Why on earth would you compare these to the human genome, instead of the smallest archeobacteria? Evolution adds complexity over time, we know this, we just need something to replicate first.
A better analogy than the Linux kernel might be a quine.
There is still an enormous jump in complexity between a few nucleobases and the smallest archeobacteria. Maybe a full Linux kernel is not the best illustration, but the idea of a kernel in general is a remarkably apt one. At the least, we're talking stuxnet or something.
Even the simplest life is incredibly sophisticated.
Why wouldn't one make the comparison with the human genome? If the evolutionay mechanisms credited with responsibility for the jump from nucleobases to archeobacteria (which is still incredible, if true) can't be fully and comfortably accredited with the nucleobase to human genome jump then said evolutionary mechanisms are suspect. According to the notion of evolution inherently adding complexity over time, nucleobases to human genome just requires some very large time. If we do know this with such certainty, then my comparison is both apt and reasonable.
A better analogy than the Linux kernel might be a quine.