Don't take my word for it, please see Post #98.
Ichneumon was kind enough to think that he could respond without spelling every detail out; I realized that was not the case.
PS. I did read his post before I replied. Feel free to read his post again and identify where it actually disagrees with mine.
...and be sure to read *all* of post#98, which goes on to explain that your point is not the *whole* story...
Look, life on Earth, like just about everything else in the real world (e.g. chemistry, physics, etc.), is not driven by any *one* force or natural law.
Instead, it is driven by multiple factors, many of which oppose each other, and the actual end results are found at the point where the various factors reach a balance, known technically as "equilibrium".
For example, thermodynamics tends to push air molecules *away* from Earth, gravity tends to redirect them *towards* the Earth, and as a result an equilibrium is reached in which we have a permanent atmosphere about a hundred miles deep, of the type we're used to, instead of an airless world from which all the atmosphere has long ago flown off into space, or one in which all the air molecules are heaped together into a thin film over the ground like a coat of wet paint, with pure vacuum above it.
Similarly, you're trying to analyze the "genome dynamics" of junk DNA by considering *only* the (weak) evolutionary pressure which would (left to itself with *no* other factors) tend in the long run to clean out the junk DNA, without acknowledging or appreciating the *MANY* other factors which would (by themselves) tend to *increase* or *maintain* junk DNA.
You can't get the right answer if you over simplify the problem.
The *actual* end result is an equilibrium condition whereby the factors tending to decrease junk DNA balance the factors tending to increase it, which is obviously going to be somewhere in "the middle" -- neither a pristinely "clean" DNA, nor one completely overwhelmed with garbage. And even just off the top of my head, 90% "noise" and 10% "signal" sounds like it's in the ballpark of a realistic equilibrium point under most conditions.
It's helpful to try to boil a problem down to its essentials, certainly, but not to the point of cartoonish oversimplification, which is what you're attempting to do on this issue. The metabolic "cost" of carrying junk DNA around is not only small, it's *NOT* the only factor at work here, nor even close to the most significant.