Hector Martin

There are N people in the network. As N grows, each individual person also has more potential people to transact with - transactions would be O(N²). Since each person has a copy of all transaction history, the network requires O(N³) storage overall.

Ah, so N isn't the size of the block or the storage needed because the block is bigger. It's the scaling factor of transaction counts as networks in general get bigger. So every network scales by O(n3)?

No, the problem here is requiring everyone to have the entire transaction history stored locally. I'm oversimplifying of course, but I'm saying overall throughput scales with O(n²) (and thus individual bandwidth with O(n).

Yeah this checks out for me, I want to use bitcoin more, but I need more people available to trade bitcoin with. If we double the number of people on the network we more than double transaction potential (handshaking problem).

But if we just want to double the number of *transactions* supported, naively doubling the blocksize achieves this.

Anyway, this is progress. Is your point that bitcoin is stupid *because* everyone has to store all the transactions?

That was an example of another inherent problem with the network. Another is the O(n²) bandwidth requirement for the same reason (must receive all transactions, not just your own) which blocksize doubling can't fix.

Basuically what i'm saying is the scaling properties are fundamentally broken; no constant improvement by using more efficient encodings or increasing sizes or what have you will fix it, and I do not believe a deep fundamental change would achieve consensus and work in practice.