e.g.: Imagine a fast but slightly centralized Twitter chain (only 21 validators let's say), committing hashes of every day of tweets to a decentralized public blockchain.
I'm not 100% sure whether that tweet posted by my friend 5 minutes ago is legitimate, but I don't need to be 100% certain to answer them and carry on the casual conservation. I can verify though, that the Trump tweet bordering on declaration of war posted 24h ago is indeed real, because the hash of the whole day of tweets was committed to a strong blockchain, and the author of the tweet didn't post any subsequent objection to it. I can proceed to hide in my bunker, confident in my choice.
[1]: https://en.bitcoin.it/wiki/Protocol_documentation#Merkle_Tre...
[2]: https://eprint.iacr.org/2017/913.pdf
[3]: https://github.com/ethereum/wiki/wiki/Sharding-FAQs#what-mig...
It's a strong consequence of the 0,1,N rule. The correct number of networks is clearly not "zero", and Metcalfe's Law ensures N networks are worse than N-1 networks with the same nodes in, so you get one network at the limit.
Once upon a time I'd have called it "the Internet" but that's really redundant at this point, just call it "the Network" and have done with it. One network, for everything, for everyone, for ever.
This isn't even a new idea, the Bell company couldn't build a single telephone network across the continental US until a better electrical repeater technology existed (today we barely even think of this problem because the transistor is so ubiquitous its original purpose becomes invisible) but they insisted it was a single network anyway, because that's very obviously what you actually want even if you still haven't invented the technology to do it.