Re: Getting a probe there?
As I understand it (or not) quantum entanglement does not provide faster-than-light communication.
While I wouldn't want to make a pronouncement on the possibility of faster than light or not of quantum communication (a subject even Einstein could get things wrong about: see John Bell), that explanation slightly dodges the quantum communication aspect of the question.
Undoubtedly, as in the description, if you tell two friends that you'll send them light beams of red and blue, send them off in different directions and then send the light beams, 1. they will know the other person's colour before the other person could communicate the signal to them, 2. the information conveyed only travelled out from you at c (which choice you made about who to send what colour) and <c (the setup in the first place). In a quantum context that choice about red or blue is representing the outcome of state collapse (if you believe in state collapse), which is generally thought to occur at the time of measurement. So the explanation is, in a way, a hidden variable explanation. If that hidden variable doesn't exist, then the outcome for Alice determines the outcome for Bob at the time the measurement is made.
The quantum entanglement part is this: Alice and Bob head out with entangled particles. Measurement on a particle collapses the waveform and you know from the answer what answer the other person will get. However, that's not how you intend to communicate. How you want to communicate is in deciding how to make the measurement.
Say instead of two colours we have four. Blue, green, red, yellow and Alice and Bob can both measure for either blue/red or green/yellow. If A measures for blue/red and Bob measures for blue/red they will always get opposite answers (as before). Same if they both measure for green/yellow.
Now imagine that I'm actually sending purple and orange. Purple shows up as red or green, orange as blue or yellow (you'll notice this isn't a lecture on colour theory). If A and B make the same measurement then when they compare later they get compatible answers. If they make different measurements then they can still work out what the other person would have got if they knew what the measurement was. My purple/orange choice is a hidden variable.
But this isn't what happens with entanglement. A measures r/b, and gets an answer. B measures g/y. No matter what answer A got, B has a 50:50 chance of g or y, the measurements are independent. Even if they make the measurements at the same time or at an interval shorter than the time light can travel between them. It's not just a case of the blue going one direction and the red going the other (in which case you could assume there's some hidden variable), it's that the measurement A chooses to take appears to influence the answers B can get instantly. This gives rise to the Bell inequality https://en.wikipedia.org/wiki/Bell%27s_theorem and one of the great missed Nobel prizes.
But, for communication of information it's no good. From B's point of view, while their answer does depend on the measurement A made, they can't tell which answer A actually got. And this is also due to the "state collapse" (probably it doesn't) part of the entanglement. If A always got red when measuring r/b and green when measuring g/y then you can construct a noisy channel (this is just off the top of my head, probably a more efficient way):
A sends "1", by reading in r/b requires four measurements:
A -> B
r -> (r/b) b
r -> (r/b) b
r -> (g/y) g
r -> (g/y) y
Note the first two results would be fixed, the second are 50:50 different.
Similarly A could send "0" by making four g/y measurements. Now the first two b measurements would be 50:50 to be different and the second two would be yellow. So you have to discard 50% of these quartets, but the others you know what the original measurement was.
But you can't do this either, and the reason is that you can only be in one pair of states. It's perfectly possible to have a deterministic "A always gets red when measuring r/b", we could do that with the lights example. But you can't simultaneously have the full set: "A and B get different answers", "A always gets red if r/b", "B gets 50:50 g/y if A measures r/b" and "A always gets g if A measures g/y", because of the symmetry of the entanglement. So though something is going on, the way the statistics work mean you can't use it until you compare the answers A and B got.
(I mentioned being dubious about state collapse, weak measurement experiments suggest it doesn't happen. If Alice believes Hugh Everett, she has gone out of phase with the Bob who measured red after she did, and can therefore never meet him again to compare notes.)