Hacker News Reader

Comment by djl0 | Hacker News Reader

I've never heard of this project, so I can't speak to its implementation, however blockchain to handle DB + access api for relatively slow db operations I think is one of the rare great current blockchain use-cases. If done well it has the potential to be a piece of open-sourced infrastructure. And once it's done once well, it could be replicated to every dmv-like entity (IE almost every permit or sign-off based process) in the world. And the comparison to a state-run database, I think many of us would agree that is an area where there is plenty of room for improvement.

But this is the first I'm hearing of this project, so I'm not advocating for it specifically.

View on HN

> blockchain to handle DB + access api for relatively slow db operations I think is one of the rare great current blockchain use-cases

Ok, sure, it's OK for car registration to be slow and expensive in compute terms.

But what benefit does blockchain have here? What could blockchain possibly add that a private database maintained by the state with otherwise equivalent functionality couldn't do better?

This is a textbook situation where there is a trusted third party. In fact, the state is the party whose opinion controls in this situation. It's not like people are registering cars for fun, the are doing it because it's a government requirement.

It might remove the need for the state to be involved at all, other than for litigation purposes. You’re basically saying things are fine already, which a lot of people disagree with.

In this case, I'm saying that the state can leverage the blockchain for infrastructure and DB access/permission/security. I haven't worked in this specific space, but having a state-affiliated key signing off on a blob of data and store that on the blockchain (and using a standardized open-source front-end for state-side or citizen-side view and editing) has potential to be a much more elegant and robust solution than every state agency around the world creating it's own front and back-end.

I agree that in this case it's not about abstracting trust from centralized authority.

What happens in the case of fraud or illegal activity where someone steals someone else's wallet/keys and is able to transfer the "title"?

If there isn't a really good answer to this question, then how can you seriously say that using blockchain could be a good use case?

1. At initial rollout, I don't imagine end users would be in charge of their own keys. State-managed keys un-does some of the benefit of being on the blockchain at all, however it sets up a modularity for the future where people could choose to take their own keys.

2. The blockchain can still have similar checks for fraud. IE I'd still imagine the DMV-approved key needs to sign-off on whatever transactions normally pass through the dmv, and perhaps with in-person paperwork. Some crypto maximalists imagine a world where cars and houses are on the blockchain and there are smart-keys or something that prove ownership. That doesn't seem realistic or desirable, however smartcontracts that mimic our current checks could be a big improvement over our current state administration.

So the solution is "I don't know yet, so we'll have the state be in control of the keys"

I'm sure you don't need me to say why this is an unacceptable solution. Yet again, blockchain is found to be a solution in search of a problem.

Aren't CRDTs more preferable for most open source projects such as databases for ML training, weights, etc? I think blockchains have some utility in that scope (perhaps at a meta level rather than all transactions). I would much prefer using CRDTs with ipfs/torrent (the best systems for downloading ML data) for any database rather than any centralized system.

Why would a blockchain be an improvement over a regular centrally controlled database managed by the state? What benefit is it providing?

Check out a comment i just made to a sibling comment.

Reply on news.ycombinator.com