> But blockchains do introduce a technical capability that never existed before: shared, “trustless” access to and control over data.
I don’t think blockchains control access over data at all without the integration of other technologies (i.e. using public key authentication to prove to some third-party service that you hold a key that performed a blockchain transaction). They only control permissions to update things on-chain and everything on the chain is still publicly readable anyway. PKI is capable of doing the same thing and has existed for a very long time.
> Sure, other people could keep their own copy of that same data, and sync with you somehow, but most pre-blockchain data services lack replication capability out of the box, and as far as I know there are no non-blockchain data services that can give you any assurances at all that data has not been tampered with.
Look at the existing NFT mess. The problem here is that the blockchain _doesn’t_ replicate anything other than a bit of metadata for these types of applications — the content still has to be stored elsewhere, often on a classic web server which is centralised and can go down any time.
If you want an example of something that replicates data and guarantees integrity then look at IPFS.
> Federated application services capable of competing with big tech have been a dream and an objective of open-source advocates for decades, but no one has been able to deliver. I would say the major reason is because trustless data syncing at scale is orders of magnitude harder than running a single standalone server, and all of the solutions until web3 attempted to perform data syncing at the application layer—which means a lot of duplicated effort, and almost zero opportunity for reuse between projects.
Blockchain’s answer to “federation is hard” is just to replicate the entire chain to all participants and have all participants repeat the exact same work to arrive at the same position. This is not exactly a glamouring success story of how to avoid “duplicated effort” — in fact, duplicated effort is a fundamental part of the blockchain architecture.
> Blockchain tech moves replication and tampering protection to the data layer, as a generalized service that is relatively trivial to set up. The promise of web3 is that shared-data applications can be built on top of that data layer in the same way that complex centralized applications can be built on well-defined centralized data layers like the file system and the relational database.
This isn’t really true though — it’s more like the blockchain provides a tamperproof metadata store. Putting actual quantities of information onto a typical blockchain is either prohibitively expensive computationally (it needs to be replicated everywhere) or prohibitively expensive financially (Ethereum gas costs as an example).