I've now read Koonin's paper. It wasn't as ground-shaking as it might have been, though it would appear to be yet another nail in the coffin in the conventional, Dawkinsian version of evolutionary theory, and another one up for Gould. Here's a few thoughts - feel free to correct me if I have misunderstood things.
Probably the most significant thing about it was the frank acknowledgement that the conventional Darwinist model (random mutation, natural selection) doesn't correspond to the evidence. The specific issue addressed is the fact that the "tree of life" doesn't really exist - for the most significant biological transitions, divergence from common ancestors seems to take place over a very small interval of time, and subsequent evolutionary changes are "small fry" in comparison. The most famous example of this is perhaps the Cambrian Explosion, but Koonin is more interested in more fundamental steps - the appearance of the first viruses, cells, and the appearance of eukaryotic cells.
Koonin suggests that the reason for this is that the big evolutionary steps take place as a consequence of "horizontal" transfer of genetic material, rather than "vertical" descent with modification. As a consequence, it is only when stable configurations appear that identifiable descent lines become apparent. Because genetic transfer of one means or another drives large scale diversification (the Biological Big Bangs), we are unable to see the antecedents of subsequent stable organisms.
Koonin stresses the analogy with the current, inflationary/Big Bang model of cosmology. To be honest, I can't see the point in labouring this, unless for ideological purposes - to stress that it is an alternative to teleological processes. In an ateleological universe, because the scale of the phenomena are so different, the analogy can be little more than a coincidence, and really is barely worth drawing attention to - it is an observation of the same sort as saying that electrons orbiting nuclei are a bit like planets orbiting a star. On the other hand, if we are in a teleological universe, that fact itself is much more significant than the fact that two processes are analogous!
Also, although this model provides a better explanation of the observations (of the nature of the tree of life), there are many unresolved issues. For example, is this mechanism more likely to generate new biochemical structures and functions than conventional neodarwinism? Certainly, a new protein or gene has a larger space in which to find a matching counterpart - but it also has a larger space in which to become distracted or damaged. Also, whilst it is easy to say that when a stable organism is formed, the inflationary phase switches off, more work would need to be done in explaining what triggers the inflationary phases, and how come they switch off at all. Of course, this isn't saying that Koonin is wrong, or that the hypothesis isn't worth exploring - especially since it does provide a better explanation.
Telic Thoughts (I think) also hints that Koonin may be too blithely accepting endosymbiotic models. The example they give is that the alpha-proteobacteria which are viewed of as most similar to mitochondria actually appear as bacteria substantially later than mitochondria - so the hypothetical absorption of bacteria by eukaryotic cells to get their current form may not be as simple as is suggested. (Not that it is suggested to be exactly simple!)
I hope this makes sense, and hasn't distorted either Koonin's work or the Telic Thoughts article.