Isaac Asimov’s science book View From A Height dedicates an entire chapter to explain concept of entropy. Assuming you have a decent background in the physical sciences, it does an excellent job, even better, I daresay, than my thermodynamics professor managed. So far the whole book has been a worthwhile read, but that essay in particular may be instructive to those interested in the topic.
Asimov concludes the chapter by presenting a very appealing hypothesis:
[E]ven if the universe were finite, and even if it were to reach “heat-death,” would that really be the end?
Once we have shuffled the deck of cards into complete randomness, there will come an inevitable time, if we wait long enough, when continued shuffling will restore at least a partial order.
Well, waiting “long enough” is no problem in a universe at heat-death, since time long longer exists there. We can therefore be certain that after a timeless interval, the purely random motion of the particles and the purely random flow of energy in a universe at maximum entropy might, here and there, now and then, result in a partial restoration of order.
It is tempting to wonder if our present universe, large as it is and complex though it seems, might not be merely the result of a very slight random increase in order over a very small portion of an unbelievably colossal universe which is virtually entirely in heat-death.
Perhaps we are merely sliding down a gentle ripple that has been set up, accidentally and very temporarily, in a quiet pond, and it is only the limitation of our own infinitesimal range of viewpoint in space and time that makes it seem to ourselves that we are hurtling down a cosmic waterfall of increasing entropy, a waterfall of colossal size and duration.
This is an intriguing idea. It suggests an alternative possibility for the fate of the cosmos than “eternal coldness”. Presuming that black holes don’t end up consuming all the matter in the universe, and proton decay turns out to not occur, then it might be possible to square a sort of steady-state theory with the existence of entropy.
Entropy isn’t merely disorder—though disorder is certainly a part of it. The Second Law of Thermodynamics tells us that energy does not spontaneously flow from cold areas to hot areas. Only by applying work can we force the flow in the opposite direction. Work, however, can only be extracted from the flow of energy from a hot reservoir to a cold one. Thermal efficiency is based on the temperature difference between the two reservoirs:
Where 
Note that the absolute quantity of energy in either reservoir is irrelevant. We are only concerned with their relative values. Work cannot be extracted by placing two equally hot reservoirs in contact, even if both are at 10,000 °C.
It is, of course, theoretically possible that random motion of individual particles might provide a very small about of usable work. However, this is exceedingly unlikely. Asimov gives an extreme example: could water in a pot freeze while the fire beneath grows hotter? Theoretically, yes. The laws of statistical thermodynamics do not forbid it. But even if the entire universe were filled with such pots, and we waited for eons and eons, we would not realistically expect to see a single pot significantly cool, let alone freeze.
As time goes on, we will approach universal thermal equilibrium. Extracting useful work, of any form, will become impossible. Work is energy, and life depends on a continuous source of energy, so all forms of life will perish.
This, naturally, can be a bit frightening to think about, especially if you are very young when you learn about it, as I was. A cosmological expiration date seemed like a very serious problem, because it meant that all of our efforts would necessarily be in vain. If the universe ends end regardless, social morality seems farcical. Rank hedonism looked like the only alternative, so my early attempts to reject
Objectivism helped me out of that trap, though its presumption of an inexhaustible universe remains problematic. But that doesn’t matter if morality is not social but personal, and the purpose of existence is Apollonian joy rather than a greater obligation.
Still, the possibility that the universe, even after heat death, can randomly reorganize will offer hope to the last mind that joy won’t go out of existence forever. A silent eternity would pass, and then…something. A universe appears again from darkness.
Doesn’t that sound familiar?
I’m tempted by this hypothesis not merely because it offers hope for the universe, but also because it helps get around one of the frequently-asked unanswered questions about the Big Bang: what happened beforehand?
So far, we don’t know. Did time even exist before the Big Bang? Did conservation of mass-energy already apply? I haven’t studied the astrophysics to pretend to answer such questions.
Random reordering gets around this issue. The universe we know is a ripple is the wider open of equilibrium particles. Entropy is maximized. Everything is in the lowest possible energy state. By chance, some of this matter happened to organize itself. Net entropy will now continue to increase, so I think this is allowed.
Since the time for a “dead” universe to randomly form an orderly patch of significant size must be incredible, it would be no surprise if the photonic evidence of previous ordered periods had been entirely absorbed or diffused. Photons, being massless, don’t decay, but in such a long period would no doubt either be absorbed by the near-equilibrium particles. The remainder would be spread out over such a large area that the number of photons are simply swamped by more recent light. No instruments could possibly detect them.
But just because a hypothesis would be personally comforting does not forgive a lack of evidence. We should seek contradicting evidence for all hypotheses, regardless of our feelings toward them. Falsification is how science works.
Does random reordering fit the evidence we’ve already gathered about the early universe?
I’m little more than a layman, but generally speaking, the answer is: not really.
We have a pretty good picture of everything that happened more than a second after the Big Bang, and for a good while before that. A lot is based on astronomical data, such as the cosmic background data gathered by COBE, WMAP, and Planck. The remainder comes from particle accelerator experiments, from which physicists can build up models that extrapolate back even further.
The current theories don’t look very much like the result of randomly reshuffling baryons or leptons. It looks like a lot matter being created ex nihilo, with somehow antimatter being in the slight minority. Possibly a reshuffling at a much lower scale occurred, well after proton decay and whatnot evaporate the particles we’ve come to expect—I have an idea about how that might work, but I won’t burden you with more unwarranted speculation.
More study is clearly needed: better space telescopes and more powerful particle accelerators to give us data, faster supercomputers to process it, maybe some mathematical breakthroughs. It will probably take awhile to get a better estimate on the odds, but until then, I would put a low prior on the likelihood that our universe is a temporary reprieve from heat-death.
Full-sky image of the cosmic microwave background, gathered over nine years by the Wilkinson Microwave Anisotropy Probe.
Source: NASA/WMAP Science Team
However, there is a bigger philosophical question here: a reorganization hypothesis does not explain the origin of the universe, it just moves the cosmological problem up a step. Our universe being a mere ripple on a larger heat-dead ocean doesn’t tell us where that ocean came from. Did that universe have a Big Bang? Is it cyclical? Is it Steady-State? We still have to answer the same questions, and now we have less data!
(Of course, if it does turn out to be true, then we’ll just have to make do with less data. But that’s a methodological question.)
Trying to explain why there is something at all isn’t necessarily a hard question, but to explain why existence started to exist 13.8 billion years ago is a bit trickier. At this point, perhaps the simulation hypothesis is a decent pseudo-explanation. You can’t make very many predictions with it, so I wouldn’t call it a real explanation. That said, it does manage to constrain our anticipation to some degree. And there is some evidence for it.
Whatever reality is, we’ve still got at least some distance further to walk on the path to Truth. It’s tempting to take a short-cut through speculation and a priori arguments, but those are distractions. If we want to be sure, we have to do things right. Proposing hypotheses is part of that process—but so is rejecting them. As tempting as random reorganization is, I’d be happy to reject it with a little counterevidence.
Nathaniel M. Routh 