It seems to effectively be equivalent to "Imagine we have a refrigerator with a massless, frictionless motor that also doesn't need to be plugged in. Isn't it crazy that this refrigerator would violate the second law of thermodynamics?"
If that sounds like a cop out, remember that the Demon is a metaphor that's supposed to represent an as-yet-undiscovered technology. Like, instead of a literal massless door that opens and closes, something like, "What if we discovered a clever way of designing a little gap between two chambers and ionizing the particles such that only hot ones go through."
I think you're leaving something out of your description. We have passive selective barriers that will admit low-energy particles while reflecting high-energy particles. (Cotton, say, will reflect visible light while being transparent to infrared light.)
So what do you get by hanging a sheet between two rooms? Nothing.
I think Maxwell's Demon will only let hot particles pass through in one direction, but that's nothing new either; we have one-way mirrors.
We don't, actually. We have semi-transparent mirrors that allow a small fraction of the light through, and they appear "one-way" only when one side is vastly brighter than the other.
There’s no apparent relationship to the energy gain per separated particle and the energy used by the door. That makes it hard to prove it’s not generating energy.
That’s what makes it an interesting problem.
That could be fine for a student that is learning physics but not for a scientist.
Treating something as unquestioningly true is a sign of belief and religion.
Now, for most everyday situation it is practical to assume that a device can't be creating energy from nothing, can't be ordering atoms in two containers without using energy, because we know there exists second law of thermodynamics.
But Maxwell's Demon is an experiment specifically aimed at testing thermodynamics.
We can use the second law of thermodynamics exactly because we expend/expended effort to test it to a sufficiently high degree and we are revisiting as we learn new facts and push the boundary.
This is essentially the same as:
> The second vital piece of the puzzle was the principle of erasure. In 1961, the German American physicist Rolf Landauer showed that any logically irreversible computation, such as the erasing of information from a memory, would result in a minimal nonzero amount of work converted into heat dumped into the environment, and a corresponding rise in entropy.
Per Liouville's Theorem, information about your surroundings is a conserved quantity. A physical system (persistent along a time-like curve) that has more information about a particular part of it's environment (eg "the content of this movie is so-and-so") in one light cone of a event than the other, must have correspondingly more information about some other part of it's environment (eg "this heat sink that I am erasing scratch space with has particle velocities within bounds such-and-such") in the other light cone compared to the first one.
If one light cone has much larger quantities of compressible ('cold', kind of) information than the other, such that interesting information about our surrondings always appears on the opposite side of events from it, we call that light cone "the past".
I don't know that we have a better explanation than the anthropic principle for why one one of our light cones is like that, though.
Hmmm. I had the impression that the flow of time was implicit in the fact that we talk about "light cones" instead of "light spheres". Is that not the case?
0: Quantum field theory makes this more complicated, since particles are more like amplitude distributions over possible graph edges, except events also aren't vertices but distributions over vertices, and if you can figure out a way to explain this that actually makes sense you might win a nobel prize, but it empirically is well aproximated by a graph at macro scales.
But more seriously (and with tons of respect, because these topics are difficult) - why invoke all these complicated concepts? When we discuss statistical mechanics it is kind of understood that the ferromagnet we try to describe is not placed on the innermost stable orbit around a black hole, and similarly when I look in a mirror there is little need to worry about the difference between P and CPT.
I guess one can say that the essence of a physical system is not always in the fundamentals of physics?
Having said that, I do not know what "intelligence" is, so I do not know how to relate it to entropy.
I don't need LSD!
> When we discuss statistical mechanics it is kind of understood that the ferromagnet we try to describe is not placed on the innermost stable orbit around a black hole
If your theory of statistical mechanics requires that, then your theory is wrong, or at least badly incomplete. That might not be a problem you need to care about for practical reasons, but it is a problem.
> I guess one can say that the essence of a physical system is not always in the fundamentals of physics?
Depends on what you mean by "essence" and "fundamentals" - you're not going to have good time trying to analyze a computer [0] running a game-of-life sim in terms of it's constituent quarks, obviously - but there's usually something useful to be learned from the underlying physical laws.
0: (whose primary descriptive quality is that it's)
Or do you mean you also have a generator to recover the energy from the temperature differences it creates? So it's not really doing refrigerating because you're using all the cooling power to run the motor?
However, reversing entropy requires an amount of energy commensurate with the amount of entropy you are reducing. Maxwell's demon would appear as a violation of this.
A massless, frictionless motor could trivially reverse all the entropy in the universe given enough time, but it is not clear that Maxwell's demon REQUIRES something equivalent.
https://www.nature.com/news/the-unavoidable-cost-of-computat....
I'm also not sure why Maxwell's demon would require erasure.
https://en.m.wikipedia.org/wiki/Maxwell%27s_demon
I think it's clear that by now the 2nd law always wins. Always.