Specifically I was replying to the argument that went like "the bitcoin system, in any of its futures with a bunch of non-zero transaction fees, is going to be replaced by a decentralized system that can commit to transactions that have lower or zero transaction fees, and which also otherwise provides the same benefits as bitcoin". My reply was that decentralized systems are going to have physical limitations that force their solutions to look certain ways, which would do something like, for example, explain why there were "$10 fees" in that original scenario in the first place. Your reply does not seem to share this context? 

Also, I don't mean to start a discussion about internet architecture, but ISP peering agreements do not look particularly like a cryptographic, decentralized system to me at all. I agree that the internet needs better architecture. I would call the IETF about this but I think Greg would be the one to answer or something :-). Would be sorta redundant, heh.