Hacker News Reader

Comment by idlewords | Hacker News Reader

Sure, but a lot of early technologies move laterally in some unexpected direction, and it was reasonable to expect blockchain stuff to do the same. Maybe the original idea could never fly (even in 2008 you didn't have to be a genius to see that evading financial controls would be the chief use of a working cryptocurrency). But likely someone would find a really neat twist on it and find the blockchain equivalent of VisiCalc, and that would happen again and so on.

What's interesting to me is that the original idea was clearly brilliant, but never found any of those lateral steps. It's kind of a string theory for programmers now.

View on HN

There was no reason to believe it would move laterally, for both technical and social reasons.

First, the underpinnings of Bitcoin are ideological, namely in Austrian economics. This means that any change that caused any inflation would be socially resisted.

Second, most of the issues with Bitcoin stem from the lack of a central authority to do things like chargebacks and handle conflict resolution. It is not possible to solve many of the issues of Bitcoin without a central authority, and that was not going to happen for obvious reasons.

Yeah, but when you took away the money part and just looked at the technical idea of a distributed chained ledger with no central point of trust, that was cool. At least, it was cool to me. It seemed like it might go in some interesting other direction.

Bitcoin was a fascinating technical trick, but it's also been a worked example of the difference between "interesting", versus "feasible" or even "a good idea".

As soon as it claimed to be money, then it was all about the money. If you look at early BitcoinTalk, you'll see a pile of scams that are extremely similar to the jargon still used in crypto to this day - the scammers got in very early indeed.

I never quite understood how that might happen. I kept hearing how it was distributed when it isn't. How the blockchain was going to change the world, when not a single example beyond Bitcoin itself could be given. Healthcare was mentioned, but no application... It's right up there with all the other crazy ideas of the last decade or two (Mongo, agile, Google Glass...)

Augmented reality has plenty of potential commercial applications, and Glass is still quietly trying to exploit them.

The problem is that delivering AR with enough fidelity to fool the eye is capital-H Hard. Michael Abrash had a good series of blog posts on this.

It's the big problem with Bitcoin. The USD dollar exists for the US. Bitcoin only exists for itself.

There's one application that I always wondered why it didn't take off as a blockchain based element, a DNS replacement. It's one of the cornerstones of the internet and the web (which to many people is the internet) and the centralized control over it is definitely a drawback and a shortcoming.

ENS has largely replaced DNS in crypto. Most dapps integrate ENS to display human readable names for addresses and it supports arbitrary TXT records so some apps are using it as a shared profile of yourself that you own.

You're probably aware of: https://en.wikipedia.org/wiki/Namecoin

I think they meant, "why didn't Namecoin replace DNS?"

It's all a significant risk to the internet.

It’s because DNS already has a centralized authority, which is necessary to create gTLDs and manage the sale of brand new domains. Once you have an inherent central authority, a lot of the benefits of a blockchain goes away.

> the technical idea of a distributed chained ledger with no central point of trust, that was cool.

So... like git?

The novel thing about blockchains isn’t the Merkle tree, it’s the mechanism that’s used to “secure” the tree against double spends.

And that algorithm is indeed novel! It also has a lot of drawbacks that ultimately makes the whole system not worth the trouble.

I could write a bunch of words to explain why a non-financial pivot isn’t technically possible, but instead I’ll just ask this: “if the blockchain is so useful, why hasn’t anyone done anything with it beyond Ponzi schemes and fraud?”

how is defi a ponzi scheme or fraud?

Technically your git repo is a blockchain. Those are pretty useful.

I hear that darknet markets tend to use cryptocoins. Which while probably mostly illegal, isn't a Ponzi scheme and isn't (necessarily, depending on what's being sold) fraud.

A merkle tree is distinguished from a blockchain (as the distributed part is effectively removed).

Amusingly enough the darknet use of crypto may only be possible because of the Ponzi-beginnings. Criminals are unlikely to accept Pokémons as currency until after they see others doing it.

Git is called a distributed VCS for a reason. And a blockchain is of course a Merkle tree. This part of the basic ideas behind crypto-coins is useful without doubt. But that's not the point.

The actual interesting and novel thing about crypto-coins was solving the "trust problem" of a distributed ledger. But this solution, to be honest, is quite clumsy: PoW. But nobody ever found a better solution until now. Any other proposal is kind of centralized somewhere.

The question that needs to be asked is now "Do we really want, or even need, a fully decentralized 'value moving system'"?

In a world where you could ultimately trust someone else the answer would be clearly "no".

But when there is on thing to learn about humans than it's "never ever trust someone ultimately as you're going to be scammed by those people eventually". Just look at what government do around the globe…

As long as this fact doesn't change (and it won't change until we create a new species of intelligent beings, better than us) there will be room for something like BitCoin. And as those are distributed systems there even doesn't need to be any agreement on that point.

The Bitcoin is the scam. Its "price" grows only due to Ponzi con men involved. By design it is not suited to be used for anything on scale but the Ponzi schemes. It can't be a substitution for money even if it claims to be. Its number of transactions per time unit is very limited. It's not a solution for fast "double spending" resolution as one doesn't know when the consensus can be final. The technology is on another side unsuited for small scale deployment because then the "proof of work" is small enough that it whoever wants can indeed overpower more than 50% of the rest but using more computers, even from the cloud.

And then there is what we do know about what happens in practice:

https://www.cnbc.com/2019/03/22/majority-of-bitcoin-trading-...

https://www.theregister.com/2018/11/30/blockchain_study_find...

"We found a proliferation of press releases, white papers, and persuasively written articles," Burg et al wrote on Thursday. "However, we found no documentation or evidence of the results blockchain was purported to have achieved in these claims. We also did not find lessons learned or practical insights, as are available for other technologies in development."

> Technically your git repo is a blockchain. Those are pretty useful.

Only if you have a wildly incorrect view of what a blockchain is.

Merkle trees are part of a blockchain, but they are not blockchains on their own. What you’re missing here is that blockchains also include something to determine which branch is the correct one, to protect against double spend attacks. This is why “blockchain” gets a different name rather than just being called Merkle trees.

> What you’re missing here is that blockchains also include something to determine which branch is the correct one,

I'm pretty sure that if I posted a fork of Linux somewhere, approximately everyone would know that it's not the correct one.

This is being downvoted, but it hits the point on the head. The differentiating feature of blockchains is removing the reliance on social capital from a technical system. While interesting from a theoretical PoV, it's mostly an answer begging for a question when you consider that social capital is still the driving force in each level of the stack except the technical level.

I saw a guy brag about swinging from altcoin to altcoin making a profit but then talks down Monero because "hard criminals" use it.

I thought that Monero at least served some purpose as a currency even if it goes against the wishes of some countries' governments. The fact that there are people that talk down cryptocurrencies that are actually being used as a currency really put things into perspective.

It's just multi level marketing schemes all the way down.

When it was actually spendable without huge fees, and before any sane person thought of it as an investment vehicle, Bitcoin had a killer app: Poker and probably fair online gambling. That, however, did not fly with certain governments who didn't want to declare crypto a "thing of value," let alone a currency, but needed to in order to control and tax offshore and onshore gaming. Once you can't spend it on entertainment and it lost its function as a medium of quick exchange, it lost most of its reason for existing. That just happened to coincide with the deluge of suckers who wanted to hodl it. Source: I built a startup Bitcoin gaming site in 2011.

https://www.ibm.com/blogs/blockchain/2019/12/blockchain-beyo...

They also didn't actually produce anything that can't be implemented better without "the blockchain". Even with the news like these:

https://theintercept.com/2021/03/24/andrew-cuomo-covid-ibm-b...

"Blockchains are typically public, their contents transparent to anyone with an internet connection, but the one behind Excelsior Pass will be private, meaning only parties sanctioned by IBM will be able to check the contents."

But the blockchain is the protocol for updating the information where other players aren't trusted. As soon as any "party" can be "sanctioned to be able" there's just no need for blockchain at all. It's just selling of snake oil, and yes, IBM is indeed doing it.

Now who'd expect that? The technology company uses the hyped technologies to sell their services and products, even if the said technologies don't make the products any better?

I've read somewhere that banks use blockchain internally. I mean levels of trust can be different within a company. Maybe you don't want anybody to be able to modify the ledger's history.

The usual use case isn't intra-company controls but replacing the clearinghouse / inter-bank settlement layer. This still does seem like a potential actual use for a blockchain, but "we can replace ACH" gets you maybe a couple hundred million at most, and for a lot of work. Easier to just scam people.

If you were a bank or a set of banks and just "don't want anybody to be able to modify the ledger's history" and you know what you are doing, you actually don't want to use blockchain. Because blockchain doesn't mean "nobody can modify the the ledger's history". It means "for it we have to use "proof of work"". But the "proof of work" means "whoever has 51% of the computing power can take over all".

Of course, blockchain promoters would tell you that "there will be something else, comes Really Soon (tm) which won't use "proof of work"."

One can more efficiently use public key cryptography, hashing and signing without the blockchain for any other problem than "implementing bitcoin."

Casino. It's a very successful casino.

Taking away the money part you'll probably want to replace "distributed chained ledger" with "distributed database" or "state-machine replication". There's actually a nice academic literature on this topic of having distributed databases that have strong notions of "commit". This is never the case with stuff that gets called "blockchain". I've yet to encounter something that isn't a variation on "rate-limit the ingest so we have time to replicate operations world-wide" or "create internal monkey-money incentives for agents not to diverge from corretness", both of which are trivial hacks that don't advance anything from a scientific point of view. Not my field tho, so here's the few refs i know.

https://en.wikipedia.org/wiki/Paxos_(computer_science) An old algorithm used for sharding big-data databases.

https://en.wikipedia.org/wiki/Raft_(algorithm) A simplified version of paxos.

http://www.scs.stanford.edu/~dm/home/papers/losa:stellar-ins... The new paper about SCP which allows replication with subjective notions of trust (no central participant-stake list like paxos/raft).

ps: actually i remove "chained" from the idea. If you want to keep the whole authenticated linked-list thing and do it right you'll end up somewhere near https://irmin.org/ (a generalization of git), which is very useful, but doesn't tackle the "strong decentralized commit". You'll probably want to either centrally manage the "lastest-hash(s)" (like git is usually done), or pair it with a state-machine replication protocol. In that last case it may seem useless to use the chained thing if you already have a real replication protocol, but it is much like using public-key crypto for encrypting small ephemeral symmetric keys and encrypting actual data with symmetric crypto: it's an optimization.

There’s a reason why the blockchain algorithms are quite different from Raft et. al.; blockchain systems need to deal with the existence of hostile actors, while most consensus algorithms need to deal only with unreliable actors. The existence of hostile actors changes the requirements a lot, and goes to explain why cryptocurrencies went a different way.

Respectful disagree, i believe you are factually wrong. Wikipedia already lists 2 paxos variants that have byzantine security. Not sure about their implementation, but in production there is also SCP, which is used for real-money transactions by the stellar network. All these actually have proofs of some security and liveliness properties under some byzantine faults. Additionaly, blockchains themselves don't solve this problem properly: they don't even have a notion of "commit", any transaction could be reverted, it is just more and more intractable to do so, thus every property is much weaker and only probabilistic.

>Wikipedia already lists 2 paxos variants that have byzantine security One of those mentioned is Fast byzantine consensus (FBC) [1]. FBC provides an optimal algorithm for a 2 phase commit system w/ byzantine agents. But FBC requires that each correct process sends a message to each other. This is O(n^2) and does not scale. Bitcoin (specifically, others differ) maintain some of the invariants with extremely high probability over time. High enough to be useful. In exchange, bitcoin's network scales much better. Really, it just needs O(n) messages to confirm a transaction.

Of course, PoS like Ethereum's takes much inspiration from byzantine paxos. But I wouldn't call this a solve problem, and crypto has innovated in this space. We can just debate whether or not the innovation has been worthwhile ;).

[1] https://www.cs.utexas.edu/~lorenzo/papers/fab.pdf

You don't need centralization to do chargebacks and conflict mediation. What is lacking is a decentralized version of a justice system.

Please dont mix Austrian economics into the picture. The few "austrians" that took the bait for BTC are either stupid or scammers at heart. I consider myself the student of Austrian economics, i work as a gold dealer, and both me and hundreds of my clients (whom many likewise are "austrians" or libertarians) believe BTC and cryptos are a scam.

> string theory for programmers

This is gold, I'm gonna steal it

We could get really meta with it and have @idlewords sell that phrase/post to you as an NFT.

String theory for programmers is close to my take. I always think of the STTNG episode where the Federation comes up with a weapon to kill the Borg by giving them an impossible but deeply fascinating problem to ponder forever.

Cryptocurrency seems like that for endless numbers of tech nerds. It’s even worse than the impossible shape they were going to use on the Borg since it plugs right into humanity’s biggest cognitive vulnerability: gambling.

Maybe it’s a weapon sent by aliens to make halt our technological progress. 25 years from now cryptocurrency will have soaked up all surplus capital and locked it into a Ponzi and all our best minds will be churning out increasingly Byzantine coin hacks. Then the hypervelocity impactors start arriving…

Reply on news.ycombinator.com