Hacker News Reader

Comment by bo1024 | Hacker News Reader

I'm not a blockchain "expert", but this seems somewhere between misguided and insane to me. How could this possibly better than a secure web server and http user interface by any metric?

- security: how could this possibly be better than a system under your own IT's control?

- security part 2: introduces a whole new set of vulnerabilities and ways for things to go wrong. e.g. a miswritten "smart" contract.

- speed: how could this be faster than a web interface and a single server?

Blockchain solves one main problem: Decentralization. The reason for decentralization is the lack of a single trusted authority who says e.g. whether a particular transaction has occurred or not. In this setting you still have and need Daimler to be the trusted authority (unless I misunderstand), so the sole raison d'etre for blockchain is gone.

When you add that to all the downsides and security risks ... let me know if there is an easy way to bet against this.

View on HN

The whole point is to cryptographically be able to prove ownership of something to the network. Given that bitcoins, litecoins and ether aren't magically disappearing out of people's wallets I think its pretty safe to say these systems can be designed in ways which are secure (at least secure enough for the marketcaps of each of those cryptocurrencies which is billions).

If I get a piece of paper that says I own some asset there is offices full of people who are dedicated to ensuring that piece of paper properly represents some piece of land, car, security etc. Why would you not want to automate such a system?

Not to mention once you can create such a system the chains of trust can extend beyond just your bond issuance, you can tie actions from other trusted systems into your bond issuing system in a fully automated way.

> If I get a piece of paper that says I own some asset there is offices full of people who are dedicated to ensuring that piece of paper properly represents some piece of land, car, security etc. Why would you not want to automate such a system?

I've been saying this for couple of years now: the most likely "disruption" from blockchain technology is going to be in notarial functions. The concept of coloured coins[0] allows to use a globally available, publicly verifiable, and for all intents and purposes a forgery-immune register to verify contracts.

Cutting through all the [templated] legalese, we can say: "Contract A, between parties K and L, concerning transaction T: T shall be considered done and legally binding once Coloured Coin X has been paid from address [foo] to address [bar], and has been verified by no less than 20 subsequent blocks."

There will be lawyers involved, of course. Someone has to actually enforce the contracts and deal with off-blockchain disputes. But dealing with the actual physical transactions and transfers of ownership could be streamlined.

0: https://en.wikipedia.org/wiki/Colored_coins

The legal side will be one of the last barriers I agree, if a court doesn't recognize the immutability of the system then it can't be used as evidence of anything. If blockchains can't be used as an audit trail for legal means then a lot of the value goes out the window.

I don't quite understand your last paragraph, but for the rest:

> The whole point is to cryptographically be able to prove ownership of something to the network.

I don't think that's the problem that blockchains solve. That problem can be solved without blockchain just by using public/private key cryptography and digital signatures, along with for instance a public ledger. But a public ledger doesn't need to be distributed and subject to consensus protocols, mining dynamics, etc. It can simply be hosted on a website.

It is though. You cant ensure ownership if there is no third party the point that of the blockchain is that it solves that issue by turning the entire network into the notarius. Public/private key cryptography does not solve that.

You're talking about a different problem - not having a trusted third party or authority. I agree that blockchain solves that problem (well, it tries).

As for the last paragraph a good example of this is the closing of a contract from an iot device signal. The device signal may be on the iot chain or chain code instance and the contract logic on another.

The only problem with your theory is how to do that with a turing complete system. I can send signatues for anything but actions which occur based on those signatures you would want a somethingwhich continues the chain of trust. Im not sure of any way to do this how you are describing while still maintaining turing completeness.

  > The whole point is to cryptographically be able to prove
  > ownership of something to the network.

Ownership is not the same thing as possession. Using a blockchain for bonds is analogous to bearer bonds, or even just plain cash. There's a reason why both bearer bonds and large denomination currency bills are no longer issued.

  > Given that bitcoins, litecoins and ether aren't magically
  > disappearing out of people's wallets

They do magically disappear. Both traditional fraudsters (e.g. Mt. Gox.) and criminal hackers take bitcoins all the time. The blockchain doesn't care, of course.

Expect in the next couple of years movies to start featuring Bitcoin heists. I'm sure the scripts are already being passed around at the studios.

I would say the problem they are trying to solve is security settlement between different institutions. I guess right now each institution has their own database of who owns what and it's difficult to transfer ownership between them.

What if there was one or many massive global shared ledgers all these institutions ran nodes for? You could transfer ownership of a security between someone in China and Europe instantly.

I am really confused by this logic.

So it's better to ask companies to invest in implementing "blockchain" within their network instead say upgrading their databases or their infrastructure to the current technology level? Then they could simple have API/Middleware calls between systems which can do this even more efficiently.

Yes cause it allow for a distributed validation or concensus model rather than centralized. This open up for a number of opportunities as you dont need centralized validator

If I woke up and found that my securities had been instantly and irreversibly transferred to China overnight, I would not be happy, at all.

Interesting point. A big problem with blockchain-style ledgers is that the definition of "ownership" is nebulous and far from instant, only increasing in certainty over time. We each might have a different time frame before we're sure a transaction has taken place.

If there's a particular group responsible for paying out the security, then it makes more sense to me for that group to just host a ledger on their site and officially endorse all transactions (which can use the same crypto that blockchains use). This would be instant and transparent.

I don't understand how security trading works, but if there is truly no trusted authority responsible for either paying out the security or enforcing that it gets paid, then I guess I see the possibility for using a blockchain (but probably you also would want cryptocurrency...).

This is what people said about the cloud a few years ago. For good or bad, companies are more likely now to use the cloud that to set up their own server room.

Blockchain is not perfect now but the basic problem of Trust is solved. It's only a matter of time before the bugs are ironed out and people start using it.

Cloud skepticism was never on the level of blockchain skepticism.

Cloud solves a real problem of scalable and efficient resource allocation and provisioning for distributed applications, as you note, in both public and private cloud settings, a problem and concept well-understood in the research community that long predates "cloud" as an enterprisey business buzzword.

And as applied, e.g. to centralized financial institution ledgers, cloud skepticism has born out. Workhorse mainframes are still doing bread-and-butter work in mission-critical settings just fine. Where they are "modernized" it is not always to anything resembling "cloud" configuration.

Blockchain, on the other hand, "solves" the invented problems of cypherpunk, crypto-libertarian fantasy. There is essentially no application where it has met with unqualified success or demonstrated advantage outside of the criminal underworld. Bitcoin did very much help make monetized ransomware a thing, so there's that.

Reply on news.ycombinator.com