Kindle. I went in cold, forgetting that with Egan the blurb on the back is not an accurate guide to what he has in mind. The afterword tells us that he was playing Raymond Smullyan's didactic game for Einstein's general theory of relativity, and not the promised Star Trek V mystical journey into the centre of the galaxy to meet some isolationists. Being lazy I quickly stopped following the lessons too closely, and I'd say it's not worth reading unless you do try to keep up.
Initially I liked how his universe respected the physics we know, with life (not as we have it) moving around on a galaxy-wide data network with transit delays of millennia, but of course the didacticism gave him no other option. I was less of a fan of the sentients having backups, given his broader conception of consciousness in Permutation City and readiness to lean on quantum encryption. Why should consciousness remain local in such a system? More fatally Egan struggles to keep his epistemics straight: how do the Splinter critters know about the hub, the void and all that when they haven't been outside for generations? Why ever would they expect their local observations to hold universally? Why would they ascribe a speed to light? So much (too much) ontology is taken for granted.
As a story it reminded me of McGahan's posthumous sapient eco disaster. The two-track structure takes the edge off the didacticism, and many explanations are too hurried; more diagrams would have helped, as Egan admits on his companion webpages. Around the mid point I realised that the stories must be very separate in time, which while skilfully suggested made at least one into deterministic history, robbing them both of forward momentum. For all his understanding of science, Egan shows less awareness of and interest in how societies may need to be structured to thrive; it seems likely that some basic empathy, truth telling or goal sharing is necessary for collaboration to arise, and that seems like a prerequisite for technology. He did make me wonder if nature knows about binary search. Oftentimes there's a bit too much science and not enough engineering.
Goodreads has many words spilt on this one. Egan's page suggests this work is insufficiently self contained. He emphasises that (as is usual with physics) almost everything he says are white lies, implying that he is not playing with the right foundational concepts. For all his protestations I didn't feel like the Aloof ever turned up.