I suspect you misunderstand my objection and that I may have used only half of the appropriate analogy
A universe in which your proposed ubiquitous low-matter low-energy interstellar computers exist is one in which space-based self-replication and manufacturing is a thing that happens. This implies the existence of a whole slew of 'ecological niches'. Indeed, the sort that is generally thought of in these circles (more-or-less industrially turning large amounts of matter near stars into stuff that intercepts light and uses the resultant energy for something or other) is rather simpler, is more similar to the demonstrated cases of terrestrial biology / human industry, and has more matter and energy available than what you propose. The low temperature low energy devices would be more akin to crazy deep extremophile lithotrophic bacteria or deep sea fish on Earth, living slow metabolisms and at low densities and matter/energy fluxes, while things in star systems would be akin to photosynthetic plants and algae at the surface, living at high densities at high flux.
In any situation other than perfect coordination, that which replicates itself more rapidly becomes more common. You will have adaptation and evolution. It doesn't matter if more computation can be done in one place than another - in terms of sheer matter and energy, that which uses high energy fluxes and large amounts of matter will replicate to large numbers and be dominant in terms of amount of stuff and effect on the physical universe. Other stuff could still exist, but most stuff would be of this faster heavier type. Niches will be filled. And a stellar system niche is not akin to the deep ocean if an interstellar niche is compared to the surface of the Earth, if anything it's the opposite. The deep sea niche may be where you see all kinds of fascinating bioluminescence and long distance signaling epiphenomena that these organisms care about and of a sort you dont see at the surface, but in terms of biomass the surface niche dominates. Furthermore, competition amongst different things mean they often do things inefficiently so as to gain advantages over each other - those that do become more common faster.
The low temperature low energy devices would be more akin to crazy deep extremophile lithotrophic bacteria or deep sea fish on Earth, living slow metabolisms and at low densities and matter/energy fluxes,
Hmm I think you misunderstood my model. At the limits of computation, you approach the maximal computational density - the maximum computational capacity per unit mass - only at zero temperature. The stuff you are talking about - anything that operates at any non-zero temp - has infinitely less compute capability than the zero-temp stuff.
So your mode...
After a 6+ month hiatus driven by grad school and personal projects, I am finally able to continue my sequence on astrobiology. I was flabbergasted by the positive response my last post got, and despite my status as a biologist with a hobby rather than an astronomer I decided to take a more rigorously mathematical approach to figuring out our biosphere's position in space and time rather than talking in generalizations and impressions.
Post is here: http://thegreatatuin.blogspot.com/2016/03/space-and-time-revisited.html. Seeing as this post is an elaboration on the last one, I am posting a link rather than reproducing the text.
To summarize, I found some actual rigorous observational fits to the star formation rate in the universe over time and projected them into the future. These fits show the Sun as forming after 79% of all stars that will ever exist, and that 90% of all stars that will ever exist already exist. This makes sense in the light of recent work on 'galaxy quenching' - a process by which galaxies more or less completely shut off star formation through a number of processes - indicating that the majority of gas in the universe probably won't form stars if trends that have held for most of the history of the universe continue to hold. It relies heavily on analysis I began in comments on this site a few months ago.
I then lift two distinct metallicity normalizations from a paper that was making the rounds here a while back ("On The History and Future of Cosmic Planet Formation"), in an attempt to deal with the fact that that is a measurement of STAR formation, not terrestrial-planet-with-a-biosphere formation. Depending on which metallicity normalization you use (and how willing you are to take a couple naive assumptions I make in order to slot the math that is too complicated for me to comment on on top of my star formation numbers) the Earth shows up as forming after either 72% or 51% of all terrestrial planets.
These numbers are remarkable in how boring they are. We find ourselves in an utterly typical position in planet-order, even if I am wrong by quite a bit. We are not early. Of interest to many here, explanations of the so called Fermi paradox must go elsewhere, into the genesis of intelligent systems being exceedingly rare or the genesis of intelligent systems not implying interstellar spread.
Now that I seem to have a life again, I will be getting back to my original plan next, talking about our own solar system.