Hacker News Reader

Comment by davesque | Hacker News Reader

> I don't understand the blockchain hype.

> A startup has certified my artwork & placed their verification on the bitcoin blockchain.

> Now art dealers & auctioneers can feel secure that I am the original artist.

> One small problem… I am not Leonardo da Vinci!https://www.verisart.com/works/23f2c64a-08c6-4a42-8013-84ac8... …

Mkay, so the system is working as designed? It seems the issue in this case is that the people behind that startup are dumb. How does that translate to blockchain tech not being useful? If you don't properly verify the information you commit to a blockchain, don't trust the information on the blockchain. Meanwhile, you can rest assured that the information was committed by someone who controls a specific private key. Why would anyone expect anything more?

View on HN

> How does that translate to blockchain tech not being useful?

By proving that without a central authority (centralization) verifying everything going into the blockchain, it is actually useless?

I have seen tons of threads where people like to mention shipping as a good use-case for blockchain. Supposedly it ensures manifest etc can't be doctored. And as this example shows if the data is tainted at source, blockchain is just a database, a rather complex one at that. In that case, it is better to use a traditional and controlled database.

How about if you need some central authority to verify everything going into the blockchain don't use a blockchain and use a database instead?

> By proving that without a central authority (centralization) verifying everything going into the blockchain, it is actually useless?

So then I guess there's no use for cryptographic signing, hash chains, and graph-theoretic consensus protocols? Because that's what you're saying. Your characterization of blockchains as "useless" only makes it seem like you have no idea what it actually is. It's not that complicated. It's just a handful of known and widely-accepted-as-not-being-useless-on-their-own algorithms which have been combined in a creative way.

> In that case, it is better to use a traditional and controlled database.

So then you get...what? You get the same thing but without cryptographic integrity checks. If incorrect data is committed, don't use that data. Or, better yet, commit a revocation of that data. At least you don't have to worry about Larry's shipping company cooking the books after the fact to their own advantage.

> So then I guess there's no use for cryptographic signing, hash chains, and graph-theoretic consensus protocols? Because that's what you're saying.

What's being said is that it's not useful for shipping (and physical asset provenance) applications. The validity of the components is a necessary but not sufficient step in validating the application, and the latter cannot be derived from the former. The usefulness of blockchains for currency does not, in itself, imply their usefulness in shipping.

>At least you don't have to worry about Larry's shipping company cooking the books after the fact to their own advantage.

This is not as much of an advantage as it seems, if Larry can defraud you by going around the blockchain. Your risk is then determined by Larry's inclination to fraud, which is the same in either case. Fixing one hole in a fence doesn't help if the cattle can get out through another that you cannot fix.

What this argument needs is the outline of an actual solution to the problems.

You're technically right. It is not the problem of the Blockchain technology. Once something is persisted in the Blockchain you can (depending on the concrete implementation) be sure that it cannot be changed without the original owner's private key.

I feel however that the whole "Blockchain Startup Universe" always comes up with the same bullshit claims. It usually goes like this: "Here is a difficult problem. To solve that problem we use Blockchain. The End." No one really explains how Blockchain is so much better at solving this than a regular database. But everyone just believes it is.

So far I have not heard a satisfying explanation on how anyone will securely verify that a physical object/subject is indeed this one particular hash on the Blockchain.

You will always need a third party to verify this. And then I wonder: If you need a trusted third party anyway, why do you need a Blockchain in the first place?

> I feel however that the whole "Blockchain Startup Universe" always comes up with the same bullshit claims. It usually goes like this: "Here is a difficult problem. To solve that problem we use Blockchain. The End." No one really explains how Blockchain is so much better at solving this than a regular database. But everyone just believes it is.

You're wrong to claim this. The "whole" blockchain startup universe is not coming up with bullshit claims. It's the bullshit portion of the blockchain startup universe that is doing this. The rest of us understand what we're dealing with and why it's novel. Perhaps you've only encountered bullshit claims about blockchain. Does that mean that hash chains are bullshit? No, we've been using them for years to power git. Oh, then does it mean that distributed conflict resolution schemes are bullshit? Nope, we've been using them for years with replicated databases. So then why do these things suddenly become bullshit when put them together? Beats me. Ask the haters.

> So far I have not heard a satisfying explanation on how anyone will securely verify that a physical object/subject is indeed this one particular hash on the Blockchain.

Why does that need to be done? That's not universally accepted as a problem that every blockchain company is trying to solve.

One of the major points that people like to push to justify the valuations of blockchain is that it will make things that verifying authenticity near free and trustless. Then people make the leap to say that you can basically assign ownership of physical items or IP on the blockchain then you increase the efficiency of all of the actions taken on it.

What they seem to miss is that you need a real authority, or oracle, to issue and enforce the connection. It only works if that authority can be trusted and doesn't lose control of their private keys.

This is an example of someone not to be trusted.

Its a distributed trustless ledger that relies entirely on trust.

That's a very crafty over-simplification. You can choose to interpret "trustless" literally like it's magic, or you can actual work to understand what blockchain advocates mean when they say that. It's "trustless" because you don't have to trust a central authority not to tamper with the data once it's been committed to the ledger, not because there's no need for trust of any kind or at any level.

I have yet to see things like key revocation protocols on the blockchain. Mistakes will inevitably be made somewhere along the line. You can bolster the fragility of being unable to cope with these mistakes with external processes, but that doesn’t change the underlying problem.

The most sensible thing is usually to trust - and pay - a third party to manage the data. The net cost of a blockchain secured by PoW is quite high for a hopefully trustless consensus model.

An immutable ledger has some properties that make it of questionable value to real world applications. The author calls out a number of scenarios that make it impractical for tasks that people keep trying to apply it to.

If a system can't cope with "dumb" users who occasionally make mistakes - I have to question whether the technology works at all.

Any technology which doesn't take into account its potential to be misused is also deficient.

>If a system can't cope with "dumb" users who occasionally make mistakes - I have to question whether the technology works at all.

Which technologies meet this standard?

Any mutable data store or versioned immutable data store that allows either the user or an administrator to manually correct the wrong input.

> immutable data store that allows either the user or an administrator to manually correct the wrong input.

Blockchains are this. What's to prevent a user or administrator from committing a "deletion" of some piece of data on the blockchain and then for users to refer to the most recent commit? That's how immutable data structures work.

Reply on news.ycombinator.com