I agree that Payment Channels is a good step forward but there's the hard limit on how many channels can be open. I'm confident that there will be proposals similar to Poon et al. [1] to create a third layer for networks of payment channels where a Blockchain tx can be used only in case of uncooperative channel counterparties.
I’ve read that paper, and I don’t quite understand how it would work when you need multiple clients to cooperate for a single client to make a payment. I.e. if the funds of 100 clients is at some multi-sig address, you’d save blockchain space, but payment from any of these 100 clients would require simultaneous cooperation from the remaining 99 clients. I have a hard time seeing how this would work in practice.
[1] https://www.tik.ee.ethz.ch/file/a20a865ce40d40c8f942cf206a7c...