Facebook Libra Is Architecturally Unsound • Hacker News Board

Re: Facebook Libra Is Architecturally Unsound

#271

Post by magnamerc » Tue, Nov 05, 2019, 8:07 PM UTC

briatx wrote: post #265

magnamerc wrote: post #130

As soon as blockchain is even mentioned on HN, everyone loses their freaking minds. For some people that tout themselves as intellectuals, they just seem to parrot that 'blockchains are useless' without even taking a cursory look into the technology. It seems to me that a lot of people here fall into the category of ultracrepidarianism.

Or maybe after much research and some testing they conclude (correctly) that "blockchain is useless."

If that were true, then they would come to a different conclusion. There are several examples of useful applications that are live today that would be otherwise impossible without a public programmable blockchain.

Re: Facebook Libra Is Architecturally Unsound

#272

Post by wmf » Tue, Nov 05, 2019, 8:08 PM UTC

robbya wrote: post #97

matthewdgreen wrote: post #56

I don't disagree with the article per se, but I think many technologists are missing the forest for the trees when it comes to the motivations here. Or perhaps they're being charitable and are evaluating Libra on purely on its stated motivations rather than the ulterior motive that Libra can't name out loud. For example, take this quote: "The possibility that a Libra node run by Mastercard or Andressen Horrowitz woul…

> The plan is to build a ledger that no single party (or coalition of parties in a single legal jurisdiction) has the capacity to edit or alter, and to make such alterations so technically challenging that it's beyond the capacity of any single court or legislature to do so. I'm not seeing why they need a new cryptocurrency for this. If you grow Bitcoin (or any other existing cryptocurrency) to be "too big to fail",…

If Libra is integrated into Facebook Messenger as planned it would have more users on day one than Bitcoin has after a decade.

Re: Facebook Libra Is Architecturally Unsound

#273

Post by thesz » Tue, Nov 05, 2019, 8:08 PM UTC

pron wrote: post #37

The answer to the question implied in the article -- why does Libra make such unjustified design decisions -- is simple. Some people have become enamored with blockchain despite it having almost no good use cases, and this certainly isn't one. It seems like a classic example of focusing on the technology rather than on the problem. -- Regardless of the other, far more important sections of this article, I find the se…

>PLT does not have any tools whatsoever to determine which language is more or less suitable and it is not interested in that question.

You are slightly off here. Look at [1], it views over a metric which correlates with language utility and provides 1) result of the metric applied to various languages and 2) reason why some languages can be more of utility (less errors, for example).

[1] http://www.cs.stir.ac.uk/~kjt/techreps/pdf/TR141.pdf

Re: Facebook Libra Is Architecturally Unsound

#274

Post by robbya » Tue, Nov 05, 2019, 8:11 PM UTC

wmf wrote: post #272

robbya wrote: post #97

Earlier quoted context omitted.

> The plan is to build a ledger that no single party (or coalition of parties in a single legal jurisdiction) has the capacity to edit or alter, and to make such alterations so technically challenging that it's beyond the capacity of any single court or legislature to do so. I'm not seeing why they need a new cryptocurrency for this. If you grow Bitcoin (or any other existing cryptocurrency) to be "too big to fail",…

If Libra is integrated into Facebook Messenger as planned it would have more users on day one than Bitcoin has after a decade.

I get that, but if Bitcoin is integrated into Facebook messenger then it would also see a huge boost in users.

Re: Facebook Libra Is Architecturally Unsound

#275

Post by short_sells_poo » Tue, Nov 05, 2019, 8:12 PM UTC

kragen wrote: post #166

rossdavidh wrote: post #116

Earlier quoted context omitted.

Blockchain gives the impression that the primary threat it is designed to defend against, is a government deciding (like Germany between world wars) to inflate their currency for whatever reason. So, not surprisingly, in order to defend a currency against being decreased in value by a central authority, it is vulnerable to all of the threats that a central monetary authority might help you with. Which of these threat…

Right, here in Argentina our currency has lost more than half its value this year, and the US dollar has lost 96% of its value since the end of the gold standard in 1973. I carry a Zimbabwe 100 trillion dollar bill in my wallet to remind people what real hyperinflation is. I don't think that's the main threat Bitcoin is designed to defend against, though; I think there's a whole spectrum of confiscation threats, rang…

> I don't think that's the main threat Bitcoin is designed to defend against, though; I think there's a whole spectrum of confiscation threats, ranging from thieves tunneling into bank vaults (I know a woman here who lost her savings that way), to immigration authorities confiscating jewelry, to pirates, to trumped-up "money laundering" charges. And of course if we're mentioning Germany and World Wars, we must not forget the confiscation of Holocaust victims' entire possessions, including jewelry and fillings after the gas chambers. Bitcoin can't make genocide impossible but maybe at least it can make it unprofitable.

You cannot mean this seriously. Every example you write is worse for bitcoins. If your coins are stolen, that's it, you can pretty much kiss them goodbye. If the government wants to confiscate someone's bitcoins, they will beat the private key out of the holder with a $5 wrench. If WW3 breaks out and the world order collapses, a wholly digital asset that relies on a large network of high upkeep, high-tech infrastructure is surely not a particularly safe bet. You can bet that datacenters will be among the first casualties in such an event, whether outright government confiscation or a denial of service attack by the enemy.

If a bank gets robbed, for most cases it won't impact account holders at all. Even if the bank goes bankrupt, there are various government schemes that will cover account holders up to some limit (e.g. $250k currently for the US). There is none of this for bitcoins. There could be of course, but then the argument is: what do cryptocurrencies actually offer beyond what is already possible with normal currencies?

To the best of my knowledge, the only practical benefit of cryptos is that they completely sidestep KYC/AML regulations and therefore one can easily move money between regions that draw regulatory scrutiny. Countries with at least semi-modern banking infrastructure already have instant transfers so let's not bring that argument up. In any case, that's not a feature uniquely enabled by cryptocurrencies.

The fact that bitcoins are deflationary I consider a bug, not a feature. Perhaps history will prove me wrong, but in my opinion money is not supposed to be an investment asset. People should not hold cash expecting it will appreciate. It should be reinvested as much as possible.

Re: Facebook Libra Is Architecturally Unsound

#276

Post by phlip9 » Tue, Nov 05, 2019, 8:15 PM UTC

kragen wrote: post #257

phlip9 wrote: post #253

Earlier quoted context omitted.

+ FWIW, 51% of the mining power only allows you to double-spend, not arbitrarily modify the existing ledger. + Since Libra uses a HotStuff variant, you need to control 2/3+ of the voting power to violate safety, i.e., double-spend. In exchange, however, you only need to control 1/3+ to halt progress (liveness).

If you can halt progress, you can make any particular person's holdings impossible to spend (if you can identify them), by refusing to include any blocks that include transactions from them. That isn't quite as lucrative as simply confiscating someone's holdings, but it's close; you can demand a ransom of a sufficient fraction of their holdings, perhaps 5% to 50%, depending on human nature.

I wouldn't say it is quite the same threat model, but certainly a powerful attack vector if you happen to control 1/3+ of the voting power. In the unlikely event that someone does compromise 1/3+ of the voting power, the remaining validators can always hard fork to a new quorum, though this is an expensive and highly synchronized affair.

Re: Facebook Libra Is Architecturally Unsound

#277

Post by stale2002 » Tue, Nov 05, 2019, 8:24 PM UTC

arcticbull wrote: post #264

stale2002 wrote: post #228

Earlier quoted context omitted.

> why they think money should be worth more later Because it is my choice? If you want to use an inflationary currency, go ahead. But other people should have the option of opting out, and using a different currency, with different properties

That’s not a reason. Give me a justification for why you think money should be worth more later. Because it would be sweet for you isn’t really compelling. I gave you a reason for inflation and your counter was “because I would be rich” - see if you can make a better case and if not, maybe reconsider your position.

> Give me a justification for why you think money should be worth more later

My justification is freedom of choice. Or, in other words, the reason is "because I want it to, and I am justified in making my own free decisions".

I believe it is people's right to choose which money system that they want, and you should not have the right to prevent other people's voluntary choice to use a different money system.

I am justifying why people should be allowed to make their own decisions about what money system that they use.

Freedom of choice is a valuable principle in and of itself.

You cannot just dismiss this important concept of freedom of choice. It applies to all parts of society.

The only justification I need, is that I do not want people's freedoms to be infrindged upon.

Re: Facebook Libra Is Architecturally Unsound

#278

Post by _57jb » Tue, Nov 05, 2019, 8:28 PM UTC

matthewdgreen wrote: post #56

I don't disagree with the article per se, but I think many technologists are missing the forest for the trees when it comes to the motivations here. Or perhaps they're being charitable and are evaluating Libra on purely on its stated motivations rather than the ulterior motive that Libra can't name out loud. For example, take this quote: "The possibility that a Libra node run by Mastercard or Andressen Horrowitz woul…

Thank you.

Well said and paints a clear lack of understanding from OP on the value this provides.

Re: Facebook Libra Is Architecturally Unsound

#279

Post by the_gastropod » Tue, Nov 05, 2019, 8:29 PM UTC

kragen wrote: post #261

the_gastropod wrote: post #256

Earlier quoted context omitted.

> When do you mean? Right now Bitcoin is US$9400, which is almost exactly half of its all-time high value of US$19891 (in 2017). There are several times it has lost more than half of its value, but I don't remember a time when it has lost 90% of its value. Looks like I overstated it slightly . It lost 85% of its value between its all time high December 17th 2017 ($19,891) and December 16th 2018 ($3,159). The rest of…

Anyone who invests in a stock is taking a significant risk of losing 100% of their investment, a risk on the order of 5% per year. Anyone who invests in gold is used to seeing it fluctuate by a factor of 2 or 3 most years. Yet not only are these popular investments — they're better investments than dollars are over all but the shortest time periods. Anyone who invested a large amount of money in stocks or gold in 197…

You're moving the goalposts. Are we comparing BTC to currency or investments?

People invest in things, generally understanding the risk. Risk is not all equal. BTC is far riskier (more volatile) than virtually all publicly traded stocks or commodities. See again the very recent 85% drop in 1 year.

And let's be clear: BTC is not comparable to a stock. Company stock is valuable because it gives you ownership (and therefore a stake in) the company's profits. Gold is a much more reasonable comparison, but even that has real practical value (e.g., use in electronics, jewelry, dental work, etc.) And funny enough, Gold is worth less today than it was in 1979. So... it's not a great investment. It's just volatile, and "investment" in it is basically a 0-sum game. Just like Bitcoin.

Re: Facebook Libra Is Architecturally Unsound

#280

Post by pron » Tue, Nov 05, 2019, 8:41 PM UTC

michaelmrose wrote: post #267

pron wrote: post #73

Earlier quoted context omitted.

> but your final paragraph seems to imply the formalism isn't key to end quality? That's not what I said. I said that merely creating a verified formal specification of a language is not in itself a good path to increasing correctness, and I wouldn't focus on that as a significant cause for concern given all others. But even when you use formal methods in the development of your software -- and I'm certainly a propon…

>That's not what I said. I said that merely creating a verified formal specification of a language is not in itself a good path to increasing correctness, and I wouldn't focus on that as a significant cause for concern given all others. This seems to be poorly considered. Isn't it easier for an OS and a language to be verified once even at great cost than hoping future billions of lines of developer code already prod…

A formal specification for the programming language doesn't automatically verify all programs written in it; in fact it verifies none, or at most one. It serves two purposes: it defines what each language primitive does so that proofs of programs written in the lang can make use of those definitions, and sometimes it is also used to verify that a compiler indeed preserves the language's semantics, as defined in the spec, when translating it to some other language; in other words, it can be used to verify the compiler.

Focusing on a formal language spec is like saying that a good way to make your software more correct is to work hard to ensure your compiler doesn't have bugs. I'm sure you'll agree there are more important things. As to use by program proofs, often an approximate ad-hoc specification is good enough.

BTW, the article you linked to is about one particular formal method -- deductive proof. There are many others, with varying costs and benefits.