You've made a few errors in your fun theoretical musing:
Oh goodie, someone who talks like this...
1) Most of our DNA is, in fact, superfluous, as far as we can tell. Less of is superfluous than we thought a few years ago, but more than we thought ten years ago.
Sounds like we've got it right this time, though! Assuming you're referencing Junk DNA, there's a world of difference between "no discernable function" and "superfluous". Additionally, even with an upper bound in DNA functional density, there's no reason to assume there isn't also an optimal upper bound to superfluous-to-functional DNA ratio. Adding a massive chunk of DNA to an organism is going to have some effect, you have to agree, and with no functional purpose there's very little evolutionary reason not to just whittle it down to nothing. After all, if there was actually a benefit to more superfluous DNA, evolution's had plenty of time to add it.
So, I guess, thanks for really not saying anything at all.
2) Evolution does not tend towards optimization. It trends towards "good enough". Extra DNA only matters if you're a bacterial cell, and the rate-limiting step in your growth is the replication of your entire cellular DNA. In many ways, for a human, noncoding DNA is beneficial - random errors and strand breaks are less likely to corrupt important parts of your file if a good chunk is noise anyway.
There is a lot of naiveté in this part of your response. First, "good enough" is a form of optimization; it's just an optimization across factors other than straight efficiency. Second, there is a cost to copying useless DNA, bacterial cell or not, and unless there is a benefit to offset the cost, an organism that sheds that DNA will be fitter than one that doesn't. If, for example, I stuffed a kilogram of extra DNA into your cell, it'd probably matter, even if you aren't bacteria. You're asserting, without any logic, that this cost fits into some magical "good enough" threshold you have just conjured. Cool threshold bro.
3) It has, technically, already been done (although not released). Venter's synthetic life form has genetic "watermarks" embedded in it. Nothing as awesome as an entire book, but the premise is there.
It's painfully obvious that my "what if it's already been done" statement was not referencing other synthetic human works, but rather the natural genome. Just a heads up, but your genes may be missing some padding around your Broca's Area expression ;)