The fallacy of a Boltzman Brain is 'theoretically by sheer chance all the atoms could bounce off and stick to one another in such a way as to assemble a functioning human brain' is NOT an accurate depiction of human intelligence. Our human brain us supported by cellular functioning, metabolism, mitochondrial energy, and the generation of synaptic potentials using neurotransmitters. What you have is a DEAD human brain, which of course is not intelligent.

This is the only way an intelligence indistinguishable from something we would recognize as an AI construct that wouldn't have to be considered life (which anything that evolved would have to be) could possibly come about ~ & it would still have to be considered natural rather than artificial because it came about from natural processes (even if extremely unlikely random ones).

However. if by some 'Boltzman' process, the atoms all bounced off and stuck to one another in such a way as to represent a ROBOT, and that robot were somehow powered. and there was some useful program in it, then I could see your point. It would be an inanimate intelligence. Could that robot replicate and survive?

@pelinore has it. It is (infintessimally) possible, but would still be considered a 'natural' process. Although, much like 'natural' ingredients I do struggle somewhat with the definition of natural and artificial with these things. Considering we're 'natural' and are by our nature creating things from other 'natural' things it all seems a little...arbitrary. But hey ho ;)

@JustinThymetheSecond I don't see why it couldn't. We're talking about infintessimally snall probabilities here, turned into almost-certainties by the anticipated scale of the universe. Also, there's no reason it couldn't create a perfect human brain plus some apparatus to keep it alive. It's all mindbogglingly low probabilities, so as long as it's physically possible it's feasible if the universe is big enough :)

And if you read the article, a Boltzman Brain is expected to last no longer than a fleeting thought before the normal process of entropy causes it to lose fuctionality. Could such a 'Boltzman Robot' continue to exist without maintenance, power, and support? Or do we posit an entire working 'Boltzman Robot Factory'? Methinks we are talking 'Hitchhiker's Guide...' here.

@Ynneadwraith - I hear you, what's "natural" & what's not is perhaps a question for English language or Philosophy SE but it's (unfortunately) integral to this question (as it's been typed) & its answers, you gotta have a common frame of reference for what the words in the question "mean" to the asker to know what's actually being asked & to express an appropriate answer & have that answer properly understood :/

@JustinThymetheSecond Oh we're well into infinite probability whales coalescing in the upper atmosphere here. But I can't see any reason why it must be ephemeral. If any feasible structure may coalesce and it's possible for structures to be self-perpetuating then it's possible for a self-perpetuating structure to coalesce :)

^ As a rule of thumb I generally think of anything we (or any other animal) makes rather than excretes or otherwise produces through "natural" (there it is again!) or ordinary bodily functions as artificial ~ so that includes pointy sticks made by chimps or crows, even though you can get naturally pointy sticks, they're called thorns ~ everything else is natural ~ a wax honeycomb looks artificial at first sight but is made by instinct so I call that natural (& that's the point at which I feel my definition perhaps gets a little arbitrary).

It is that 'self-propagating' bit that was the conundrum to a Boltzman intelligent thing. Under the typical physical laws of entropy, there can be no such thing as a self-perpetuating entity. That was my original point in the OP. As soon as a complex structure is formed, it begins to 'entropy' and can no longer beget something more complex than itself. Intelligence should immediately become unintelligent, not more intelligent. By what method can it become MORE intelligent? How is it possible, under entropy, for intelligence to beget more intelligence?

@JustinThymetheSecond Interesting. I'd ascribe more weight to it if life so far hadn't proven it at least locally false throughout its existence, producing more and more complex and structured forms as time goes on. If it's actually known to have happened at least once then it's probably orders of magnitude more likely than a Boltzmann Brain :)

Oh, and I suppose a key point is that entropy in total still increases (everything life does ultimately results in entropy). This allows for localised and continued structures that reduce entropy.

@JustinThymetheSecond Let's say we have an intelligent self-replicating thing called a "bot". When a bot replicates itself the bots it makes are randomly different. Because of these changes, most of the new bots will be less intelligent, but some of them will be more intelligent. The more intelligent bots are better at making more bots because they are more intelligent. Because the more intelligent bots make new bots in their own image the next generation of bots is more intelligent than the last generation. Thus, the intelligence of the bots generally increases. This is how evolution works.

But what if life grows so exponentially that eventually all matter in the universe becomes part of something living? A rather intriguing thought, the entire universe becomes a living, intelligent thing? If life locally defies entropy, wouldn't this living thing defy entropy universally? The more these conversations are developing, the more they point to the possibility that life and intelligence are the same thing.

@Mike Nichols But my question is, where did the FIRST bot come from? it's one thing to start in the middle, to assume the existence of the bot, and then explain all further bots, quite another to explain the existence of the first bot.

@JustinThymetheSecond So you agree then that self-replicating bots can give rise to ever more complex self-replicating bots according to the principle of natural selection? Then all we need posit is the random creation of the simplest possible self-replicating bot. This self-replicating bot will then give rise to ever more complex bots that are better and better at replicating themselves until we may eventually reach an intelligent bot. The complexity of this first self-replicating bot and the likelihood of it spontaneously forming is dependent on its form and environment.

@ Mike Nichols In which case it would seem to me all you are doing is positing a different form of life than human life. Replace the word 'cell' with 'bot' and you have the same scenario. However, I do not see in this scenario where the original programming came from? How did the bot get the instruction set that allowed it to replicate? It's the origin if that intelligence that I want to understand. Can a random confluence of particles, elements, and whatever result in intelligence? By what process?

The difficulty with the 'ad infinitum' bit is losses. All life is inefficient, and these inefficiencies when applied on a universal scale ensure that a sort of 'life singularity event' cannot occur. Life uses energy to create structure, causing entropy equivalent or greater to that somewhere else. Ultimately, it'll all fall to entropy (although, I suppose heat-death is sort of the ultimate homogenous structure so perhaps it will all fall to order instead).

I suppose in the final analysis I am not completely convinced about the entire concept of heat and entropy. It is like gravity. Will gravity ever be used up? Does anything replace the supply of gravity? It alway seems to be 'pulling' everything, but that 'pull' never expires. Why isn't there an 'entropy for gravity'? Okay, so we invent some fanciful concept that gravity is just some distortion of this thing called a space.time continuum, which doesn't really exist. As far as I can tell, entropy and heat are the same kind of nebulous concept as gravity, so we invent concepts that explain them.

The problem is that 'losses' and 'infinitum' just are not compatible. if you have an infinite supply of something, then even if you suffer losses, how can you run out of it? The only infinite thing I can sort of grasp is a mobius strip, such that if you start drawing on it, you never come to an end and your line is infinitely long, even though you go over the same section time and time again. How do you measure the length of the mobius strip? Certainly not by measuring the length of the line until it comes back on itself.

And I am still trying to come to grips with the fact that you can have an infinite amount of something, and not be able to count how many you have because you can never run out of counting numbers to count it with, but no matter how many you have you can have MORE of something else. That even when you run out of your infinite number of counting numbers, you still have more of these other things to count.