Live data from Hacker News

Blockchain study finds zero success rate and vendors don't call back

theregister.co.uk

121–130 of 247 posts

Re: Blockchain study finds zero success rate and vendors don't call back

#121
post #118

Earlier quoted context omitted.

> Yes but how do I know that you didn't swap the barcodes or you didn't fiddle with the GPS or the thermometer was working or a million other things? Game theory. Without DLT you make more money by doing these things. With DLT it just becomes the cream skimming problem, where you're just paying money to shuffle assets around but you're actually losing money because the total value of all the assets is still the same.…

The fact that a blockchain can’t always be guaranteed to always represent the ground truth about physical assets is certainly a limitation. But I agree with you. This isn’t about 100% guarantees. It’s about making fraud harder and more expensive, increasing visibility, and streamlining the whole process. I remain something of a skeptic but there’s certainly a lot of interest in and money for things like provenance an…

> Just reducing fraud and otherwise cutting costs could be a big win.

I mean that alone is a several trillion dollar a year improvement that will continue on indefinitely, and that's only one of the many applications for blockchain.

Consider that roughly 75% of American adults of working age are either unemployed or underemployed, and in most countries it's vastly worse than that. The reason most of them are unemployed or underemployed isn't (for the most part) that they're unable or unwilling to work or that they don't have valuable skills, but rather that double entry accounting doesn't enable a broad enough spectrum of economic relationships to capture the long tail of human talent.

When I say this is going to be a multi-quadrillion dollar technology I'm not kidding around. 99% of the stuff you own today wouldn't have been possible without the invention of double entry accounting, which vastly expanded the types of economic relationships that were possible and thereby vastly expanded the types of goods and services that could be produced. DLT isn't a replacement for your MySQL instance, it's a replacement for double entry accounting, and a 10x improvement at that. When people look back 500 years from now they're going to see that the vast majority of the (then) modern world wouldn't have been possible without it.

It's going to largely replace the middle class in the same way that automation has largely replaced the working class, because if you go down the Bureau of Labor Statistics list of jobs it's pretty clear that most middle class American jobs are basically just fulfilling glorified accounting functions.

Even just considering supply chain, as the profitability of creating counterfeit items decreases that doesn't just reduce the percentage of things you buy that are counterfeit, but it also makes it possible to make entirely new classes of (genuine) products that wouldn't have been economically viable previously.

Re: Blockchain study finds zero success rate and vendors don't call back

#122
post #118

Earlier quoted context omitted.

> Yes but how do I know that you didn't swap the barcodes or you didn't fiddle with the GPS or the thermometer was working or a million other things? Game theory. Without DLT you make more money by doing these things. With DLT it just becomes the cream skimming problem, where you're just paying money to shuffle assets around but you're actually losing money because the total value of all the assets is still the same.…

The fact that a blockchain can’t always be guaranteed to always represent the ground truth about physical assets is certainly a limitation. But I agree with you. This isn’t about 100% guarantees. It’s about making fraud harder and more expensive, increasing visibility, and streamlining the whole process. I remain something of a skeptic but there’s certainly a lot of interest in and money for things like provenance an…

> It’s about making fraud harder and more expensive, increasing visibility, and streamlining the whole process

It may cut fraud costs, but there are other ways to do that. Blockchain replicates computing costs relative to centralized data stores, so implementations have non-zero scaling costs.

Re: Blockchain study finds zero success rate and vendors don't call back

#123
post #19

I said this in another thread but I'll say it again here. I am downright shocked at just how much of a failure this technology has turned out to be. How can this many brilliant people work on an area for this long and produce almost nothing of value? I suppose it happens in areas of "pathological science" like infinite energy machines, but those are areas where the tech doesn't work and probably violates laws of phys…

If money is the root of all evil, give technologists and futurists the ability to print money they can swindle muggles into pouring their life savings into as a get rich quick scheme or espouse at a pitch about how life changing it is and how much excessive amounts of money they need to get paid to copy paste half the Bitcoin Core or Etherium repo.

You see pretty fast the dissolution of values in the face of potential fortune. Its an ancient aspect of humanity - the first European explorers in the Americas would slaughter whole countries on the whisper of treasure to be had.

Re: Blockchain study finds zero success rate and vendors don't call back

#124
post #122
post #118

Earlier quoted context omitted.

The fact that a blockchain can’t always be guaranteed to always represent the ground truth about physical assets is certainly a limitation. But I agree with you. This isn’t about 100% guarantees. It’s about making fraud harder and more expensive, increasing visibility, and streamlining the whole process. I remain something of a skeptic but there’s certainly a lot of interest in and money for things like provenance an…

> It’s about making fraud harder and more expensive, increasing visibility, and streamlining the whole process It may cut fraud costs, but there are other ways to do that. Blockchain replicates computing costs relative to centralized data stores, so implementations have non-zero scaling costs.

> Blockchain replicates computing costs.

Which is about as relevant as the cost of tea in china.

Or, to steal another analogy, counting on database efficiency as the competitive advantage for your business is like picking up nickels in front of a steamroller. Especially since DLT efficiency will eventually be figured out. Even if it takes another five or ten years or even twenty years, that's nothing in the grand scheme of things.

Re: Blockchain study finds zero success rate and vendors don't call back

#125
post #8

I've been working on DLT / enterprise blockchain technologies since 2014 and have insight into hundreds of projects, a small number of which made it to live production. Here is the bottom line: yes, it's mostly hype, but this technology does have genuine use cases - when you want to build an interparty database-driven application, and cannot find a suitable place to put the database, because of business concerns or r…

Can you let me know how these "private" blockchains utilize proof of X to determine consensus? E.g. proof of work, or proof of stake, or something else? The thing I never understood with private blockchain tech is that the "traditional" blockchain (i.e. Bitcoin) relies on proof of work, and the only way this is viable is to have tons of resources working on these proofs so that you don't get a 51% attack (there have…

HN discussion about 51% attacks: https://news.ycombinator.com/item?id=17173051

Re: Blockchain study finds zero success rate and vendors don't call back

#126

Earlier quoted context omitted.

> when you want to build an interparty database-driven application, and cannot find a suitable place to put the database, because of business concerns or regulation If you're operating a legal business, a centralized database hosted by an industry mutual or regulator beats a blockchain.

Yes, that's true. But in some cases an industry mutual doesn't exist, and the regulator doesn't want to manage the database. Then what? It can be cheaper and easier to deploy a blockchain than to build the necessary organizational structure to run a central database. Like I said, it's niche but it happens.

> But in some cases an industry mutual doesn't exist, and the regulator doesn't want to manage the database. Then what?

Then you have a people & process problem, not a technical one. The Blockchain, A Real Database™ or even a CSV file won't solve your problems here.

Re: Blockchain study finds zero success rate and vendors don't call back

#127
post #38

Earlier quoted context omitted.

There are a variety of formal consensus algorithms uses for enterprise blockchains, but they are all some variant of voting schemes based on validator signatures. Nothing like proof-of-work is needed to ensure that one bad actor, or a small number of bad actors, cannot break the network's consensus. If you have just one validator, like QLDB, then you're back to a centralized scenario.

The problem is that having a non-BFT consensus is the same that using database nodes and if you use BFT consensus you don't have the performance required for most use cases. BFT-SMaRT supposely achieve 80k tps for a few nodes but when you scale to more nodes the performance degrades substantially. Another issue against the DLT hype is that the blockchain does not provide security where 90% of the whole architecture i…

There might be some constant factor penalty, but we can shard BFT systems much like we do with Paxos etc. See Shasper for example -- it won't have any practical limits on throughput.

Re: Blockchain study finds zero success rate and vendors don't call back

#128

Earlier quoted context omitted.

> the "farm" selling you their romaine might actually be selling you romaine from somewhere else that they've repackaged. Think about it from a game theory perspective. Under the status quo, each party in the supply chain maximizes their profit by lying about where their supplies came from. Whereas with DLT, each party maximizes their profit by being honest about where their supplies came from. As an example, let's s…

Doesn't this require you to trust the source? And if you trust the source, there is no need for a blockchain?

> Doesn't this require you to trust the source?

It allows you to know the source, which enables you to make a data-informed decision about whether or not you trust the source.

As opposed to the current status quo, where you can never even know the source so trust doesn't even come into the picture.

Re: Blockchain study finds zero success rate and vendors don't call back

#129

Blockchains make no sense when the "asset" they are tracking have to be imported from the outside world because they lose the trustless property which is the whole point. For example supply line tracking blockchains. "Look this piece of beef was at place X on time Y with temperature Z, I have a hash and everything". Yes but how do I know that you didn't swap the barcodes or you didn't fiddle with the GPS or the therm…

What we need is an edible blockchain, perhaps based on a proof of steak model.

Re: Blockchain study finds zero success rate and vendors don't call back

#130
post #86
post #63

Earlier quoted context omitted.

How does the blockchain keep brokers from lying?

Maybe you trust the source, but not the broker?

Exactly - if the diamond mines are honest then you don't really need to trust the middlemen. You could also see how many diamonds a certain mine was claiming to produce, and if that number seemed disproportionate to the size of the mine, people might stop treating diamonds coming from that mine.
Post reply on HN