The notion that starlight is "out there," and "on its way," is wrong. By experiment!

Many quasars are a billion light years away from us.
What we usually say (but quite incorrectly), is that the quasar's light "set out on its way a billion years ago,"
and that when it enters our eye, we see it.
That, in fact, NOTHING can POSSIBLY be "on its way" is shown by the experiment that is illustrated below.
The blue circle represents the quasar.
This particular quasar happens to lie almost exactly behind a massive cluster of galaxies.
The mass of the cluster bends spacetime, so that "light going one way" is bent one way toward us,
and light going the other way is bent the other way - as shown!
As a result, we see TWO images of the ONE quasar.
The dashed lines show the two paths by which "the light travels."
We detect the light of these two images on the sky, using the red detectors shown;
the detectors go "CLICK" every so often (randomly), signalling detection of light.
The "light travel time" for the two routes is equalized by the Delay line.
We have carefully and cleverly arranged matters so that our two beams cross each other before hitting the detectors.
That allows us, if we wish, to put a beamsplitter at the intersection.
Try rolling your mouse over the image: that's how you put in the beamsplitter!

What is a beamsplitter? It is a piece of glass that has enough aluminum deposited on it to reflect HALF the light that strikes it.
The other half of the light is transmitted. (That is the way we normally describe it; however, none of it is true, as this experiment shows!)

The result is astonishing.
Now only ONE of the detectors goes off!
(Note: if you were doing the experiment, and you adjusted the detector positions slightly,
you could reverse the role of the two detectors,
or even, if you wished, get the uninteresting result of both detectors going off randomly)

The illustration, of course, shows the "astonishing" case.

Why is it astonishing?
If you block either of the two beams, both detectors resume going off randomly.
 This means that when neither beam is blocked, information MUST be arriving by BOTH paths.
Remember that "clicks" occur only every few seconds, and light "goes at 299,792,458 meters per second."
So it isn't that light is "always arriving by both paths."
Yet with both beams un-blocked, only ONE detector CLICKS.
If our picture is correct, what causes the clicks MUST be arriving by BOTH paths.
But how can that possibly be?
That would require that a billion years ago, the light leaving the quasar "knew" that you were going to put in the beamsplitter,
and so it divided into two parts going in just the right directions so as to meet at the beamsplitter and interfere with each other constructively and destructively to produce the observed result.
This tells us that all the notions that we commonly use is discussing the universe are WRONG.

The good news is that we DO know the RIGHT answer, and that is Quantum Mechanics.
In fact, this experiment would never have been performed, were it not that quantum mechanics predicted this absurd result.
This experiment was conceived by people who felt quantum mechanics could not be right (because it denies the reality of the world).
Hey: keep on using the old language (I do too), but––never forget that it is ultimately WRONG!

[Note: the experiment as described has not yet been done, simply because we do not yet have optical delay lines of weeks or months.
But, the equivalent experiment has been done many times in the laboratory.]


(By the way, the fact that there is no actual universe is well known!)