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

Bitcoin's Shared Ledger Technology: Money's New Operating System

forbes.com

Post reply on HN
Print view View original

21–30 of 98 posts

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#21

Post by EGreg »

Why do blockchains need proof of work at all? It seems that this is only to solve the Byzantine Generals problem, not to prevent spam or something like that (which HashCash used).

Given a network where every node eventually communicates with every other node, a vector clock is enough. You can simply have append-only trees! This data structure is ideally suited for distributed applications.

In a smaller network, question about double-spend is simply solved by having a MAJORITY of parties report that they've seen a certain precondition at time A (ie Person X has Y coins) before any subsequent transaction is considered validated.

What is also beautiful about appending to a TREE instead of a ledger is that you can have subsets of the network which care about a particular subtree. You don't need everyone to store everything!

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#22

Post by TheDong »

EGreg wrote: post #21

Why do blockchains need proof of work at all? It seems that this is only to solve the Byzantine Generals problem, not to prevent spam or something like that (which HashCash used). Given a network where every node eventually communicates with every other node, a vector clock is enough. You can simply have append-only trees! This data structure is ideally suited for distributed applications. In a smaller network, quest…

Proof of Work does disincentive spam a little because you can only spam as quickly as the network can incorporate transactions (6tps iirc) due to the slow nature of the network. If the network were more efficient and used vector clocks, you could spam much more efficiently too.

If the majority of parties can decide such things without a proof of work, then a single person can just create a large number of accounts and own the network. You need to avoid the problem of fake accounts somehow. Proof of Work does that and you don't propose a solution.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#23

Post by brighton36 »

TheDong wrote: post #19
natrius wrote: post #5

Earlier quoted context omitted.

People pay Nasdaq to make their trades reliable and safe. Reliable and safe trades can now be done by a swarm of computers running free software, and they only charge a tiny margin over the costs of bandwidth, energy and storage. How is Wall Street supposed to make money now? They're all trying to find out, and they all need to be the first. Financial profits will be much smaller, and the slow movers will have a smal…

> only charge a tiny margin over the costs of bandwidth, energy and storage. The margin charged in terms of energy is massive . The bitcoin network can't handle many TPS because each transaction needs significant computational work to secure it. On the other hand, if I trust my database I can do the tiny work to write it, make sure I've got consensus among my DB nodes it's written, and be done with it. I'd say that b…

That's not even a little true. The transaction requires little/no work whatsoever. The miners require precisely as much work as there is speculative value for bitcoin. If Bitcoin were worth $10,000 per bitcoin - the energy burned per transaction would be 40 times higher. But that is not to say that the transaction itself required this expense.

Mt Gox had as much to do with bitcoin as it did to do with http. I wouldn't argue with you that centralized databases are way more efficient than blockchains (hence why some nodes went down), but you simply don't understand the basics of what's going on with bitcoin.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#24

Post by brighton36 »

TheDong wrote: post #22
EGreg wrote: post #21

Why do blockchains need proof of work at all? It seems that this is only to solve the Byzantine Generals problem, not to prevent spam or something like that (which HashCash used). Given a network where every node eventually communicates with every other node, a vector clock is enough. You can simply have append-only trees! This data structure is ideally suited for distributed applications. In a smaller network, quest…

Proof of Work does disincentive spam a little because you can only spam as quickly as the network can incorporate transactions (6tps iirc) due to the slow nature of the network. If the network were more efficient and used vector clocks, you could spam much more efficiently too. If the majority of parties can decide such things without a proof of work, then a single person can just create a large number of accounts an…

Miner fees mediate spam, not the transaction count. (which btw is likely to be enormously higher than 6tps)

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#25

Post by mwilcox »

kolbe wrote: post #4

One of two things: I either do not understand Wall Street's obsession with blockchainesque technologies, or Wall Street's obsession with it is totally emblematic of their buzzword-obsessed technology-ignorant groupthink that fried the tech industry in the late 90's. I'm seeing companies like this pop up all over the place. Blythe Masters's new venture[1], which made the cover of this month's Bloomberg Business, seems…

It's not actual blockchains they're finding useful. It's the reasonably robust framework that people are building around it, where they get a whole lot for free and just need to optimise that annoying blockchain bit in the middle. Rather than do all of the crypto, security R&D they would need to do by not using 'blockchain technology' they just turn off the mining and get it all for free.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#26

Post by brighton36 »

natrius wrote: post #11
TheDong wrote: post #7

Earlier quoted context omitted.

It gives you a publicly writable database that is only secured by people desiring to throw away large sums of computation in an environmentally unfriendly manner. Writes are also slow (10 mins...) and requires relatively high bandwidth and space to be used securely. There's no reason that AirBnb or Uber would want their database to be public, slower, and broken as soon as people lose interest in splitting up mining.…

Blockchains can be built without 10 minute block times or wasteful proof-of-work for mining. You can just pick people the users trust to make blocks or use proof-of-stake mining. Blockchains are harder to break than most corporate databases. Reading up on Ethereum will answer most of your questions.

Proof of work is entirely required. Proof of stake is at best a pipe dream - read Andrew Poelstra's writing on the matter.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#27

Post by brighton36 »

TheDong wrote: post #18
natrius wrote: post #11

Earlier quoted context omitted.

Blockchains can be built without 10 minute block times or wasteful proof-of-work for mining. You can just pick people the users trust to make blocks or use proof-of-stake mining. Blockchains are harder to break than most corporate databases. Reading up on Ethereum will answer most of your questions.

If your "blockchain" doesn't have a proof-of-work beyond trusted people, that's hardly different than a company publishing their database full of signed transactions. I'm aware that 10 minutes is a configurable number, but there's a reason it's as high as it is. I'm also familiar with Ethereum, and I still don't see it as better than a company-run database for anything like uber/airbnb/etc with a central company behi…

There's precisely two (and maybe one) reasons where blockchains offer an economic efficiency. Immutability and/or regulatory arbitrage. When you can buy crack on Google and Amazon, then we'll no longer see an efficiency with blockchains.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#28

Post by Kinnard »

kolbe wrote: post #4

One of two things: I either do not understand Wall Street's obsession with blockchainesque technologies, or Wall Street's obsession with it is totally emblematic of their buzzword-obsessed technology-ignorant groupthink that fried the tech industry in the late 90's. I'm seeing companies like this pop up all over the place. Blythe Masters's new venture[1], which made the cover of this month's Bloomberg Business, seems…

>"Bitcoin technologies solved some very important problems in the creation of currencies: double spending, decentralization, cost distribution, &c. These are not problems that need to be solved in securities trading. One company running on one platform doesn't have a double spendng problem. Decentralization is irrelevant. And cost is no issue."

You're thinking of the securities trading of today, which has its roots in the 16th Mercantile Revolution of Europe. Securities are still essentially pieces of paper: stock certificates. If you've started a company(assuming it was a C-Corp) you know this system is shit.

While this is the case, equity markets, for all that's been developed are like rocket ships tethered to the ground.

In the future you will be able to do business globally the way you can communicate globally today. That means not just transacting money, but assets as well. Bitcoin is a system of digital tokens, which can represent anything. For this, you do need the trustless transactions that the blockchain enables. Would you trust a Russian stock market? Russia has very different ethical standards and very different laws. What if you are declared a persona non grata and your assets are to be confiscated? What if Russia and the United States go to war? What about a Nigerian stock exchange? Would it be reliable? Your reluctance to engage in these markets means entrepreneurs there do not have access to capital.

Math and the laws of computation are the same everywhere. This allows for new types and a level of global commerce that will change the world the same way global communication did.

Bitcoin is way, way more than a currency. It allows assets to flow like email. And that is just the beginning.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#29

Post by natrius »

TheDong wrote: post #18
natrius wrote: post #11

Earlier quoted context omitted.

Blockchains can be built without 10 minute block times or wasteful proof-of-work for mining. You can just pick people the users trust to make blocks or use proof-of-stake mining. Blockchains are harder to break than most corporate databases. Reading up on Ethereum will answer most of your questions.

If your "blockchain" doesn't have a proof-of-work beyond trusted people, that's hardly different than a company publishing their database full of signed transactions. I'm aware that 10 minutes is a configurable number, but there's a reason it's as high as it is. I'm also familiar with Ethereum, and I still don't see it as better than a company-run database for anything like uber/airbnb/etc with a central company behi…

> I'm also familiar with Ethereum, and I still don't see it as better than a company-run database for anything like uber/airbnb/etc with a central company behind it.

It's better because no one is taking a cut off of the transactions that the network enables. No one has to approve a new application that uses the network, so every application can build on top of the network. The people own the network, not companies.

This isn't just better, it's better to the degree that the Internet was better than private networks. That difference was world changing, and this one will be too.

> Additionally, if you change your mining to be "people users trust", your security is now equivalent to corporate database security -- a lead pipe to the head to get a password in.

Blockchains have programs that validate transactions. You'd have to get every validator to incorporate a fraudulent transaction. Hacking three financial institutions at the same time, for example, is pretty hard. Ten is even harder.

Re: Bitcoin's Shared Ledger Technology: Money's New Operating System

#30

Post by hippich »

EGreg wrote: post #21

Why do blockchains need proof of work at all? It seems that this is only to solve the Byzantine Generals problem, not to prevent spam or something like that (which HashCash used). Given a network where every node eventually communicates with every other node, a vector clock is enough. You can simply have append-only trees! This data structure is ideally suited for distributed applications. In a smaller network, quest…

blockchain is a tree. ledger is just analogy to make it easier to understand.

hashcash needed to prevent swarm of fake nodes making majority vote.

Post reply on HN