How Ethereum can be stolen through DNS rebinding

Re:How Ethereum can be stolen through DNS rebinding by kerkeslager on Sunday March 11, 2018 @01:15PM (#16563793)

> Most of the apps we are seeing today don't make any sense anyway. Very much smacks of the dotcom bubble, where people who suspected that the Internet would change the world were absolutely correct, but then the companies and ideas they put their money behind were missing the mark by miles.

This is the crux of the problem. The cryptocurrencies with the highest valuations right now are ones with serious technical and social flaws--people are investing in cryptocurrencies based on hype rather than merit.

Ethereum is the most egregious example: the idea that a JavaScript-like language is what you'd want to write your security-intensive code in is mind-boggling-ly stupid, and after the community showed themselves to be willing to perform a 51% attack on their own blockchain to fix a problem with the DAO, it's not even decentralized any more. So technically you have all the downsides of writing secure software in JavaScript, and socially you have all the problems of centralization AND all the problems of decentralization. The security flaw this thread is about would be a critical vulnerability to any competent security developer, but it's really minor compared to the more fundamental flaws.

But even in competently-developed cryptocurrencies, the problems are still there because people don't understand or value decentralization, which is literally the entire reason Blockchain is worthwhile. The number of people who buy cryptocurrencies and then leave their coins sitting on an exchange is incredible. You're literally donating money to the world's biggest bug bounties when you do that.

[ Parent | on HN ]

Re:How Ethereum can be stolen through DNS rebinding by dorgo on Monday March 12, 2018 @07:31PM (#16572644)

In my opinion, decentralization is not the point of crypto-tech. Maybe it's some minor feature, but not the selling proposition. The major feature is the formalisation of the concept of consensus. Forks are a manifestation of this feature. "a 51% attack on their own blockchain" is a major feature of crypto-tech. In the last century we would call this a revolution. In this century issues like this can be resolved enterily in software. It's an opt-in economy and in some years an opt-in society.

The only reason you think the cash in your portemonai is worth something is that everybody else things the same. As soon as everybody stops believing it, the paper in your wallet can still be used to make fire. Same goes for authorities. Judges, policemen, presidents are what they are because everybody believes it. We have already an informal consensus of who is a policeman and who is not. Most of the time it resolves to whether somebody wears the uniform. Crypto tech promises to make this verifiably. No need for uniforms.

[ Parent | on HN ]

Re:How Ethereum can be stolen through DNS rebinding by kerkeslager on Tuesday March 13, 2018 @07:39PM (#16580896)

> The major feature is the formalisation of the concept of consensus. Forks are a manifestation of this feature. "a 51% attack on their own blockchain" is a major feature of crypto-tech. In the last century we would call this a revolution. In this century issues like this can be resolved enterily in software. It's an opt-in economy and in some years an opt-in society.

Huh? No. Distributed consensus was formalized in 1978 by Leslie Lamport and the first version of Paxos was introduced in 1989. If you're okay with centralization, Paxos or it's successor, Raft, are much faster and more scalable algorithms than blockchain, and have a wide variety of mature implementations.

[ Parent | on HN ]