I am new to this forum even though I have read a lot on the subject of boltzmann's brains and have come to a terrifying conclusion, the classic argument which for example Sean Carroll used, "The cognitive instability" (https://arxiv.org/abs/1702.00850), is not enough to dispel the doubt, his argument states that if thanks to scientific observations I come to the conclusion that at 99% I should be a Boltzmann brain, then follows that my perceptions are more likely to be random and therefore inconsistent, one falls into a contradiction where if it is true that I am a Boltzmann brain then it does not sense to think so because all the observations that led me to think of it are problably wrong because they are random.

This makes sense for heaven's sake, but there are several problems. The first is that one could argue that a Boltzmann brain could not even survive, it would die immediately from lack of oxygen, blood and nutrients. The second is that it would not be able to form in the sense that, being the Boltzmann brains slowly form (same link above Sean Carroll), the "meat" of the brain would freeze or should decaying for lack of support vital.

Now, a digression, I did not understand how exactly boltzmann's brains are formed, on Wikipedia, they cite 2: by quantum fluctuation and via nucleation, Sean Carroll cites 3: Vacuum fluctuations, Measurement-induced fluctuations, Boltzmann fluctuations. Now there is a problem, I don't understand too much about physics, if please someone can explain to me at least those from Wikipedia, from what I "understand" is that in quantum fluctuations, the object appears already fully formed, while for those on nucleation, are slowly formed with the accidental union of atoms, although I'm not sure because Wikipedia mentions strange concepts such as de Sitter's cosmological horizon. A question that follows is: particles and atoms in a vacuum can really unite to form macroscopic objects such as a brain, a stone, a computer, etc.. without energy (like in endoergonic reactions), so it seems strange to me that atoms in a vacuum can be connected at random. It would seem even more strange to me if atoms such as carbon and nitrogen for example, but any pair of atoms with different atomic numbers, can join together to form a heavier element, for example: hydrogen + hydrogen = helium, as occurs in nuclear fusion. It seems strange to me that it can happen in a vacuum without great energy.

When I realized that on its own it is impossible (I think) for a Boltzmann brain to survive, an idea came to my mind. Being that eternity or nearly so, (if the proton does not decay, it is estimated that before the "heat death" will pass 10^10^120, source: https: //en.m.wikipedia.org/wiki/Future_of_an_expanding_universe) it is very long as time, also form other objects, such as computers, and being that in a possibly infinite period, the most likely events will occur more frequently, computers will be formed , most will have a meaningless code, but there will be a small part of them that will have a meaningful code and that another fraction of this, will want to form human brains (because an entire human needs more nourishment) and simulating an external world in their mind. That's a tiny part of computer, but compared to the only one I'm supposed to be living on Earth, there are a lot more that are going to be living in the distant future, how likely are I to be in the only life on earth?

Now, there are several obstacles that these computers have to face in order to do such a thing, I will divide them into 5 themes:


particles to create a human mind

Computing power needed

Computer knowledge

5)Errors in the simulation

a computer could take energy from Hawking radiation from a black hole, a neutron star or a star of iron (iron beads that are formed in the distant future, see link: https://en.m.wikipedia.org/ wiki / Future_of_an_expanding_universesince blacks holes will rule for more time, so I assume that they take energy from there, maybe a black hole rather small that releases more energy.

a computer could travel for a long time and a lot of space, collecting atoms and then combine them and create a brain, would not be as impossible because there is no friction being empty space, so it would take a small initial boost and a small to dodge blacks holes or objects dangerous

3)since the fluctuations that form smaller objects are more probable, a traditional computer is to be excluded given the size of our universe, we would need a quantum computer that would require 615 qubits to simulate our universe(https://medium.com/data-driven-fiction/how-many-bits-to-simulate-the-universe-f60489f5ddce), and being the temperature already close to absolute zero, you don't even need a cooling system, so the size is very modest

4)here (fortunately) you have the greatest problems, being a computer created by chance, the more coherent information it has, the less likely it is, so it could form with "only" the knowledge necessary to form a human brain and keep it alive for a moment (in this so if really the most probable scenario is this and I am clearly not dead then we can say a priori that for some reason these computers will not form, without testing it with experiments), even if an AI could also be created that understands and is able to deduce through attempts and study of the remaining matter, how to create a human brain (a small percentage will do so, albeit in quantities larger than the one human life) and how to keep it alive and make them live a life in a simulated world. Now, I don't know which of the two will happen more ooften

I think errors occur but they will still be hidden from my computer I have thought of different solutions, but none of them seem convincing to me, and you have to understand one thing, you have to understand what is the most likely situation that will happen in which a brain that lives, at least for a short time. I greatly fear, I'm terrified because I think that the situation most likely expected that I survive long in a simulation because if a computer is able to form a brain and keep me alive, then it means that it is already very intelligent, and therefore will deceive completely, because if the computer was more "stupid", it would not even survive and create me and make me a moment.

however the solutions I found are: 1) due to quantum tunneling, all the baryon matter will become iron and furthermore there will not be enough energetic photons to form matter through the pair production, in this way at least for one way of formation by mixing particles a chance, it won't be possible, I don't know about quantum fluctuations. 2) particles do not join in vacuum (reason explained better above). 3) it is more likely that a computer will be formed that only knows how to do what is necessary to create a human brain and make it survive for a short time than a computer that understands the universe and also manages to deduce more or less what happened in the distant past . 4) if the goal of the computer is to simulate, I don't think it wants me to find myself in a horrible life that will surely lower my life span and therefore simulate for less time. It could at least, I don't know, create a brain less predisposed to anxiety and a society that feeds it less. 5) the protons decane and therefore the universe goes into thermal death first, so even if boltzmann computers are created, they will not have the energy to maintain a human brain. 6) maybe there's something I'm forgetting that prevents all this.

I know I wrote a long text but I really afraid, please if you can answer me at least something, thanks

  • Assuming the code in these computers formed completely at random, why would you think computers programmed to create brains receiving signals from a consistent, lawlike simulated world would be more common than computers programmed to create brains receiving more chaotic or unpredictable signals, given that there are far more possible combinations of inputs that would fit in the latter category than the former?
    – Hypnosifl
    Commented May 30, 2022 at 17:36
  • Thanks for the answer, you are right, if we assume that it is more likely a code statically with the information already present rather than a code that allows the computer later acquire the information, the latter in my opinion, it may be easier, although I sincerely hope no
    – Zeruel017
    Commented May 30, 2022 at 17:52
  • 3
    Forgive me for asking, but what you are afraid of? I'm confused as to what bad thing you think might happen here. Commented May 30, 2022 at 18:09
  • 3
    The essay that says you can simulate the universe with 615 qubits is wrong. Extremely wrong. You shouldn't use it as a source.
    – benrg
    Commented May 30, 2022 at 18:12
  • 2
    Re the 615 qubit thing: they calculate the number of atoms and photons and Planck-volume-sized regions in the visible universe. Then they add those three numbers, which makes no sense. The sum is almost exactly equal to the number of Planck-sized regions since it's the largest, so the other two calculations were pointless. Then they take the base-2 log. The result is basically the number of bits you'd need to identify one Planck-sized volume, not to represent the state of the universe. The author knows nothing about quantum computing, they just read some Wikipedia articles.
    – benrg
    Commented May 31, 2022 at 1:09

3 Answers 3


A Boltzmann brain doesn't need to survive for long. From Wikipedia:

The Boltzmann brain thought experiment suggests that it might be more likely for a single brain to spontaneously and briefly form in a void (complete with a memory of having existed in our universe) rather than for the entire universe to come about in the manner cosmologists think it actually did.

Theoretically, over a sufficiently long time, random fluctuations could cause particles to spontaneously form literally any structure of any degree of complexity, including a functioning human brain. In this thought experiment, a Boltzmann brain is a fully formed brain, complete with memories of a full human life, that arises from a state of thermodynamic equilibrium. The scenario involves only a single brain and not an entire person, because a single organ is less complex than an entire body and therefore more likely to spontaneously occur. Also, an entire body would include sense organs which would immediately disprove the brain's misperception that it is a normal brain in a normal situation. Like any brain in such circumstances (the hostile vacuum of space with no blood supply or body), it would almost immediately stop functioning and begin to deteriorate.1

The idea is named after the Austrian physicist Ludwig Boltzmann (1844–1906), who, in 1896, published a theory that tried to account for the fact that we find ourselves in a universe that is not as chaotic as the budding field of thermodynamics seemed to predict. He offered several explanations, one of them being that the universe, even after it had progressed to its most likely spread-out and featureless state (or at thermal equilibrium), would spontaneously fluctuate to a more ordered (or low-entropy) state such as the universe in which we find ourselves. Boltzmann brains were first proposed as a reductio ad absurdum response to this explanation by Boltzmann for the low-entropy state of our universe

In other words,

  1. Boltzmann brains are supposed to be an absurd thought experiment to counter arguments that our universe is just a giant thermodynamic fluctuation.

  2. Boltzmann brains wouldn't live long, as you have correctly said. At the same time, none of us know whether we will die in the next second, we know only the now. So the fact they would basically immediately die doesn't help. All memories fake, and time to think a single thought is all it takes.

The consensus amongst cosmologists is that some yet to be revealed error is hinted at by the surprising calculation that Boltzmann brains should vastly outnumber normal human brains.[6] Sean Carroll states "We're not arguing that Boltzmann Brains exist—we're trying to avoid them."

Brian Greene states: "I am confident that I am not a Boltzmann brain. However, we want our theories to similarly concur that we are not Boltzmann brains, but so far it has proved surprisingly difficult for them to do so."

So, you should expect that eventually, the theories we believe in exclude Boltzmann brains.

  • okay, I'll explain better what I mean, what prevents particles from joining together in a vacuum to form, for example, a computer, which persists over time unlike a biological brain, and what prevents that computer from forming a brain. You could tell me that it is very, very, very difficult and unlikely that such a thing will happen, and you are right, but in an infinite or near infinite time, it will surely happen, and a lot of times.
    – Zeruel017
    Commented May 31, 2022 at 15:54
  • @Zeruel017 Yes but still vastly outnumbered by the small, minimal brains. The whole point is to dispute that our (big) universe is a random thermodynamic fluctuation. Because Boltzmann brains would vastly outnumber oberserver that actually live somewhere. Simple brains would outnumber more complex (like your computer). Brains with abnormal experiences would outnumber normal experiences. It doesn't matter whether a computer is possible to exist (that it is possible is exactly the idea of Boltzmann brains), you are still more likely to just be a shortlived brain that just popped into existence.
    – kutschkem
    Commented Jun 1, 2022 at 10:01
  • 1
    @Zeruel017 Again, the point is you want a theory that excludes Boltzmann brains but don't have one so far. You shouldn't doubt your existence and your very reality just because we have faulty physical theories that we don't quite know yet how to fix.
    – kutschkem
    Commented Jun 1, 2022 at 10:03
  • @D.Halsey Evolution has nothing to do with the question, nor the answer.
    – kutschkem
    Commented Jun 1, 2022 at 10:05
  • @kutschkem then, starting from a point. If a brain is very simple, I don't think it can be considered me, I think "I" is a configuration of cells with the same DNA, so if it is simpler or more complicated than this (for example with a simpler DNA and short) it's not me. 2) a brain would not even have time to form, it would first die from frostbite or the tissues would die while they are formed. 3) how do we know that the theories that predict boltzmann brain are falty and not simply that they are correct and therefore we probably are?
    – Zeruel017
    Commented Jun 1, 2022 at 12:18

Boltzmann brains are still talked about by prestigious physicists and philosophers so I don’t know why people are dismissing you. Leonard Susskind, Sean Carroll, and David Albert and many others continue to discuss them at very deep levels.

The main reason they discuss them is not philosophy of mind (or only barely). I’m following their lead and focusing on their cosmological power.

The two above physicists and philosopher discount any cosmology which posits more BB’s than “natural observers”, i.e. ones which were not formed via fluctuation.

Many current cosmological models predict infinitely more BB’s than any other type of observer. Those models must be wrong according to the above scientists and philosophers.

The models’ failures are because if there are infinitely more observers of this type why aren’t we one? The odds of being a natural observer are 0 in any a typical heat death single universe model.

That lead Susskind, Carroll, and Albert to say those models must be wrong. I am not aware of what exact changes Carroll and Albert favor, but Susskind’s I am aware of.

He says a special case of the multiverse of eternal inflation offers a model which avoids this problem. Regular eternal inflationary models fail. Susskind has in mind an eternal inflationary multiverse where the probability of spawning a new universe overwhelms the number of bubble universes in heat death. Each non-degenerate bubble universe has a chance at fluctuation a daughter universe, an idea from Sydney Coleman and Delucca. If the probabilities are played with, you still get infinite BB’s, but infinite natural observers as well. Perhaps infinitely more with the right probability measure.

This is where I think the solution to you worry is to be found. There are potentially many bubble universes dominated by BB’s or your digital version. But there are so many more observers like us.

How do we know we are “observers like us” or naturally evolved? Say BB’s don’t just fluctuate into existence knowing they are BB’s. Maybe the fluctuation is so great they fluctuate with memories like ours. They think they are natural but aren’t. How can anyone know they didn’t just pop into existence a second ago with all their memories? Can we solve that epistemic problem and how does it connect with Susskind’s model.

I don’t know that Susskind discusses it but I think he may implicitly believes what Carroll and Albert state. That if we begin to worry about that kind of BB, we can’t make the usual scientific progress. There really is no way to absolutely prove we aren’t such beings. Science is not interested in radical doubt at every level. Something has to hook into reality. And we carefully follow those leads and assumptions scientifically to better and better theories.

  • I am of the opinion that it is a major test for cosmological theories, exactly because it has relevance to Brain in a Vat arguments. Really no one would care if randomly a carpet was created, the problem is that we may be BBs and it is very likely statisticaly. So theory of mind is necessarily involved
    – Nikos M.
    Commented Jun 11, 2022 at 16:12
  • @NikosM. People would care if a carpet fluctuated because those odds are so minuscule according to stat mech we’d begin to doubt them. Every additional particle in a fluctuation exponentially raises its unlikelihood. So unlikely we’d probably posit many other explanations before believing it did fluctuate. And if we did believe it fluctuated, we’d probably revise some theories.
    – J Kusin
    Commented Jun 11, 2022 at 16:21
  • Ok I agree, but a carpet fluctuation would not make us doubt our own existence.
    – Nikos M.
    Commented Jun 11, 2022 at 16:27
  • @NikosM. This goes back to your statistically likely comment. If BBs were statically likely I’d doubt myself as a human yes. But that statically likelihood depends on what models we entertain right?
    – J Kusin
    Commented Jun 11, 2022 at 16:34
  • Not if BBs are unlike natural human beings, regardless of how likely they are, right? We dont doubt ourselves
    – Nikos M.
    Commented Jun 11, 2022 at 16:35

I would like to give a different answer to this question.

A Boltzmann brain is a brain that has formed by chance (probability theory does not forbid that). Many things may be formed by mere chance, what is interesting in this scenario is that it is a brain, it experiences things (or at least it should).

Really it doesn't matter how it is formed, it is irrelevant to the questions posed by this thought experiment.

All that is important is that it is a brain that experiences itself as a normal human while it is not a normal human

Let me introduce you to another thought experiment called "Brain in a Vat" which is the hypothetical scenario of a human brain inside a jar in a laboratory where it is fed electric signals so it thinks it is actually a normal human experiencing real life.

What these thought experiments are about is the hypothetical existence of perfect simulations (regardless how the simulation comes about or how it is sustained).

There are conditions such and such that give the perfect impression of something (in this case full experience of a human being) while it is not so in reality.

Let us ask ourselves now, are perfect simulations even possible?

For example in the brain in a vat experiment, where are the interactions of the brain with the liquid in the jar and with the jar itself? Really they can't disappear, they must leave their trace (energy which cannot disappear physically is contained in these interactions).

Where are the interactions of the Boltzmann brain with its environment? Really they can't disappear, they must leave their trace too.

These environmental interactions is more than possible that they are so overwhelmingly intense and important that they override any false memories or other electric signals. Even if the interactions are not so intense, they still leave a non-zero residual trace.

So if they leave such trace, this means the trace of these factors is a distinguishing feature between real thing and simulated one, that is, the simulation is not perfect.

What would a perfect simulation consist of? Really it would have to eliminate these traces completely, so brain in a vat would really need to have a functioning human body experiencing real life and similarly for Boltzmann brain, so all factors are truly identical.

But if this is the case, it is not simulation any more, it is the real thing!

Aside from such purely physical arguments against perfect simulations there are also purely logico-philosophical arguments. Like the contradiction with the law of identity. A perfect simulation is identical to the real thing (as perfect) while at the same time it has to be different from it (as simulation and not the real thing).

So we come to the conclusion that perfect simulations are a contradiction. Either something is a (imperfect) simulation or is (perfectly) identical, but cannot be both at the same time.

People have pondered on perfect simulations even before modern science (eg using dreams in similar thought experiments). The hypothetical existence of perfect simulations poses the theoretical problem of not being able to know, in principle, what is really real. Well, dreams are also not perfect simulations, if you search the interwebs you will find many methods to recognize when you are dreaming.

In simple words: a brain to experience real life as a human, needs to be embodied in a functioning human body operating in real life. Period.

"I live in a lousy lab and have a vat for a body, for crying out loud!!" - Brain in a Vat

PS: There is a humorous but deep essay (The Story of a Brain) in the book The Mind's I that reflects on the problem along these lines. Also related Brain in a Vat or Body in a World? Brainbound versus Enactive Views of Experience. Check them out.

  • 1
    If Plato is right all we’re learning via “real” experience is remembering from somewhere such as world soul, then how I could detect I was just in a perfect dream when I didn’t doubt at all one of my teeth was broken last night? To be conditioned to look for some food to eat to test out? Any conceivable “test” method/trick can be interpreted as already exist outside our “real” world long ago within the world soul, and any “simulation” trace is in fact there but out of our current reach… Commented Jun 11, 2022 at 23:24
  • Positing unobserved entities (and unconnected to any real aspect, regardless how you define "real"), like Plato's Ideas can be dismissed easily and the onus of proof is on them, which they can never really carry. As for the dreaming example I answered that elsewhere
    – Nikos M.
    Commented Jun 12, 2022 at 7:45
  • We demonstrably live in a world where we can utter literal nonsense, lies and all in between. Thus any propositions articulated do not by default constitute possibility.
    – Nikos M.
    Commented Jun 12, 2022 at 21:58
  • I smell (logical) positivism here which was once very certain and popular, but seems now out of fashion especially after Carnap's death and Popper's attack... Also see a recent perhaps related post... Commented Jun 12, 2022 at 22:30
  • 1
    thanks, I really appreciate your answer, yes, my hypothesized scenario was just brain in a vat
    – Zeruel017
    Commented Jun 14, 2022 at 23:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .