That's actually kind of the point of GP: the geographic limit here doesn't apply to the power of the sovereign around whose notional domain the line is drawn, it applies as a rule adopted governing relations between sovereigns but does not in practice apply where no sufficiently powerful sovereign has an interest at stake for which they will Lodge a protest (backed with the implicit threat of force.)
Parties using geographic games as a way to reject affiliation with recognized sovereigns are not evading the power of such sovereigns, only the protection of them.
Another key difference is that your phyle follows you around, geographically, while your nation's jurisdiction ends at its borders. If I'm an American expat interacting with another American expat in Shanghai, I'm still subject to Chinese law. If I'm a Neo-Victorian (to use one of the phyles from Diamond Age) interacting with another Neo-Victorian in Hong Kong, I'm subject to Neo-Victorian law. A block over could be a Han Chinese (also using the Diamond Age phyle, but in this case it corresponds to a real ethnicity in our world) enclave, and its citizens would be subject to Han Chinese law when interacting with other Han Chinese. If, as a Neo-Victorian, I cross the street to do a business deal with a Han Chinese, our dealings are governed by the Common Economic Protocol, the set of laws common to the federated system as a whole.
A nation-states jurisdiction does not end at it's borders unless it chooses for it to, which few do entirely. It ability to practically enforce it's rules may or may not do so, depending on the power of the state and power and interests of other states who may be concerned with the specific matter (especially, but not exclusively, the one in whose borders the target of the enforcement is located.)
It also was very pessimistic on technocratic solutions for the "good nurture" problem. Basically nothing can replace a real parent coaxing a self-expanding mind into beeing.
He's pretty good at imagining it too - The Diamond Age was published in 1995, just as the Internet was beginning to take off, and yet it envisioned The Internet, videoconferencing, smart paper, drones, and a whole bunch of other technological innovations that didn't catch on for 15 years or so. Cryptonomicon was published 10 years before Bitcoin but fully described two of its biggest use-cases, remittances and digital gold. Snow Crash had the "America is good at only 4 things: music, movies, microcode, and high-speed pizza delivery" line in 1992, 20 years before it became true and right as the rest of the country was celebrating us becoming the sole superpower.
I've never read anything else quite like it, and honestly don't know what I think of it. Are all of Neal Stephenson's books like that?