I saw @kelseyhightower talking about git and web3, and I've talked about a similar topic before, so I'll rephrase it here in this thread.

In my vision of web3, the application's data is stored in a local repository like git, attached to a p2p network, and synchronized in real-time. For example, @kelseyhightower writes a blog and saves it to his local device repository, broadcasts the root hash update notification to subscribers on the web. My device listens to it, and it automatically synchronizes the repository. Then, I opened a local application to see my subscribed blog.

This solution has several advantages:

1. The user is guaranteed ownership of the data so that his data is on his device, and anyone else cannot delete it.
2. Any intermediate node can cache data, speed up synchronization, improve the experience, such as home routers, such as buying a cloud service automatically backup data. But there are git checksum mechanisms to ensure that intermediate nodes can not manipulate the data.
3. The centralized blogging system can also play a role in this system, such as search, aggregation, and recommendations, but these services only have deletion permission of replica. The original content is always on the user's device.

The infrastructure of this system is almost available. Public key distribution through Blockchains and DIDs, Service discovery with Decentralized Name Service, Payment with crypto, Membership cards with NFT, and the business model is also available.

But the experience of this system compared to web2 applications shows a large gap. To solve this, we need to answer two key questions:

1. How does the blockchain support massive users? That is the blockchain's scaling problem.
2. In what way should applications be combined with the blockchain? That is the relationship between the blockchain and Web3 applications.

What do you think about these two questions?