whether you're a Boltzmann brain, you've probably lasted longer than one and aren't one.

A Boltzmann brain can form with a false memory that it had paused to consider whether it is a Boltzmann brain.

In other words, take a real human considering whether they are a Boltzmann brain and deciding that they are not. Any single moment in that human's existence can be simulated by a Boltzmann brain existing momentarily, including the moment where the human concludes that they are a real human, and the moment where the human reflects on its conclusion that it is not a Boltzmann brain.

You're right. It can, and I see this has been discussed elsewhere in the thread. So it's certainly in play.

But each variant of a boltzman brain that accounts for the passage of time, passes that 'test' at the cost of being less probable overall. Most boltzman brains won't pass that test, but some will. And then you throw in the other tests people have mentioned: a feeling that your memories are somehow internally coherent and consistent. Most won't pass that test either, but some will. And 'some' could be so many that, if you were picking out of a hat, they are still more likely than that your experiences have really happened.

I guess the interesting question is whether you can collect together enough 'tests' that filter out BBs, that you can pare down the range of possible BBs that pass the tests, to the point that they become a less probable explanation than the actual spontaneous emergence of life in the universe.

That question only makes sense in reference to the model of the universe you choose and you estimate parameters. It's not a precise argument, it's a warning about theories that make life too improbable.

And of course probability doesn't matter for events that occur only once in the universe. This isn't a scientific theory.

I take the boltzman brain convo as one that takes BBs as forming within the universe we are familiar with, and thus grounded in probabilities based on physical interactions of the kind we know about that happen in our universe. A BB could spontaneously form out in space, and the rules of space, entropy, etc are things we are familiar with. We can make reasonable estimations of the probability of things that happen within our universe. A boltzman brain forming within our universe is one thing, and life forming is another.

And if we've thrown that core reference frame out the window we've stepped entirely outside of the rules that are the basic structure of the whole boltzman brain thought experiment.

A Boltzman half-brain with random fluctuations filling in the missing hemisphere's activations is much more thermodynamically likely than a full Boltzman brain. And a quarter of one even more so, etc.

That certainly explains why I'm so forgetful.

I think the proof that I am not a Boltzmann Brain is that my memories are chronological and all feature me as the self.

If it is infinitely more likely that I am a Boltzmann Brain than not, isn’t it also extremely more likely that I’d be one with any number of random memories rather than a group of memories that form a more or less coherent whole?

A dream is a totally incoherent result of random fluctuations, but it seems real and coherent until you wake up. A Boltzmann brain only exists for a very short period of time — on that small of a timescale, your existence is really just a single instantaneous snapshot of a state of mind, and there’s no time to recall a sequence of memories or think through logical propositions. Even though existence feels coherent, that feeling would be just a false memory.

>> there’s no time to recall

Why do you think that? A Boltzmann brain will immediately start to degrade but that does not mean there is zero time for thought.

If I were a Boltzmann brain, I would spend most if not all of my existence dwelling, contemplating, ruminating over the things that took me to my current point in space and time. Then I would quickly type something up and post it on a forum to confirm that I

How long could you think for if you were dumped in the vacuum of space?

I would disagree that a dream is 'totally' incoherent or truly random, and I don't think it's necessarily inevitable that existence would feel coherent, which I take to be GP's point. There's a wide swath of possible sets of memories I could be given. Some coherent and belonging to a self. Some that seem to but that, given sufficient relfection, would turn out not to be, as when you wake up from a dream, and then some that, even when lucidly regarded in the moment appear to be nonsensical on their face.

Of this range of possibilities, ones giving you a seemingly coherent and ordered timeline of experiences, or one that is disordered but close enough to order that it can be readily can be interpreted as ordered, would appear to be more improbable.

>on that small of a timescale, your existence is really just a single instantaneous snapshot of a state of mind, and there’s no time to recall a sequence of memories or think through logical propositions.

I think this is very true. Even being around long enough to pose the question to yourself gives you reasonably strong confidence that you've 'tested out' of being a boltzman brain. Though I think that consideration is being momentarily set aside for purposes of this variant of the thought experiment.

No, I don’t think so: https://news.ycombinator.com/item?id=22082864

Coherent memories are the convergent result.

> my memories are chronological and all feature me as the self

I’m not sure exactly what you mean by this, but the way I read it, I disagree with this as a description of human memory.

> I’m not sure exactly what you mean by this,

All my memories point to be me being not just human, but a particular human. In every memory I have, I am male, human, on Earth, live in the United States, live between the period of my birth and now. Speak English. Have the same number of siblings. etc. There's a more or less consistent story there.

And the story is chronological. You could argue that the timeline of events is arbitrary and my brain is just remembering them in the order of that makes sense. But I also have memories of remembering things, and in each one of those, I am remembering things that happened before the time I am remembering them.

The odds against a Boltzmann Brain having even two random memories that relate at all are incredibly small. The ratio of ordered:non-ordered Boltzmann Brains would be roughly on the order of actual Brain:ordered Boltzmann Brain. That's true even if we include a little wiggle room for the way the brain deals with memory.

The likeliest scenario is then that we are not Boltzmann Brains and any explanation of the universe that includes the probability that we are Boltzmann Brains is probably wrong in some way.

Preface: It took me years (decades?) to realize that what happens in each person's brain when they 'think' or 'remember' is very different.

I often have perfectly clear memories that don't fit into any cohesive story. I reject them under the assumption that they're a dream leaking into my waking subconscious, or neurons going crazy, or whatever. (If this were a sci-fi movie, they might be critical foreshadowing!)

I wouldn't describe memories as inherently chronological. They're an unordered heap, with various indexes. Time is one such index, but a weak and incomplete one. I often misremember the age (or relative age) of events. Location and mood are much stronger indexes.

I haven't had any memories of being a Martian, true, but given that there seems to be some automatic filter already (most dreams don't become memories), it doesn't sound unreasonable that my brain has a GC of sorts. Perhaps my filter threshold is set too low, but still high enough to reject the Martian memories.

I agree that Boltzmann Brains are not the most likely explanation, but reflecting on how my memory works doesn't seem to be a good argument against that scenario.

Yes, that is true. But those all will occur too (or so this theory says), you just happen to, right this instant, be a coherent one.

I don’t think that argument really works. The theory is built on statistics, the anthropic principle, and rejecting implausible coincidences, so you can’t reject the coincidence of the existence of the universe while accepting the (apparent) coincidence of random memories being somehow coherent.

But if you are a Boltzmann brain (Bb), it would be a very big coincidence if all of your memories were internally consistent. Since they presumably mostly are, you are likely not a Bb.

You can hide a lot behind "presumably mostly"!

My memories aren't entirely internally consistent. Memories fade over time, too, so even if they formed perfectly, I certainly misremember many events. The earliest ones are from when I was a kid, when I didn't understand the world, so I naturally place the least faith in their accuracy.

When I do have conflicting memories, I use logic and common sense to try to figure out which is correct, and hedge my bets against memories which might be false.

So if I were an intelligent brain formed with random memories, over time my natural tendency, in order to remain sane, would be to correct my memories to be internally consistent, as much as needed, and not examine the rest too closely.

That conclusion isn’t quite logically rigorous. You need to have a prior on the odds of being a Boltzmann brain in the first place. If that prior is extremely high, your conclusion won’t hold.

What if you are a recurring sequence of Boltzmann brains pooping into existence, lasting mere seconds but correctly temporally sequenced as to form a continuous consciousness?

There's no reason to expect that your decisions would affect the next Boltzmann brain in the sequence, so there's still no need to put value of the Boltzmann brains' decisions. I don't think it makes sense to put more weight on the specific series of Boltzmann brains where the decisions seem to be followed by a subsequent Boltzmann brain than on the much more likely series of Boltzmann brains that just resemble your mental state but don't have any seemingly-causal relationships between each other.

If you're allowed to put more weight on specific extremely-improbable sequences of Boltzmann brains, then you can propose all sorts of nonsense scenarios like "Imagine the series of Boltzmann brains where if one thinks of a purple elephant, then the next Boltzmann brain experiences the subsequent thought but also a lot of pain. Should you avoid thinking of purple elephants because of the fact this unordered sequence will happen eventually in an infinite universe?".

I think in order to really reason about this sort of thing, you need to start considering theories of the universe like UDASSA (https://news.ycombinator.com/item?id=20119513, http://fennetic.net/irc/finney.org/~hal/udassa/), which provides a framework for how we should expect to anticipate different events, and Boltzmann brains and especially specifically-selected sequences of Boltzmann brains have so little measure compared to other things including our human selves that they're not worth considering, even if their decisions actually did have a causal effect on some part of a world (which would also have minuscule measure compared to ours).

If you like these thought experiments, consider that I’m not actually responding to your comment, but I am actually mashing the keys randomly and it only appears to form a response by coincidence. We are not actually communicating with each other.

You are not mashing the keys, because nothing happens except that God wills it so. It just happens that God keeps making miracles that you have interpreted as the effects of your actions on the world.

This is the first time I've been downvoted for paraphrasing David Hume :)

https://en.wikipedia.org/wiki/David_Hume

Is the phrase "correctly temporally sequenced" even a necessary phrase? I mean, if you imagine there is a set of them where the "frames" are each at some random point in time or space, then what really changes? It's like if you had a simulation of the universe, the passing of time inside wouldn't have any particular relationship to the passing of time outside because the former is part of what's being simulated.

So if you had random sequenced ordering of Boltzmann Brains that add up to the experience of a single life experience, does that collection experience having a life?

If a random stone formation somewhere in the universe happens to look exactly like a statue of Abraham Lincoln, is it a statue of Abraham Lincoln?

Actually, more like "if a million random stone formations somewhere in the universe would be a statue of Abraham Lincoln if put together, are they such a statue?"

What if there is a meteor shower that happens to leave glowing craters on the moon, that spell out "John 3:16" in a dot matrix manner?

> What if you are a recurring sequence of Boltzmann brains pooping into existence, lasting mere seconds but correctly temporally sequenced as to form a continuous consciousness?

This is similar to the premise behind the novel Permutation City, except that it’s not just one brain, but an entire virtual world.

And while fictional evidence is still only fiction, the way that book ends strongly hints that they got something crucial wrong.

They weren't running "on dust", they were running in the simulation of some being that's particularly interested in universes that embed simulations of sapient beings. Such as their own, and then later the "experiment"...

I've thought about this.

And you could reason that because you observe your reality, then the only series of boltzmann brains that could comprise you are those that sequentially form consciousness.

You would never experience any other configuration.

Given some description of the real world that defines how physical states progress through time (like, say, quantum mechanics), aren't we all sequences of such states?

That's not necessary (and is even less likely): the "you" at any given instant would be a Boltzmann brain with the memory of existing at previous instants.

>> recurring sequence of Boltzmann brains pooping into existence

I think this finally explains current pop music to me.

Justin Beibers pooping into existence...

Almost all commenters below here seem to be stuck with how insufficient appearing like a continuous stream of consciousness is. It's a dhamma, impermenant, easily fakeable.

The shape of this argument reminds me of Pascal's wager.

Right, dissecting why that answer would be worthless is part of the fun of the thought experiment. It seems like it should be ... spectacularly interesting to somehow discover that your entire memory is a lie and the world you're in is radically different than you thought, but the fact you can't do anything with that knowledge completely neuters it. It implies that in the calculation my mind does about how much I care about a possibility, there's something like a term for "how much control (or even a vague casual connection) do I have over the world?" that everything or nearly everything else is multiplied by. Even if a deity told me I was a Boltzmann brain, I would still hang on to the minuscule chance or delusion that I was a human and still process decisions that way, because in some sense I fundamentally can't care about a situation where there's zero utility to be gained besides as a curiosity. It feels like this might be a useful intuition pump to think about concepts like motivation and learned helplessness. The fun in a thought experiment is how it gives you a new lens to view other concepts with.

I don't understand the point you're trying to make in this post.

Yeah, I kind of went off the rails there, didn't I?

My point was merely: How do you recgocnize a brain (and its very special cousisn the BB) at all? I wager it's just because you see a human before you. But if you really want to question things... then you're in for a very bleak time.

Just like one can't disprove solipsism, one can only brush it aside as pragmatically useless. Just because a theory is internally consistent or even accurate, does not mean it is a good all-around theory: simplicity and elegance counts too. And having to agree on who is the solipsist among all figments is inelegant and an impossibly complex task.

> According to the most extreme form of this view the only way by which one could be sure that a machine thinks is to be the machine and to feel oneself thinking. One could then describe these feelings to the world, but of course no one would be justified in taking any notice. Likewise according to this view the only way to know that a man thinks is to be that particular man. It is in fact the solipsist point of view. It may be the most logical view to hold but it makes communication of ideas difficult. A is liable to believe ‘A thinks but B does not’ whilst B believes ‘B thinks but A does not’. Instead of arguing continually over this point it is usual to have the polite convention that everyone thinks. -- Alan Turing

> Just like one can't disprove solipsism, one can only brush it aside as pragmatically useless.

Agree.

> Just because a theory is internally consistent or even accurate, does not mean it is a good all-around theory: simplicity and elegance counts too. And having to agree on who is the solipsist among all figments is inelegant and an impossibly complex task.

Indeed not... but the true measure of a theory/hypothesis is how hard is it to vary to explain new facts. (I think this idea is due to Peter Deutsch, originally.) If you can change your explanation by saying things like "Ah... well, Aphrodite wouldn't do that because she's pregnant in Spring"... then I'm sure we can agree that it's a badly or unfounded theory of the world.

This is actually an extremely high standard that very few pass.

Thanks. I'd never considered adaptiveness of a theory before and I like it. Perhaps adaptiveness at times emerges spontaneously from simplicity / regularization. But it is good to keep it in mind from the very start.

Don't we experience reality as if we were a Boltzmann brain anyways?

How many of us remember the first moment we became conscious?

Yes... and there's really no way to tell unless the simulation actually tells us.

An easy test for this is to ask anyone who's either on mind altering chemicals, or schizophrenic[0]. That makes it clear that you are really perceptually bound.

[0] I don't mean to say that these are even close to the same thing. I just want to pick two extreme examples.

It's very easy to tell you're not a Boltzmann brain. As long as your consciousness lasts more than a few seconds, you're not a Boltzmann brain. A Bolzmann brain appears in the vacuum of space and will quickly die due to the vacuum and not being in a body.

> Typically, a quantum Boltzmann brain would suddenly appear from the vacuum (alongside an equivalent amount of virtual antimatter), remain only long enough to have a single coherent thought or observation, and then disappear into the vacuum as suddenly as it appeared.

But at any given moment you don't know if your consciousness has lasted more than a few seconds or if you've just appeared with the false memories of having been conscious for a long time.

It begs the question, how few "true" memories and perceptions does consciousness require? I don't know 5^4, but I know I can know it; but couldn't my knowledge of all math be a false memory as well?

Could a consciousness comprise of just a single, false perception— that there are further memories and perceptions? What would it take for such a "mind" to manifest in our universe?

Probably almost nothing; I can't imagine the difference between this and an inert object, or even an instantaneous event. It has no real memory, so it doesn't need to exist for more than an instant.

(Uh, you can stop reading here if you want. Just a warning.)

This has some parallels to Bill Gosper's "HashLife", an implementation of Conway's Game of Life that effectively hashes every event at every scale and timescale. The simulation accelerates rapidly by skipping events whose cause and effect are already coded into those of larger events, and so on, up and up and up; every portion of the simulation ends up being a key for a value in the cosmic hash table. So who's to say what in the simulation is real, and what is symbolic of something more complex?