Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

blog.smartdec.net

Post reply on HN
Print view View original

331–340 of 415 posts

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#331

Post by naasking »

warkdarrior wrote: post #305
naasking wrote: post #266

Earlier quoted context omitted.

That's why smart contract languages should be simple and analyzable by third-party tools and provide statically verifiable guarantees that are readable by ordinary humans. That's also why current smart contract languages are largely failures.

And how do you trust those third-party tools? How do you know they are not colluding with one smart-contract writer to steal your money?

Three points:

1. Smaller trusted computing base. That's the whole reason I'm suggesting crypo might provide a more reliable foundation than a layered social structure.

2. Open source tools will exist.

3. Verification is more feasible now than ever. The more money that's involved, the more people will want formal guarantees. The Ether "thefts" of the past few years have highlighted exactly the deficiencies that I've been pointing out in these past few comments. Already some new contract languages have been proposed that are formally verified, simpler and provide meaningful static guarantees, with little meaningful loss in expressiveness. This situation will only improve in the coming years.

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#332

Post by perl4ever »

AnthonyMouse wrote: post #137
gregschlom wrote: post #33

I've said this several times before on this site but will keep repeating it: there's exactly one use case where the blockchain is a superior (and, in fact, the only) solution: when you can't use contracts and the legal system to ensure trust between the parties. In other words, anything illegal. For any other conceivable use case, a database and a contract between parties are a superior solution. Edit: To clarify wha…

> In other words, anything illegal. Not exactly. It's anything for which the legal system is not effective. For example, you can have a perfectly legal contract, but the other party is in a corrupt foreign jurisdiction that would never find in your favor in the event of a breach. Then you can't contract with them because they have no incentive not to breach, without an alternative method of ensuring compliance that d…

"What you're arguing is that a perfect system of government removes the need for any alternative. But we haven't got a perfect system of government."

Yeah, but maybe something like bitcoin is a cancer on an imperfect system.

If you asked a tumor, maybe it would say, well, the infrastructure of the body I'm living in sucks! There are all these use cases that the immune system and the circulatory system and so on don't handle!

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#333

Post by pnw_hazor »

narrator wrote: post #222
gregschlom wrote: post #33

I've said this several times before on this site but will keep repeating it: there's exactly one use case where the blockchain is a superior (and, in fact, the only) solution: when you can't use contracts and the legal system to ensure trust between the parties. In other words, anything illegal. For any other conceivable use case, a database and a contract between parties are a superior solution. Edit: To clarify wha…

There's one other situation, when there is a business where nobody trusts the brokers. The diamond industry is a good example. Where does a diamond come from? Sellers don't want to know when buying from shady brokers and obfuscating or altering the origin is extremely valuable and hard to detect, so brokers have huge incentives to cheat and thus the database can't be centralized.

The international diamond market is run by opaque cartels and criminal governments that prop up a false market.

Perfect for blockchain.

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#334

Post by biot »

skilgarriff wrote: post #209
biot wrote: post #169

Earlier quoted context omitted.

What are you trading the 1 bitcoin for? Given that the typical use case is to purchase some product or service (rather than just transferring bitcoin for fun), how do you secure that aspect of the transaction?

The same way that I secure my transaction when I pay the waiter for my meal in cash by giving it to them when I'm done eating. I've paid for a lot of services in Bitcoin and been paid in Bitcoin. It's generally been a very smooth and enjoyable process.

A successful transaction is when one party pays and the other party delivers the good/service. If one party delivers but the other fails to hold up their end, blockchain technology doesn’t help and the solution is to seek recourse via the courts. You can mitigate this via escrow but that’s no different than non-blockchain transactions.

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#335

Post by naasking »

pnw_hazor wrote: post #330
naasking wrote: post #266

Earlier quoted context omitted.

That's why smart contract languages should be simple and analyzable by third-party tools and provide statically verifiable guarantees that are readable by ordinary humans. That's also why current smart contract languages are largely failures.

Complicated deals (hell even simple deals) require complicated concise language. If you use simple language. You will still have to go to court to argue about what 'what' means in the context of your deal. Legal documents are complicated because we live in a complicated world. And, humans are terrible communicators who have no idea what they want but they somehow "know" what the otherside meant when they agreed to th…

> Complicated deals (hell even simple deals) require complicated concise language. If you use simple language. You will still have to go to court to argue about what 'what' means in the context of your deal.

The lambda calculus, the SK combinator calculus and Turing machines are all very "simple" languages, but suffice to describe any computable function. There is also no ambiguity in their execution. Lack of ambiguity is exactly one of the benefits of smart contracts.

> Legal documents are complicated because we live in a complicated world.

Legal documents are complicated mostly because natural language is ambiguous, thus requiring considerable verbosity. Programming languages, and specifically smart contract languages, can be designed to be unambiguous, and thus avoid one significant complicating factor.

> If it was easy to do, it would be easy to do.

First, no one said it was easy.

Second, computers haven't even been around for a century. The earliest conceptions of smart contracts were in the mid to late 90s IIRC. That's barely 20 years. Are you advancing the argument that if we haven't solved a problem in 20 years which took millennia to solve via social institutions, then it's unsolvable?

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#336

Post by MildlySerious »

The anti-blockchain sentiment seems severely short-sighted when news like Youtube coming up with new ways to push people off its platform are posted within three hours of this submission.[1]

Where do we move the endless amount of unique content to, once that previously trusted party breaks down? Be it greed, bad management, economics that just won't work out or whatever. Are we just going to watch it disappear? Is that the course of action you condone and suggest? Or are we moving to the next monopoly that will ultimately fail and disappear, hoping that time won't be the one where it's all lost?

The amount of data we generate only goes up, and just throwing a dart to pick which megacorp is responsible for not losing or abusing that data is clearly unsustainable.

In my eyes, Blockchain and other decentralized approaches are the way to move forward in terms of owning your own data, and keeping things that are worth preserving from disappearing at the whim of some third party that couldn't care less.

[1] - https://news.ycombinator.com/item?id=19227294

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#338

Post by wmf »

MildlySerious wrote: post #336

The anti-blockchain sentiment seems severely short-sighted when news like Youtube coming up with new ways to push people off its platform are posted within three hours of this submission.[1] Where do we move the endless amount of unique content to, once that previously trusted party breaks down? Be it greed, bad management, economics that just won't work out or whatever. Are we just going to watch it disappear? Is th…

Most of that user-generated content only exists because the megacorps offered to host it for free. Decentralized platforms aren't going to be free so 90% of that stuff just can't be saved.

But speaking of decentralization there are plenty of proposals like personal servers or federated co-ops that aren't blockchain related.

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#339

Post by pnw_hazor »

naasking wrote: post #335
pnw_hazor wrote: post #330

Earlier quoted context omitted.

Complicated deals (hell even simple deals) require complicated concise language. If you use simple language. You will still have to go to court to argue about what 'what' means in the context of your deal. Legal documents are complicated because we live in a complicated world. And, humans are terrible communicators who have no idea what they want but they somehow "know" what the otherside meant when they agreed to th…

> Complicated deals (hell even simple deals) require complicated concise language. If you use simple language. You will still have to go to court to argue about what 'what' means in the context of your deal. The lambda calculus, the SK combinator calculus and Turing machines are all very "simple" languages, but suffice to describe any computable function. There is also no ambiguity in their execution. Lack of ambigui…

You are always going to have disagreements about what the words mean in the context of particular a case.

I can see smart contract tools continuing to evolve to help draft or validate contract language. There are some enterprise contract management systems that do some this already. But, if things go sideways, IRL facts have to be interpreted in view of the contract you will still need lawyers and courts.

Also, not sure how equitable grounds for breach could be accounted for. A person can have the law (contract, statutes, etc.) squarely on her side and still lose a case because of equitable reasons that are not anticipated by the contract or statutes at play.

2 + 2 doesn't always equal 4 in a court of law.

And, if such systems become available, it seems unlikely that contracts will more understandable to layman. Instead of trusting their lawyers or legalese, people will have to trust the code/proofs.

edit: typos 2x

Re: You Do Not Need Blockchain: Popular Use Cases and Why They Do Not Work

#340

Post by simonw »

nootropicat wrote: post #314

The main argument is a logical fallacy: just because some trust is still required doesn't mean there's no value in minimizing it. The land registry argument is the best example of this fallacy >the regulatory authority can make its own record and thus rewrite yours, which means that blockchain doesn’t work. No, the value is precisely the fact that it's the regulatory authority that forcibly changes ownership, and not…

The thing I don't understand about the land registry / stock ownership scenarios is this:

If someone hits me with a rubber hose until I give them my private key, do they get to then keep my house?

Post reply on HN