441–450 of 577 posts
These incidents really illustrate the main flaw of smart contracts: a single bug in your code can lead to incredible losses. I simply don't think it's possible for human beings to write good enough software for smart contracts.
I've seen them described as "self-funding bug bounties".
Earlier quoted context omitted.
Because this is the kind of story that cryptocurrency detractors love, and that aligns with the general sentiment towards cryptocurrencies on HN.
Odd: HN is made up of the people closest to cutting-edge technology on the planet, and yet sentiment is negative on cryptocurrency. Hmmm...
Earlier quoted context omitted.
Formal methods are about matching the spec. Errors in the spec are still an issue. Moreover, there is very, very little formally verified code in the world, much less than you'd believe. My former professor (RIP) oversaw the formal verification of the F-16 computer software and it still had significant bugs in the end where the specification itself was incomplete or in error. And that was a multi-year, team-scale eff…
A nice practical example from TLS 1.3: TLS 1.3 has been formally proven (not an implementation, but the standard itself). To the extent the mathematicians correctly explained what the TLS 1.3 RFC says, and correctly told the machine what TLS 1.3 is supposed to do, the machine proof says this protocol does what we intended. That work assumes a bunch of components are black boxes, they must work. If we ever lose confid…
Why not just have a version 1 smart contract, that inside of it has an amount check that, if passed, rewrites the smart contract (or issues a new one) - version 2 - that rectifies the problem? So people who "agree" with the change can just insta-exchange their cash, and those who don't can keep their cash on version 1. Then you don't have bugs like this as easily anymore. This whole thing will no go very far without…
In fiat world, if big industry or finance makes a mistake, the taxpayers are forced to bail them out instead of allowing them to fail, because of some perceived “systemic risk” that could take down the entire national economy.
Earlier quoted context omitted.
That’s not even close to the main flaw. The main flaw is that it incentivizes crime, human trafficking and accelerates environmental destruction. None of these are huge problems if you can afford a private bunker in New Zealand. It’s a major problem for ordinary citizens, especially those of poorer countries. Everything involving code has bugs. Bugs aren’t a reason not to use code. Bitcoin isn’t a problem for the ver…
From a humanitarian and economical point of view, Bitcoin doesn't hurt the poor. It allows the unbanked to finally store wealth, as a counter to a local strongly inflationary currency. Besides storing wealth, also growing wealth, as pretty much the only option they have. Unless you don't believe the narrative that Bitcoin stores/grows wealth very well, in that case...poor people can simply not use it. Perhaps you mea…
Earlier quoted context omitted.
The problem here is really Ethereum and the language Solidity. Chuck full of gotchas and completely horrible to work with. The mechanics of stable coins is really trivial and thusly should be trivial to implement and validate. Unfortunately it is not, and it is sad that the hype and investment rush inside the broader crypto field does not foster languages and platforms that actually are suitable for the few realworld…
EOS tried to use on-chain Ricardian Contracts and pair that with upgradeable smart contracts in C++ with DPoS consensus so that intent-of-code could be enforced. That's pretty "real world" but most people feel it's not unstoppable enough. Ethereum is pretty unstoppable, but unless you use a contract proxy, your contracts are immutable and bugs that can be exploited will be exploited.
Earlier quoted context omitted.
Time for LawyerCoin (TM). The smart-contract is written in English (or other human language) and the oracles are powered by a bunch of lawyers who pinky swear to enforce them. I'm doing ICO next week, if anyone wants in.
We can do better than pinky swearing. Each lawyer gets mining fees when they enforce a contract fairly. "Fairly" is defined by consensus, where each lawyer is forced to predict the consensus before voting in a private ballot that is only revealed when the rewards are issued. Lawyers who predict the consensus get rewarded. If you start it off with people who are trying to be fair then predicting fairness will lead to…
In this design, determining the threshold for similarity may require the participation of a third-party "oracle" or "judge" component.
These incidents really illustrate the main flaw of smart contracts: a single bug in your code can lead to incredible losses. I simply don't think it's possible for human beings to write good enough software for smart contracts.
Would an "upgradeable" crypto-coin solve that flaw?
Earlier quoted context omitted.
It’s more regulated though. That’s the point is there’s some control in usd, for better and worse.
Cash is the ultimate anonymous method of payment.