Great article Gav…as to be expected from one of the original architects of all this. Realistic in terms of the technical and political challenges required to enable Web 3.0.

To Stephen Miracle’s question in the Responses, my thoughts are that all data need not be, nor is necessary to be, stored ‘on chain’. One possible solution includes API linking to existing (centralised) databases with the hashes ‘on chain’ providing a form of user controlled access (eg. in case of linking to existing Medical Records).

In a sense your reference to a large intranet is accurate, particularly in the case of a private blockchain. Of course there is an argument that a private blockchain is not too far from a centralised database as the nodes are permissioned. But you do get a distributed architecture and potential for a wide consensus mechanism.

Another potential solution to your question would be a public ‘main chain’ utilised as ‘pointers’ to some kind of p2p data sharing (as you point out), similar to IPFS, to keep large data out of the consensus requirement. But that requires some trust in the data held off chain.

Just my thoughts for all they’re worth.