8 ms·

> The reason for this fundamental conclusion is as follows. All instances of smart contract theft or loss – in fact, the very definition of smart contract theft

by nbadg 10y ago

> The reason for this fundamental conclusion is as follows. All instances of smart contract theft or loss – in fact, the very definition of smart contract theft or loss, is fundamentally about differences between implementation and intent. This is, I think, to its core the fundamental problem with any kind of rigid language in contracts. It's nice to see an acknowledgement of its existence, but disappointing to see no acknowledgement of its severity. We see this all the time, not just in code, but in law and policy as well: an overly-prescriptive law can be just as problematic as an overly-vague one. And imperatively written smart contracts are about as prescriptive as you can get. This is of course in addition to another critical shortcoming for smart contracts: they are capable only of interacting with informational assets (of which "money" is a subset). You can't possibly create a smart contract to purchase a house, because at the end of the day the person that physically occupies the house is in the position of power. Any smart contract enforcement mechanism would necessarily rely upon some outside system, which defeats the whole purpose of the thing in the first place. I have high hopes for cryptocurrencies, in no small part because they inherently deal with only information (again, money is a subset thereof). But I'll be honest: I remain extremely skeptical that smart contracts as we currently understand them will ever see general adoption. It strikes me as an emotionally understandable, but rationally ill-advised, radical departure from hundreds of years of political philosophy and sociology... not in a way that inspires the term "disruption", but rather in a way that evokes "out of the frying pan, and into a boiling pot of lava". I'm all for "a better contracts system", but given that a contract (even one with a non-human entity) is by its very definition a social instrument, I just fail to see a reasonable way forward for what is essentially contractual bytecode.

nemild 10y ago
I agree broadly - and the current DAO was a terrible idea on many fronts, but I'll discuss one quibble in case it starts a broader discussion. > You can't possibly create a smart contract to purchase a house, because at the end of the day the person that physically occupies the house is in the position of power. It's likely that if something like this ever succeeded, there'd be many parties that would provide bridges between real world assets (dollars, a house, entry to an electronic lock) + real world contracts that might bridge to crypto contracts (so a trustee may hold a house under US law for a crypto title owned by others - similar to how a custodial bank holds financial assets on your behalf). You're right that the enforcement mechanism might be decoupled, but that doesn't make it worthless, as the trustee could still use the enforcement mechanism on the real world side. On the cryptocontract side, parts could be enforced like payments. So why might crypto contracts be interesting in this case: - There are too many parties involved and coordination costs are high (crypto contracts that millions of people collectively negotiate together) - The adherents are international, and may prefer a standardized form of law accessible cheaply and easily to all There are clearly an infinite number of risks - but this is just an example of how it might still be beneficial. One other point. Everyone in this recent DAO hack discusses cryptocontracts as set in stone - and the be all, end all once published. Still, you could always inject in a party that adjudicates, but only in extreme scenarios (technically, they might have root access when a few parties with conflicting incentives agreed). You might ask, doesn't this defeat the whole purpose of crypto contracts if third parties are involved - but there still may be value if the third party role is substantially diminished, or if previous forms of negotiation that never happened due to coordination cost now occur. (the canonical Bitcoin example is escrow, where the two parties can verifiably escrow money which in the real world always required a third party actively involved, and a third party in the cryptocurrency world steps in only if there are issues that the two parties can't resolve and cannot take the money itself - 2 of 3 multisig).
- ethbro 10y ago
  - The adherents are international, and may prefer a standardized form of law accessible cheaply and easily to all I think this is undervalued, and probably going to become even more important in the future. Largely because it's something that the current legal framework simply doesn't provide. We have tenuous legal agreements between nation-states (largely of the "we choose to obey now because it's in our best interest" variety). And we have legal systems that apply in a single country. But we really don't have an ironclad anything in the ether (ha!) between those.
  - nickpsecurity 10y ago
    More than you think. One of the selling points for why some organizations, especially financial or legal, advertise they are in specific countries is a strong, consistent enforcement of specific things like banking integrity, contract law, I.P. law, whatever. It's important enough to business sector that it's on the international ratings for what countries are best to do business in. Also, Switzerland makes a killing in the financial sector using this approach. So, picking the best jurisdiction for something like this plus getting all parties to agree to settle disputes according to its laws would be a positive step for such organizations. Knocks out a whole category of risk. These decentralized schemes often pretend they won't need lawyers but they exist in real world for good reasons. Best to know what they can or can't do in at least all the common, criminal situations.
colordrops 10y ago
> You can't possibly create a smart contract to purchase a house, because at the end of the day the person that physically occupies the house is in the position of power I can imagine a future where blockchain-based logic is used to drive robots, bridging the gap with the real world. To me, blockchains seem like a proto-DNA set to evolve as a base for AI; perhaps we are at ground zero for the primordial soup of future digital lifeforms.
- cel1ne 10y ago
  I don't think anything can evolve in a space that wastes that much energy with proof-of-work.
  - colordrops 10y ago
    Your DNA doesn't change very much throughout the course of your life. One a program is on a blockchain, much less energy is expended to maintain it compared with the POW used to create the block..
    - cel1ne 10y ago
      But the definition of evolution is many many changes.
      - colordrops 10y ago
        You should look up the definition of evolution.
        cel1ne 10y ago
        "Evolution is change in the heritable traits of biological populations over successive generations." I claimed that there are many generations needed for programs to evolve, which POW limits because the generation of new contracts is not cheap.
pash 10y ago
As I explained in another thread [0], there is a reasonable argument that so-called "smart contracts" are not legally contracts at all, because they are not intended to give the parties recourse to legal enforcement of their terms. (This is the basic raison d'être of smart contracts—enforcement de machina rather than de jure.) If a judge accepts that argument, then contract law would not apply to smart contracts, and the usual considerations of intent in contract law would not be relevant. Of course the intent underlying the code that implements the smart contract would remain relevant, but unless the "contracting" parties agreed on meta-provisions for arbitrating disagreements about the intent of the code (legal or otherwise), it seems reasonable to accept the default interpretation supported by many members of Ethereum's community: what the code says goes. In short, legal arguments may not be dispositive, because smart contracts may not be legal contracts. In that case, determining the intent behind the code may be a morass best avoided as many assumed it would be, by ignoring it and letting the code decide. (Of course, the ideal form of arbitrating meta-provision is code compiled against a spec enforced by a tool-chain that includes formal verification—both because it is self-enforcing and because it bridges the gap between intent and what the code actually does. But that seems a long way off.) 0. https://news.ycombinator.com/item?id=11932003 https://news.ycombinator.com/item?id=11932003
joosters 10y ago
Any smart contract enforcement mechanism would necessarily rely upon some outside system Absolutely. This drastically limits the scope of smart contracts. You can do all kinds of interesting stuff with the contracts and digital money, like ponzi schemes, simple gambling games, and so on, but you can't link them with the real world without relying on some third party trust. And once you require trust, then you might as well implement the whole thing in the real world. Incidentally, this is also the killer problem with Slock.it's (the DAO creators) other project, their 'digital locks', physical padlocks that would open or close based upon smart contracts. They were fundamentally flawed because it's impossible to prove that a physical lock is tied to the blockchain. There's no proof that it will honour a smart contract, so the whole system again relies on third party trust.
- ethbro 10y ago
  I don't see why this is a fundamental problem. At worst, a simple legal shim of the "third party law firm legally bound to do X if smart contract dictates so" would bridge pretty effectively. Smart hardware locks sounds interesting. I'll have to read up on what they tried -- it sounds like a technically interesting problem.
  - notahacker 10y ago
    It's a pretty fundamental problem for a "trustless", "decentralised" system to require either a significant amount of trust, a third party to fix things or both in most conceivable practical use cases. If you're requiring a third party law firm to ensure offline parts of a smart contract are enforced and mediate disputes, you might as well write the contract in something the law firm actually recognises as a contract.
    - nemild 10y ago
      See my other comment, there are gradients of trust not simply trusted or trustless. A simple example why is that 10 MM parties across the world want to coordinate together for a custom contract - this might be challenging depending solely on the current system (coordination cost, different jurisdictions), but there may be benefits to having elements of both systems involved. Again, this is all speculation - and only time will tell, but I also think it's a debate to be had, rather than suggesting that such a system is worthless if it depends on any real world legal forms.

by nbadg 10y ago

nemild 10y ago

I agree broadly - and the current DAO was a terrible idea on many fronts, but I'll discuss one quibble in case it starts a broader discussion. > You can't possibly create a smart contract to purchase a house, because at the end of the day the person that physically occupies the house is in the position of power. It's likely that if something like this ever succeeded, there'd be many parties that would provide bridges between real world assets (dollars, a house, entry to an electronic lock) + real world contracts that might bridge to crypto contracts (so a trustee may hold a house under US law for a crypto title owned by others - similar to how a custodial bank holds financial assets on your behalf). You're right that the enforcement mechanism might be decoupled, but that doesn't make it worthless, as the trustee could still use the enforcement mechanism on the real world side. On the cryptocontract side, parts could be enforced like payments. So why might crypto contracts be interesting in this case: - There are too many parties involved and coordination costs are high (crypto contracts that millions of people collectively negotiate together) - The adherents are international, and may prefer a standardized form of law accessible cheaply and easily to all There are clearly an infinite number of risks - but this is just an example of how it might still be beneficial. One other point. Everyone in this recent DAO hack discusses cryptocontracts as set in stone - and the be all, end all once published. Still, you could always inject in a party that adjudicates, but only in extreme scenarios (technically, they might have root access when a few parties with conflicting incentives agreed). You might ask, doesn't this defeat the whole purpose of crypto contracts if third parties are involved - but there still may be value if the third party role is substantially diminished, or if previous forms of negotiation that never happened due to coordination cost now occur. (the canonical Bitcoin example is escrow, where the two parties can verifiably escrow money which in the real world always required a third party actively involved, and a third party in the cryptocurrency world steps in only if there are issues that the two parties can't resolve and cannot take the money itself - 2 of 3 multisig).

ethbro 10y ago
- The adherents are international, and may prefer a standardized form of law accessible cheaply and easily to all I think this is undervalued, and probably going to become even more important in the future. Largely because it's something that the current legal framework simply doesn't provide. We have tenuous legal agreements between nation-states (largely of the "we choose to obey now because it's in our best interest" variety). And we have legal systems that apply in a single country. But we really don't have an ironclad anything in the ether (ha!) between those.
- nickpsecurity 10y ago
  More than you think. One of the selling points for why some organizations, especially financial or legal, advertise they are in specific countries is a strong, consistent enforcement of specific things like banking integrity, contract law, I.P. law, whatever. It's important enough to business sector that it's on the international ratings for what countries are best to do business in. Also, Switzerland makes a killing in the financial sector using this approach. So, picking the best jurisdiction for something like this plus getting all parties to agree to settle disputes according to its laws would be a positive step for such organizations. Knocks out a whole category of risk. These decentralized schemes often pretend they won't need lawyers but they exist in real world for good reasons. Best to know what they can or can't do in at least all the common, criminal situations.

colordrops 10y ago

> You can't possibly create a smart contract to purchase a house, because at the end of the day the person that physically occupies the house is in the position of power I can imagine a future where blockchain-based logic is used to drive robots, bridging the gap with the real world. To me, blockchains seem like a proto-DNA set to evolve as a base for AI; perhaps we are at ground zero for the primordial soup of future digital lifeforms.

cel1ne 10y ago
I don't think anything can evolve in a space that wastes that much energy with proof-of-work.
- colordrops 10y ago
  Your DNA doesn't change very much throughout the course of your life. One a program is on a blockchain, much less energy is expended to maintain it compared with the POW used to create the block..
  - cel1ne 10y ago
    But the definition of evolution is many many changes.
    - colordrops 10y ago
      You should look up the definition of evolution.
      - cel1ne 10y ago
        "Evolution is change in the heritable traits of biological populations over successive generations." I claimed that there are many generations needed for programs to evolve, which POW limits because the generation of new contracts is not cheap.

pash 10y ago

As I explained in another thread [0], there is a reasonable argument that so-called "smart contracts" are not legally contracts at all, because they are not intended to give the parties recourse to legal enforcement of their terms. (This is the basic raison d'être of smart contracts—enforcement de machina rather than de jure.) If a judge accepts that argument, then contract law would not apply to smart contracts, and the usual considerations of intent in contract law would not be relevant. Of course the intent underlying the code that implements the smart contract would remain relevant, but unless the "contracting" parties agreed on meta-provisions for arbitrating disagreements about the intent of the code (legal or otherwise), it seems reasonable to accept the default interpretation supported by many members of Ethereum's community: what the code says goes. In short, legal arguments may not be dispositive, because smart contracts may not be legal contracts. In that case, determining the intent behind the code may be a morass best avoided as many assumed it would be, by ignoring it and letting the code decide. (Of course, the ideal form of arbitrating meta-provision is code compiled against a spec enforced by a tool-chain that includes formal verification—both because it is self-enforcing and because it bridges the gap between intent and what the code actually does. But that seems a long way off.) 0. https://news.ycombinator.com/item?id=11932003 https://news.ycombinator.com/item?id=11932003

joosters 10y ago

Any smart contract enforcement mechanism would necessarily rely upon some outside system Absolutely. This drastically limits the scope of smart contracts. You can do all kinds of interesting stuff with the contracts and digital money, like ponzi schemes, simple gambling games, and so on, but you can't link them with the real world without relying on some third party trust. And once you require trust, then you might as well implement the whole thing in the real world. Incidentally, this is also the killer problem with Slock.it's (the DAO creators) other project, their 'digital locks', physical padlocks that would open or close based upon smart contracts. They were fundamentally flawed because it's impossible to prove that a physical lock is tied to the blockchain. There's no proof that it will honour a smart contract, so the whole system again relies on third party trust.

ethbro 10y ago
I don't see why this is a fundamental problem. At worst, a simple legal shim of the "third party law firm legally bound to do X if smart contract dictates so" would bridge pretty effectively. Smart hardware locks sounds interesting. I'll have to read up on what they tried -- it sounds like a technically interesting problem.
- notahacker 10y ago
  It's a pretty fundamental problem for a "trustless", "decentralised" system to require either a significant amount of trust, a third party to fix things or both in most conceivable practical use cases. If you're requiring a third party law firm to ensure offline parts of a smart contract are enforced and mediate disputes, you might as well write the contract in something the law firm actually recognises as a contract.
  - nemild 10y ago
    See my other comment, there are gradients of trust not simply trusted or trustless. A simple example why is that 10 MM parties across the world want to coordinate together for a custom contract - this might be challenging depending solely on the current system (coordination cost, different jurisdictions), but there may be benefits to having elements of both systems involved. Again, this is all speculation - and only time will tell, but I also think it's a debate to be had, rather than suggesting that such a system is worthless if it depends on any real world legal forms.