4 ms·

Reading this, it seems that the author does not understand Bitcoin, blockchains, BFT consensus, or indeed computer programming. He acts as if the people working

by woah 11y ago

Reading this, it seems that the author does not understand Bitcoin, blockchains, BFT consensus, or indeed computer programming. He acts as if the people working on Bitcoin are some kind of mages with arcane knowledge that nobody else has. They are simply well-versed in a very quirky distributed database.

asciimo 11y ago
He founded the MIT Media Lab Digital Currency Initiative, so he must have a pretty good grasp of the domain (http://joi.ito.com/weblog/2015/04/15/announcing-mit-dci.html http://joi.ito.com/weblog/2015/04/15/announcing-mit-dci.html). He also held stakes in Blockstream, who favors small blocks (http://joi.ito.com/coi.html http://joi.ito.com/coi.html).
- JustinAiken 11y ago
  > He also held stakes in Blockstream, Aaah, that explains it.
wpietri 11y ago
I'm personally bearish on Bitcoin. (It's neat tech but as a product I think it's doomed to always be on the margins.) Despite that, I'd agree with Ito here that there are a relatively small number of people who really have the background and the experience to work on the core code. Distributed systems are hard. Cryptography is hard. Financial systems are hard. Transactional databases are hard. Securing systems that handle real money is hard. Open source development itself isn't easy. The number of engineers who think they can handle all that is way, way larger than the number who actually can. As an example, just consider how many systems get breached because people build their own crypto. And look at how we're still finding critical bugs in core security libraries.
- woah 11y ago
  The thing I have issue with is not his assumption that only a few people can work on Bitcoin, but that only the Bitcoin core team understands BFT consensus and POW.
  - wpietri 11y ago
    Could you point out where he said that? I mainly noticed him saying that few people had the necessary experience to work on the code, not that nobody else could possibly understand some core concepts.
schmichael 11y ago
> some kind of mages with arcane knowledge that nobody else has. They are simply well-versed in a very quirky distributed database. Being well-versed in very quirky distributed databases absolutely makes you a mage with arcane knowledge. Edit to expand: being an expert in bitcoin requires understanding distributed databases, distributed consensus, cryptography, bytecode interpreters, and having some knowledge of economics probably wouldn't hurt. Being an expert in any one of these things is more than enough to get you a great career. Being an expert in all of them operating cooperatively is... pretty arcane knowledge.
- lazaroclapp 11y ago
  To be fair, most bitcoin developers are likely to be experts in one or two of those things and just reasonably conversant in the others. Satoshi's paper was pretty clear on the consensus and the crypto, but sort of hand-waved the network, for example. It is my understanding that there have been significant improvements to the security of bitcoin's P2P network since then...

by woah 11y ago

asciimo 11y ago

He founded the MIT Media Lab Digital Currency Initiative, so he must have a pretty good grasp of the domain (http://joi.ito.com/weblog/2015/04/15/announcing-mit-dci.html http://joi.ito.com/weblog/2015/04/15/announcing-mit-dci.html). He also held stakes in Blockstream, who favors small blocks (http://joi.ito.com/coi.html http://joi.ito.com/coi.html).

JustinAiken 11y ago
> He also held stakes in Blockstream, Aaah, that explains it.

wpietri 11y ago

I'm personally bearish on Bitcoin. (It's neat tech but as a product I think it's doomed to always be on the margins.) Despite that, I'd agree with Ito here that there are a relatively small number of people who really have the background and the experience to work on the core code. Distributed systems are hard. Cryptography is hard. Financial systems are hard. Transactional databases are hard. Securing systems that handle real money is hard. Open source development itself isn't easy. The number of engineers who think they can handle all that is way, way larger than the number who actually can. As an example, just consider how many systems get breached because people build their own crypto. And look at how we're still finding critical bugs in core security libraries.

woah 11y ago
The thing I have issue with is not his assumption that only a few people can work on Bitcoin, but that only the Bitcoin core team understands BFT consensus and POW.
- wpietri 11y ago
  Could you point out where he said that? I mainly noticed him saying that few people had the necessary experience to work on the code, not that nobody else could possibly understand some core concepts.

schmichael 11y ago

> some kind of mages with arcane knowledge that nobody else has. They are simply well-versed in a very quirky distributed database. Being well-versed in very quirky distributed databases absolutely makes you a mage with arcane knowledge. Edit to expand: being an expert in bitcoin requires understanding distributed databases, distributed consensus, cryptography, bytecode interpreters, and having some knowledge of economics probably wouldn't hurt. Being an expert in any one of these things is more than enough to get you a great career. Being an expert in all of them operating cooperatively is... pretty arcane knowledge.

lazaroclapp 11y ago
To be fair, most bitcoin developers are likely to be experts in one or two of those things and just reasonably conversant in the others. Satoshi's paper was pretty clear on the consensus and the crypto, but sort of hand-waved the network, for example. It is my understanding that there have been significant improvements to the security of bitcoin's P2P network since then...