I sometimes wonder if we couldn't have completely different public internet topologies if
a) Wireless local networking was invented and popularized earlier
b) We had transitioned earlier to IPV6 or some other protocol with an address space as huge, thus making NAT not as pervasive as it became.
b) We didn't have hordes of VCs financing walled gardens and social networks.
It's simpler than that. The only thing that would need to change is spectrum allocation. We need unlicensed spectrum with higher power limits in longer range radio bands. It's a miracle what radio engineers have done with a tiny slice of unlicensed spectrum near 2.4 Ghz. Imagine what they could do with a few unlicensed lower frequency VHF/UHF channels.
I think "wired" (really bundles of optical fiber) will win out anyway due to the vast capacities it can provide. There will always be gaps between populations and to have a reliable link, you will need these dedicated mainlines, which will also help scale datacenters. Perhaps we would have more p2p tools, but the public internet would have a similar topology too if it were to have a similar capacity.
a) Wireless local networking was invented and popularized earlier b) We had transitioned earlier to IPV6 or some other protocol with an address space as huge, thus making NAT not as pervasive as it became. b) We didn't have hordes of VCs financing walled gardens and social networks.