4 ms·

Are we overestimating how much crypto is involved? I'm not a crypto guy, but as I understand it, the crypto in Bitcoin is relatively simple and well-understood

by neilk 14y ago

Are we overestimating how much crypto is involved? I'm not a crypto guy, but as I understand it, the crypto in Bitcoin is relatively simple and well-understood. The genius idea was how to verify transactions using proof-of-work in a peer system. At least this is what an acquaintance of mine, Paul Bohm, says (http://www.quora.com/Bitcoin/Is-the-cryptocurrency-Bitcoin-a-good-idea/answer/Paul-Bohm http://www.quora.com/Bitcoin/Is-the-cryptocurrency-Bitcoin-a...). Bitcoin lets other people worry about performance and implementation; the protocol just poses a problem. And apparently the first miners, produced by the entity known as Satoshi Nakamoto, were really slow. So maybe one person, or a small group of people, could have made that conceptual breakthrough?

rdl 14y ago
It is way more novel crypto than virtually any other implemented system. Usually you have papers, then protocols, then reference implementations. Bitcoin is a pretty coherent thing, borderline what one great individual would do; more plausibly a small team.
- betterunix 14y ago
  What novel crypto is there in Bitcoin?
  - chm 14y ago
    This thread is full of claims like this... without evidence to back them up. "I know who wrote the protocol, but I can't tell you. Do your homework." "The protocol is years ahead of everything else, but I can't talk about it."
    - rdl 14y ago
      "Novel" or "difficult to implement" doesn't actually mean better or years ahead. Researchers tweak things to be novel when standard tools will work just for the purposes of experimenting sometimes. I've been arguing that 1981-tech is superior to Bitcoin, even though Bitcoin is novel.
  - rdl 14y ago
    The distributed proof of work and ledger system at the core of bitcoin is fundamentally more novel than, say, a new block cipher mode of operation, and less novel than e.g. the concept of asymmetric cryptography. It would certainly be worthy of a PhD or two in theory alone. The practical deployment could get an econ PhD too I think.
    - javert 14y ago
      The distributed proof of work / ledger system is an extremely cool insight, but that's what it is - an insight. Typically, a computer science PhD involves doing a lot of actual _work_ proving things and so on, which hasn't happened here. This isn't to put it down, at all. It's just to say that the two things are kind of orthogonal. No PhD-granting econ department would grant a PhD for this. They'll just keep laughing and nay-saying until bitcoin goes out of fashion, or proves them completely obsolete and out of touch with reality. At which point they'll just keep going on, just as the philosophy departments have despite their proven irrelevance.
javert 14y ago
From what I know, I agree with you. Bitcoin is fundamentally a heuristic to approximately solve the Byantine Generals Problem.
chjj 14y ago
Came here to say this. I really think it may have been one person. It was a brilliant idea, and as you all know, it only takes one person to have an idea, but would it really take that much work to implement? Would it really require an elite highly-funded team of programming geniuses? When the Linux kernel was first posted, a lot of people, including RMS, couldn't believe that one man got a kernel up and running so quickly. Normally, up until that point, only large well-funded companies were thought to be capable of producing robust kernels. Linux happened despite this preconception. Imagine if Linux had been released anonymously. I wonder what rumors would have started.
- neilk 14y ago
  Hang on a minute. Writing toy kernels for an OS is an exercise that grad students do. The other Unix OS for PCs at the time, Minix, was the work of a single individual, Andrew Tanenbaum. There are plenty of other examples. I don't really know what I'm talking about when I talk about crypto or Bitcoin. I was just musing out loud. In fact, don't even read this comment that I'm writing. :) BTW, I think the best comment in this thread so far is by myprasanna, who seems to be saying that the premise of the entire question is wrong. Those of us who came late to Bitcoin marvel at how this brand new idea and an implementation came out of nowhere. But it seems to have had a history on crypto mailing lists, and implementation ideas were even discussed in public. Maybe the idea has a longer gestation than it appears. https://news.ycombinator.com/item?id=5548682 https://news.ycombinator.com/item?id=5548682
  - chjj 14y ago
    I specifically said "robust" kernels for this reason. While I wouldn't call Minix a toy kernel, it definitely was not as robust as Linux after a while into Linux's infancy. I'm guessing it was considered something a single person could probably manage due to its size. I think this actually furthers my point as well. Minix was created by a world-renowned computer scientist. Linux was created by a college student. Did it really require an elite genius computer scientist to make Linux? No, it just required a kid who was interested in programming (as it happens, he was also very smart). Couldn't the same be true of bitcoin?

by neilk 14y ago

rdl 14y ago

It is way more novel crypto than virtually any other implemented system. Usually you have papers, then protocols, then reference implementations. Bitcoin is a pretty coherent thing, borderline what one great individual would do; more plausibly a small team.

betterunix 14y ago
What novel crypto is there in Bitcoin?
- chm 14y ago
  This thread is full of claims like this... without evidence to back them up. "I know who wrote the protocol, but I can't tell you. Do your homework." "The protocol is years ahead of everything else, but I can't talk about it."
  - rdl 14y ago
    "Novel" or "difficult to implement" doesn't actually mean better or years ahead. Researchers tweak things to be novel when standard tools will work just for the purposes of experimenting sometimes. I've been arguing that 1981-tech is superior to Bitcoin, even though Bitcoin is novel.
- rdl 14y ago
  The distributed proof of work and ledger system at the core of bitcoin is fundamentally more novel than, say, a new block cipher mode of operation, and less novel than e.g. the concept of asymmetric cryptography. It would certainly be worthy of a PhD or two in theory alone. The practical deployment could get an econ PhD too I think.
  - javert 14y ago
    The distributed proof of work / ledger system is an extremely cool insight, but that's what it is - an insight. Typically, a computer science PhD involves doing a lot of actual _work_ proving things and so on, which hasn't happened here. This isn't to put it down, at all. It's just to say that the two things are kind of orthogonal. No PhD-granting econ department would grant a PhD for this. They'll just keep laughing and nay-saying until bitcoin goes out of fashion, or proves them completely obsolete and out of touch with reality. At which point they'll just keep going on, just as the philosophy departments have despite their proven irrelevance.

javert 14y ago

From what I know, I agree with you. Bitcoin is fundamentally a heuristic to approximately solve the Byantine Generals Problem.

chjj 14y ago

Came here to say this. I really think it may have been one person. It was a brilliant idea, and as you all know, it only takes one person to have an idea, but would it really take that much work to implement? Would it really require an elite highly-funded team of programming geniuses? When the Linux kernel was first posted, a lot of people, including RMS, couldn't believe that one man got a kernel up and running so quickly. Normally, up until that point, only large well-funded companies were thought to be capable of producing robust kernels. Linux happened despite this preconception. Imagine if Linux had been released anonymously. I wonder what rumors would have started.

neilk 14y ago
Hang on a minute. Writing toy kernels for an OS is an exercise that grad students do. The other Unix OS for PCs at the time, Minix, was the work of a single individual, Andrew Tanenbaum. There are plenty of other examples. I don't really know what I'm talking about when I talk about crypto or Bitcoin. I was just musing out loud. In fact, don't even read this comment that I'm writing. :) BTW, I think the best comment in this thread so far is by myprasanna, who seems to be saying that the premise of the entire question is wrong. Those of us who came late to Bitcoin marvel at how this brand new idea and an implementation came out of nowhere. But it seems to have had a history on crypto mailing lists, and implementation ideas were even discussed in public. Maybe the idea has a longer gestation than it appears. https://news.ycombinator.com/item?id=5548682 https://news.ycombinator.com/item?id=5548682
- chjj 14y ago
  I specifically said "robust" kernels for this reason. While I wouldn't call Minix a toy kernel, it definitely was not as robust as Linux after a while into Linux's infancy. I'm guessing it was considered something a single person could probably manage due to its size. I think this actually furthers my point as well. Minix was created by a world-renowned computer scientist. Linux was created by a college student. Did it really require an elite genius computer scientist to make Linux? No, it just required a kid who was interested in programming (as it happens, he was also very smart). Couldn't the same be true of bitcoin?