4 ms·

It's younger than most technology is when it becomes successful. Much younger when you realise you think about the age of the concepts, designs, algorithms, etc

It's younger than most technology is when it becomes successful. Much younger when you realise you think about the age of the concepts, designs, algorithms, etc. that eventually become successful rather than a specific attempt at implementation.

JumpCrisscross 9y ago
Imagine if a bunch of investors walked into the Homebrew Computing Club [1] in 1976 and offered to buy anything and everything its members made for $1 million. With the benefit of hindsight, blockchain is the story of a neat database technology trashing its near-term prospects by marketing itself as a currency. [1] https://en.m.wikipedia.org/wiki/Homebrew_Computer_Club https://en.m.wikipedia.org/wiki/Homebrew_Computer_Club
- darawk 9y ago
  Blockchain is a shit database concept. It's use as a currency is one of its best uses. This idea that blockchains are a good database is an absurd myth. Blockchains are useful specifically for doing distributed consensus, and almost nothing else. Currencies are one of the best applications of distributed adversarial consensus. There are others, but currency is pretty near the top of the list.
  - asimpletune 9y ago
    Can you elaborate why you think it’s a terrible DB?
    - simias 9y ago
      Why is it not? It's slow, expensive and wasteful. You could host a regular MySQL on commodity hardware that could handle ten times the rate of transactions of the bitcoin blockchain. You could publish signed dumps publicly for people to replicate and monitor as they see fit. The only interesting feature of bitcoin is the trustless consensus but it's not as useful or revolutionary as the hype would have you believe. In particular it's only working as intended as long as you remain withing the digital world, as soon as meatspace is involved you need trusted third parties and arbiters. "Blochain technology" whatever that is, is a solution in search of a problem.
      - Nuzzerino 9y ago
        I guess you've never heard of CQRS then. Blockchains are not on-demand databases, but rather sources of truth from which derived, queryable databases can be constructed.
        simias 9y ago
        I fail to see the relevance of your comment. >Command Query Responsibility Segregation is a software pattern that divides the system into two distinct parts, an append-optimised command side and a read-optimised query side. My point was about the "append-optimised" part. Of course the blockchain can be arbitrarily fast to query but you can only make about 5 transactions (or "inserts") per second on the bitcoin blockchain on average. Not very impressive as far as DBs are concerned.
        darawk 9y ago
        CQRS is great. But you can implement it very efficiently using standard logging techniques. Blockchains are just a slow, inefficient implementation of CQRS.
    - Frondo 9y ago
      This is a good read: hackernoon.com/ten-years-in-nobody-has-come-up-with-a-use-case-for-blockchain-ee98c180100
      - Nuzzerino 9y ago
        30 Non-Financial Use Cases of Blockchain Technology. https://imgur.com/a/ShRoW https://imgur.com/a/ShRoW
    - SkyMarshal 9y ago
      These cover it pretty well: Why you don’t want a blockchain https://twitter.com/jimmysong/status/964172100054417409 https://twitter.com/jimmysong/status/964172100054417409 Do you need a blockchain? https://news.ycombinator.com/item?id=16315456 https://news.ycombinator.com/item?id=16315456 Avoiding the pointless blockchain project: How to determine if you’ve found a real blockchain use case https://www.multichain.com/blog/2015/11/avoiding-pointless-blockchain-project/ https://www.multichain.com/blog/2015/11/avoiding-pointless-b...
  - Joeboy 9y ago
    But irrespective of its conceptual quality, why should anybody care, outside of compsci? Why doesn't the doubly linked list get to be a household name or a hotly tipped investment buzzword?
    - pishpash 9y ago
      Because money. B-school types saw green everywhere and concluded they should stick their hands in and hype the heck out of it.
    - darawk 9y ago
      Because at a fundamental level, distributed adversarial consensus commoditizes trust. Institutional trust is an enormous bottleneck in our economy. It is a virtuous cycle that ossifies industries, e.g. banking, and consumer finance. Why does it still take longer to send money to someone in another country than it does to mail them a package? Why are credit cards a pull architecture, even though that is pretty obviously incorrect? The reason is that the institutions that created these things are extremely well insulated from competition. They are well insulated from competition, in part, because they have acquired the public's trust, through a long track record of not stealing everyone's money. Solutions to the distributed adversarial consensus problem provide a way to give new businesses access to the same, or even greater, levels of public trust. Now - obviously there are still scams in this space. When I say 'trust', what I mean is, in the sense that two counterparties can transact in a way that doesn't require intermediation to ensure the completion of the transaction. This really is a fundamentally novel and socially significant innovation. It really does reshape the competitive landscape of a number of industries. At least, it has the potential to. But it is also true that there is an enormous amount of greed and hype floating around. And this isn't in any way intended to justify the ICO scams, or the sky high prices of the existing currencies. It's not clear that anyone has yet figured out the right interface and set of practices for actually realizing all these economic gains. Nobody has yet made the iPhone of blockchain. But the basic technical problem - distributed adversarial consensus was indeed solved by Satoshi. And that really does have substantial, positive social implications for the future. And I fear that people here are losing sight of that because of the gyrations of these silly markets.
  - JumpCrisscross 9y ago
    I think this paper [1] made a decent case for permissioned blockchains (i.e. "where the participants are limited to a predefined set") in cases where one must store state, accommodate multiple writers who do not trust each other and does not have access to an always-online trusted third party. The number of cases where the final condition applies is small, but positive. [1] https://eprint.iacr.org/2017/375.pdf https://eprint.iacr.org/2017/375.pdf
    - darawk 9y ago
      > cases where one must store state, accommodate multiple writers who do not trust each other and does not have access to an always-online trusted third party Ya, that's distributed adversarial consensus. Decentralized currencies are an excellent instance of such conditions. They really are a useful and interesting application. That doesn't mean they deserve all this hype in the form of investment from the average person, but if they weren't so hyped, they'd have a serious use-case as a world reserve currency and they'd make a very nice international settlement layer. Because all of the countries in the world would know that the currency was politically neutral. I think that's a pretty sweet/useful property to have. There are definitely a few other cases where adversarial consensus problems exist too. I think decentralized prediction markets are a pretty great application of smart contracts, for instance. But yes, a lot of these things are way way over-hyped and being pushed for things way beyond their useful scope.
      - kirillseva 9y ago
        Right now such currency is USD, with the Fed acting as a trusted third party. That's one of the reasons why other countries are investing in "blockchain technology", whatever that entails
    - icebraining 9y ago
      The Blockchain is a clever combination of two preexisting concepts: distributed merkel trees, and a consensus algorithm. If you remove one of them, you no longer have a Blockchain, you just have one of its preexisting components. "Permissioned blockchains" is like someone thought "what if we took the radios out of cellphones and connected them by cables" and proceeded to call that a brand new invention.

JumpCrisscross 9y ago

Imagine if a bunch of investors walked into the Homebrew Computing Club [1] in 1976 and offered to buy anything and everything its members made for $1 million. With the benefit of hindsight, blockchain is the story of a neat database technology trashing its near-term prospects by marketing itself as a currency. [1] https://en.m.wikipedia.org/wiki/Homebrew_Computer_Club https://en.m.wikipedia.org/wiki/Homebrew_Computer_Club

darawk 9y ago
Blockchain is a shit database concept. It's use as a currency is one of its best uses. This idea that blockchains are a good database is an absurd myth. Blockchains are useful specifically for doing distributed consensus, and almost nothing else. Currencies are one of the best applications of distributed adversarial consensus. There are others, but currency is pretty near the top of the list.
- asimpletune 9y ago
  Can you elaborate why you think it’s a terrible DB?
  - simias 9y ago
    Why is it not? It's slow, expensive and wasteful. You could host a regular MySQL on commodity hardware that could handle ten times the rate of transactions of the bitcoin blockchain. You could publish signed dumps publicly for people to replicate and monitor as they see fit. The only interesting feature of bitcoin is the trustless consensus but it's not as useful or revolutionary as the hype would have you believe. In particular it's only working as intended as long as you remain withing the digital world, as soon as meatspace is involved you need trusted third parties and arbiters. "Blochain technology" whatever that is, is a solution in search of a problem.
    - Nuzzerino 9y ago
      I guess you've never heard of CQRS then. Blockchains are not on-demand databases, but rather sources of truth from which derived, queryable databases can be constructed.
      - simias 9y ago
        I fail to see the relevance of your comment. >Command Query Responsibility Segregation is a software pattern that divides the system into two distinct parts, an append-optimised command side and a read-optimised query side. My point was about the "append-optimised" part. Of course the blockchain can be arbitrarily fast to query but you can only make about 5 transactions (or "inserts") per second on the bitcoin blockchain on average. Not very impressive as far as DBs are concerned.
      - darawk 9y ago
        CQRS is great. But you can implement it very efficiently using standard logging techniques. Blockchains are just a slow, inefficient implementation of CQRS.
  - Frondo 9y ago
    This is a good read: hackernoon.com/ten-years-in-nobody-has-come-up-with-a-use-case-for-blockchain-ee98c180100
    - Nuzzerino 9y ago
      30 Non-Financial Use Cases of Blockchain Technology. https://imgur.com/a/ShRoW https://imgur.com/a/ShRoW
  - SkyMarshal 9y ago
    These cover it pretty well: Why you don’t want a blockchain https://twitter.com/jimmysong/status/964172100054417409 https://twitter.com/jimmysong/status/964172100054417409 Do you need a blockchain? https://news.ycombinator.com/item?id=16315456 https://news.ycombinator.com/item?id=16315456 Avoiding the pointless blockchain project: How to determine if you’ve found a real blockchain use case https://www.multichain.com/blog/2015/11/avoiding-pointless-blockchain-project/ https://www.multichain.com/blog/2015/11/avoiding-pointless-b...
- Joeboy 9y ago
  But irrespective of its conceptual quality, why should anybody care, outside of compsci? Why doesn't the doubly linked list get to be a household name or a hotly tipped investment buzzword?
  - pishpash 9y ago
    Because money. B-school types saw green everywhere and concluded they should stick their hands in and hype the heck out of it.
  - darawk 9y ago
    Because at a fundamental level, distributed adversarial consensus commoditizes trust. Institutional trust is an enormous bottleneck in our economy. It is a virtuous cycle that ossifies industries, e.g. banking, and consumer finance. Why does it still take longer to send money to someone in another country than it does to mail them a package? Why are credit cards a pull architecture, even though that is pretty obviously incorrect? The reason is that the institutions that created these things are extremely well insulated from competition. They are well insulated from competition, in part, because they have acquired the public's trust, through a long track record of not stealing everyone's money. Solutions to the distributed adversarial consensus problem provide a way to give new businesses access to the same, or even greater, levels of public trust. Now - obviously there are still scams in this space. When I say 'trust', what I mean is, in the sense that two counterparties can transact in a way that doesn't require intermediation to ensure the completion of the transaction. This really is a fundamentally novel and socially significant innovation. It really does reshape the competitive landscape of a number of industries. At least, it has the potential to. But it is also true that there is an enormous amount of greed and hype floating around. And this isn't in any way intended to justify the ICO scams, or the sky high prices of the existing currencies. It's not clear that anyone has yet figured out the right interface and set of practices for actually realizing all these economic gains. Nobody has yet made the iPhone of blockchain. But the basic technical problem - distributed adversarial consensus was indeed solved by Satoshi. And that really does have substantial, positive social implications for the future. And I fear that people here are losing sight of that because of the gyrations of these silly markets.
- JumpCrisscross 9y ago
  I think this paper [1] made a decent case for permissioned blockchains (i.e. "where the participants are limited to a predefined set") in cases where one must store state, accommodate multiple writers who do not trust each other and does not have access to an always-online trusted third party. The number of cases where the final condition applies is small, but positive. [1] https://eprint.iacr.org/2017/375.pdf https://eprint.iacr.org/2017/375.pdf
  - darawk 9y ago
    > cases where one must store state, accommodate multiple writers who do not trust each other and does not have access to an always-online trusted third party Ya, that's distributed adversarial consensus. Decentralized currencies are an excellent instance of such conditions. They really are a useful and interesting application. That doesn't mean they deserve all this hype in the form of investment from the average person, but if they weren't so hyped, they'd have a serious use-case as a world reserve currency and they'd make a very nice international settlement layer. Because all of the countries in the world would know that the currency was politically neutral. I think that's a pretty sweet/useful property to have. There are definitely a few other cases where adversarial consensus problems exist too. I think decentralized prediction markets are a pretty great application of smart contracts, for instance. But yes, a lot of these things are way way over-hyped and being pushed for things way beyond their useful scope.
    - kirillseva 9y ago
      Right now such currency is USD, with the Fed acting as a trusted third party. That's one of the reasons why other countries are investing in "blockchain technology", whatever that entails
  - icebraining 9y ago
    The Blockchain is a clever combination of two preexisting concepts: distributed merkel trees, and a consensus algorithm. If you remove one of them, you no longer have a Blockchain, you just have one of its preexisting components. "Permissioned blockchains" is like someone thought "what if we took the radios out of cellphones and connected them by cables" and proceeded to call that a brand new invention.