I assume you are referring to the article yesterday that quoted Buterin as saying it will handle that much.
He is saying that the various scaling methods multiply on each other so that will allow Ethereum to go from like <10 tx/sec to a million. He only mentioned plasma and sharding, but I believe the three main scaling techs are sharding, plasma, and raiden. Sharding splits the network to basically run many smaller Ethereum chains in parallel. Plasma I believe is child chains. And Raiden is basically the LN for Ethereum. If all three of those scaling techs work, they each only need to scale Ethereum by about 50x to get ETH to 1 million tx/sc. That is assuming they are all compatible, like you use Raiden Network on top of a plasma side chain that is itself off of a shard. Assuming these all work together and that they are all actually secure, then yeah I think 1 million tx/s will be possible.
wow sounds like a clusterf*ck of centralization upon centralization lol. It's centralization all the way down...
not sure why you say that. None of that is centralization.
It does sound like a clusterf*ck of complication upon complication for sure, but not centralization. Side chains, layer 2, and sharding are all decentralized.