Hacker News new | past | comments | ask | show | jobs | submit login
What is Ethereum? (etherscripter.com)
121 points by mode80 on Aug 1, 2015 | hide | past | favorite | 99 comments

Isn't fuzziness a feature of most legal codes? Treating the law like an absolute immutable contract sounds anathema to a healthy judicial system.

I'm not sure why some technologists have such a fascination with legalism.

I'm not against the idea of digital government and having laws be distributed under a free documentation license with the ability for citizens to discuss and amend laws. A standard version control and issue tracker model sounds much more appropriate, however. Perhaps even just email and diffs, so as to have a barrier to entry.

Ethereum is a currency and enforcement system built for computers. Initially, ambiguity isn't a feature because it's meant for contracts and things which are meant to be precise. Room for ambiguity in ethereum can be added later, once computers are intelligent enough to negotiate ambiguity.

The main feature of this system is removing the need for enforcement. Think about how much goes into maintaining police, bounty hunters, debt collection services, reposession services, prisons. In many people's opinions, if enforcing a contract is possible through software rather than physical force, software is always preferable. It brings its own issues of course and isn't instantly perfect, but the possibility of a world with much less need for violence to enforce the law is worth pursuing.

The biggest advantage of this is that it can be completely automated. Corporations can be programmed to run themselves and negotiate enforcable contracts with each other without any human intervention. You could have a personal digital assistant which interacts with other digital assistants to manage your finances and investments, and everything can be enforced by software. As long as enough people are running free software, everyone can share mutually beneficial enhancements to the algorithms which govern their behavior.

You might not want this kind of thing today instead of a trained financial advisor, but 30 years from now, it's likely that software will be better able to make financial decisions than any unaugmented human. Even more important would be movement of commodities around the world between people who need them. Instead of relying on profit-motivated markets to distribute goods, we could have optimal algorithms for moving resources where they're most needed, and enforcement of these agreements will not depend on the strength of a nation's military or economic power but rather smart contracts which are perfectly enforcable whether they belong to a first or third-world nation.

None of this comes free, and there are real challenges to overcome. People will try to exploit them for personal profit. Software will have bugs and make bad decisions. Those are some of the growing pains we'll need to work through to get to a more fair, less violent world.

You misunderstand why contract enforcement leads to violence. In the real world, contracts among equals are always fungable and rarely enforced with violence. It is different when the contracts are between powerful and the not so powerful. Violence propagates down.

Ethereum, by automating contract enforcement is going to end up being a tool for the powerful against the weak. A tool which brings the cost of lowering contract enforcement will lead to powerful organizations using Contracts more often against the weak.

You're right in general, but if two groups use a common software platform and write the rules of the contract into it, there's less need to build up an enforcement regime.

I'm not sure exactly how you're interpreting contract enforcement leading to violence, so I'll try to explain a little more. Two people make a financial agreement, hire lawyers, and depend on a powerful police force and and penal system to enforce the terms of their contract. This increases the need for armed police, prisons, etc. As more people and institutions are given the right to enact violence to enforce these contracts, people become more desensitized to violence and arrest and more opportunities for abuse present themselves. Abstracting this enforcement into software also distances the enforcement from the culture and removes some of the negative psychological effects on people of carrying out that enforcement.

That's on the smaller scale. On the larger scale, nations feel pressure into increasingly more arms races to keep themselves on an even level. Abstracting enforcement of agreements into software can have the same effect as the introduction of diplomacy into international trade. While it didn't eliminate war, it offered an alternative to straight-out aggression between tribes. Groups that used to stay to themselves and interact with each other only on disputed borders were able to cooperate on a much higher level by establishing representatives inside each others borders.

A smart contract system is a neutral, diplomatic ground for the Internet. Today, the Internet resembles in many ways the world before diplomacy. There is a persistent, escalating cyber-war between nations. People can move in between nations, but there's no real neutral territory that every nation would consider fair. By having a mutually-agreed and verifiably trustworthy software platform to enact contracts, there can be opportunities for new types of deals that are more mutually beneficial.

Will this suddenly eliminate war and hostility? Of course not. Neither did diplomacy. But almost nobody thinks we should go back to a world without any diplomacy, and once we have neutral and fair systems to conduct electronic business, few will want to give it up.

Contracts that are mutually beneficial typically don't need to be enforced because they are not often broken. People who consider themselves equals give each other slack in contracts. The only use case this software fits with is is handling contracts where one party does not trust the other. Typically the more powerful party is at an advantage.

There are two scenarios I see. This software is useful to powerful groups to make contract enforcement cheaper in which case they will essentially force the weaker party into compliance. Or this gives the weaker party an advantage, in which case this software never takes off. Because let's face it, I would never use this with people I like.

    Contracts that are mutually beneficial typically
    don't need to be enforced because they are not
    often broken.
This is a fundamental misunderstanding the purpose of contracts. Contracts are almost by definition mutually beneficial, at least nominally. Otherwise the only reason parties would enter into them is under duress or out of ignorance. And in fact, contracts backed by a legal system are routinely struck down if either of those cases occur.

What you are describing is a situation where neither party has an incentive to defect or default on the contract. And if that is the case then there was no need for a contract in the first place. For example, you and I agree to meet Friday night for dinner. This is a beneficial arrangement for both of us, neither of us need sign anything.

The purpose of contracts is to solve a game-theoretic problem: it is to our mutual benefit to cooperate in some way. But if we cooperate, then one of us can do even better by defecting. For example, it might be mutually beneficial for me to give you $1,000 today, and for you to give me 100 widgets on Thursday. But if I give you $1,000 today, you can do better than the original agreement by keeping the money and the widgets. So the only reason I would ever give you $1,000 is if there was some way to enforce that you give me 100 widgets. That can be by legal means, or technological means. The threat of punishment under a legal system is one traditionally effective way to incentivize you to hold your end of the bargain, and in many ways is at the very heart of what makes modern society work. A technological solution, that avoided law enforcement and judiciary systems, has the potential to be even better because there are obvious negative effects to having a large and over-reaching police force, for example.

TL;DR The enforceability of the contract is the whole point. The only reason you need a contract in the first place is because someone has an incentive to defect.

Thanks. That's exactly it. People might try to counter by saying that specific cases don't work that way, but that's not what game theory predicts. It won't tell you the outcome of a specific instance of a conflict, but it will predict the overall trend when a winning strategy exists. Traditionally, the counter to exploitable strategies was the honor system. You trusted the people you interacted with not to take advantage of you because you believed in their character. While that can work in specific cases, it's not an overall good system for running a large-scale system of social interactions because it invites exploitation. We tend to think of corruption as inevitable, but the degree of corruption in a system is actually somewhat predictable by the incentives for it built into the rules of the system when that system runs at a large scale.

Thanks for some fascinating contributions to this thread.

I see both sides of the debate, and lean toward your side. But I also think both sides are probably wrong to speak as if we will have either one system or the other in the foreseeable future. I think it is more likely that government resources and law enforcement will exist alongside any trustless contract system, acting as a failsafe (or guarantor of exploitation, depending on your POV I suppose). Few contracts would be irreversible, and few holders of contracts would resist reversal, if physical violence from the state were brought to be bear, even if the issue had already been "settled" by software. I believe the reality is that Ethereum-style contracts would increase efficiencies in more routine, less controversial contracts and the state would remain the final recourse in more extreme cases. As some have said, 30 years down the road when machines are more intelligent, all bets are off. But for the foreseeable future, the state isn't going to wither away just yet.

As far as people obfuscating with complex contracts, yes, of course they will do that. But the obvious implication is that a class of techno-jurists will have to arise to help advise users. Again, that's not much different than the current system of clueless folk consulting attorneys, and I imagine both systems would coexist for some time to come.

I definitely agree that smart contracts are going to be implemented as another layer. They'll erode the need for traditional enforcement slowly. And while there's a need for resolution when there's a bad contract, that resolution system doesn't always have to be a person.

Here's a simple theoretical example. People operate under a common, universal law system that defines the most basic rights. The rules of this system are only those which everyone who participates can agree to. For example, everyone agrees not to kill each other. People who do not accept these rules can form alternative societies, but they will not be able to access areas under common law through the use of smart locks and the like which restrict access to public spaces only to those who accept basic terms of behavior.

These rules can be under constant revision because the code that runs the base social platform is publicly audited, and anyone can submit a request to edit it. These revisions can be tested at smaller scales just by running the rules in private spaces. In your home, you can set whatever rules you want, but anyone running software with conflicting preferences will be alerted to the changes when they enter. If your home rules become popular with your friends, they may choose to implement them in their home. If enough people adopt them, the community may choose to implement them in certain spaces, and they can propagate outward.

The basis of this type of system would be mutual consent, and what a person consents to do can be adjusted at any time by their own software. It would be important that a person's personal platform was entirely under their control and only interacted with the outside world through approved protocols. Likely this would require a direct brain connection that could not be hacked because of physical safeguards.

Business could be conducted between communities that operated under entirely different rules through automatic contract negotiation. New ideas could spread from private to public spaces automatically through smart social contracts.

This is just one, very basic example of a society run by smart contracts. There's plenty of ways for these types of systems to be exploited, but that's part of the design process of any system. Particularly difficult is the problem of letting a machine which can be manipulated dictate so much of life.

But people's brains can be hacked already. Look at advertising, religious cults, mob mentality. People have always been vulnerable to manipulation, and we have developed an interconnected financial system where almost anything can be bought, especially power. We can learn from existing flaws as we develop new systems to fix the bugs in our biology and political systems that have made us vulnerable. We will definitely introduce new vulnerabilities. We just have to make sure we have a robust plan for continued development to respond to them.

If you asked Nash what he thought of game theory if he was alive today, he would tell you it have very little to do with how most people interact with each other.

It was said earlier. People get into contracts out of desperation or ignorance. Therefore there are contracts that are not mutually beneficial.

Since the vast majority of the people of this world are desperate or ignorant, ethereum can become a frightening tool.

Imagine this horrific scenario. It is the year 2030. Ethereum contracts can be written where a robot comes and steals your child for slavery automatically if you don't hold your end of the bargain.

Even today people sell themselves to pay off debts. Ethereum would just be another tool creditors can use to enslave people.

Game theory is usually presented as the science of human interaction, but I admit that I have no special knowledge about what Nash might think. I do think that individual interactions are very different from overall trends in the same way that the structure of how two grains of sand interact with each other is different from the structure of a beach as a whole.

I have no doubt that every new technology invites exploitation, and you're right to think about the consequences. That's an essential part of the design process of any technology, especially one that's general useful. That's the reason why we have such complicated software licenses.

But EvilCorp doesn't need smart contracts to send robots to steal your children. It will definitely use them if they're convenient, but they could just as easily go with the traditional human blood on dead trees contract that's ensnared many souls. Smart contracts add very little to their already efficient operation. But the people who can't afford to hire lawyers, bribe the police, and build robot armies will benefit a lot from enforceable contracts that don't require huge resources to enforce for purely digital transactions.

Yes, there are some things that make me uneasy about this attitude of Ethereum being used as a kind of law. The problem is, traditional law has a some room for the idea of something being "reasonable" or not. I fear it would be quite easy for an disreputable dealer to draw up contracts in code in an obfuscated way, perhaps having hidden subversive behaviour, or clauses that behave in a way different from how a brief inspection would suggest.

This is of course possible in standard law too, hiding clauses in impenetrable legalese. However, in standard law (as far as I know), such tricks don't tend to stand up in court as the behaviour is considered "unreasonable".

Ethereum, being self-enforcing, gives no recourse to victims of tricksters. It's all set in stone.

Good observation. There is a continuum in legal language between expressive power in the natural language and the legal formal language. "Absolute immutable contracts" are however very desirable in many cases where people interact with or through machines. Consider that modern computers, in terms of the hardware are extremely deterministic on a macro level, but a lot of the software is fuzzy (UI).

If there is a market for these, why don't they exist already?

Fuzziness is a feature when you benefit and a bug when you get burned. I'd like an option with no ambiguity.

No, it's a defect. It's probably futile attempting to write "perfect" law, and probably for the best that jurists look to the law's intent rather than always trying to apply the law literally. But how can it be a good thing if people honestly do not and cannot know what is and is not legal?

Ambiguity of legal language doesn't mean you cannot figure out if a certain phenomenon is legal or not.

For example, how would one write insurance contracts? Is it possible to write out every single eventuality? No, so the solution is generic language.

There are of course badly worded laws, but it does not mean that everything has to be concretely written out.

This is a terrible description of Ethereum and trivializes what it is.

The interesting parts of a contract are not the deadline and the transfer of funds. Even an escrow service can provide those functions.

The interesting parts of the contract are the conditions of the transfer of funds, and how those are evaluated prior to effectuating the transfer, and how the obligations are enforced after the transfer.

What makes Ethereum interesting is that it creates a public, decentralized infrastructure for standardized outsourcing of third-party verifiers for each conditional precursor of a contract, and each subsequent obligation.

A contract can specify a set of payments and deadlines, as in OPs example, but it can also specify Boolean inputs from trusted third-party institutions that provide various named state inputs into the contract.

For example, you could have an input that is Boolean, which says, "Website transferred from Seller to Buyer: Verified by Verisign: True/False", etc. Those conditions can be part of the verification process. AND even more excitingly, small micro payments can be made part of the contract, to the third-party verifiers, as part of the contract itself.

I am kind of sick of people who don't understand what the value of smart contracts is, messing up other people's perception of what it is. I award OP zero points, and may god have mercy on his soul.

Interesting additional context, thank you. If you know a better introduction to the topic, please link&submit it. From what I've seen in other discussions about Ethereum, people are kind of sick of not finding an understandable description about what it actually does except making the world a better place ;)

Smart contracts can define AND ENFORCE the transfer of smart property without intervention. The world of smart property today is limited to digital currency. But if you "skate where the puck is going" it's not hard to imagine other digital goods being smart-property-enabled and eventually physical things like cars, door-locks etc.

A smart contract CAN incorporate a digital sign-off from third party verifiers. But it's not always necessary nor is it a defining feature. God's mercy notwithstanding.

> outsourcing of third-party verifiers

Since people using the network will be in control of constructing their own contracts, they will have an incentive to draft the contract in such a way as to make sure they'll have control over this "outside verifier"; it could even be same company which it chooses to be the "verifier".

> trusted third-party institutions

The defeats the previous attempt at decentralization, by re-centralizing the information the network needs to do it's job.

This is what the network hinges on to gain information about the outside world (and remember, companies will have an incentive to build the contract in such a way to control these trusted third-party institutions as mentioned above).

(I'm aware of schelling coin/contracts; I'm don't think it will work unless it's made a mandatory part of all contracts for all third-party verifiers. For applications that need data that is publicity verifiable, such as the weather or sports outcomes, it could work, but if the information is not is publicity verifiable, e.g. a transaction amongst private parties, I don't see it. And that's only if you force the use of Schelling coin/contracts, which is usually never the first thing Ethereum folks talk about when pitching this; I'm not sure if there are actually plans to implement that.)

Your counterparty will not agree to an arbitrator that is not neutral, so there is market for third-party verifiers who have a reputation for trustworthiness.

In some cases yes, but if your counterparty is Facebook (or commercial svc) you don't have any negotiating power. Either you accept their terms or don't use their Service

That problem is orthogonal to the discussion and not absent in real contracts either.

I agree, I think the website is failing because it doesn't touch on a key topic, one that you mention: How does Ethereum handle non-blockchain items?

It's easy to see how a blockchain might handle money payments (using its own money system, that is), but far less clear how you can encode contracts for real-world items. Like the example website, for instance. Just because a blockchain entry states 'x' doesn't give it any legal weight, and certainly doesn't transfer real world possessions or ownership.

As you mention, one of the key features is verifiers. Say I am selling a website. We need to ensure that the money transfer happens only once I have control of the site. (and presumably, that the site has been built and is complete). We need a trusted party to check that this is so, and the ethereum contract needs to have this verifier written in to it.

To be honest, half the people working as developers on Ethereum itself fail to bring up this point. I don't know if they didn't think about it, or if they just assume everyone is smart enough to think of it themselves.

Bitcoin transactions are scripts that execute sequentially and are very restrictive. Ethereum instead took bitcoin and replaced that scripting language with a turing-complete language.

Now, consider the conditional operator in the given example.

"If a transaction occurs.." is the condition. ONLY the conditional of transaction types can be verified by the program because it is inherent in the network. For every other conditional, for example, "if team A wins.." the program has to consult a third-party. These are known as 'oracles' in ethereum lingo.

Now since we are trusting a third-party, why can't we simply use bitcoin and even let the decision script be hosted on the third party?

For the program to be able to take decision on a real-life event, only the occurring of a transaction is the type of event it does not need a third-party for. And trusting is all the problem there is.


Bitcoin's script not being turing-complete is a feature because it specializes the use-case. Theoretically, if ethereum allows a turing-complete language to be executed, a code could be written to hack the system itself. Ethereum handles this by executing in a virtual environment.

Thus ethereum is essentially a abstraction on bitcoin and introduction of a new currency token (ether) is unnecessary.


The vision of ethereum is very desirable. Although Ethereum does not make it possible I want such a system to exist. My limited understanding is that for a system to exist, we need an algorithm to be able exchange private keys between two parties without trusting each other.

We need a homomorphic secret sharing but without a dealer. In a Blakley scheme, both parties chose their own planes; we need the intersection of planes, the line, to be our private key and for everyone to be able to know the public key for the same without the private key being known to anyone.

Does anyone know of such an algorithm, because solving this problem will also truly decentralise cryptocurrencies like email.

> Now since we are trusting a third-party, why can't we simply use bitcoin and even let the decision script be hosted on the third party?

One reason which springs to mind immediately: the entity running the script always has complete control, whilst we may choose to grant limited control to third parties. For example, we might define a set of separate, potentially independent, verification conditions, and perform some action when 2/3 of those conditions are met. Manipulating such a contract would require collusion by many of those third-parties; and it would also make contracts more robust against technical problems.

The bits of a cryptocurrency transaction cannot be changed. Its like a signed cheque, can only be read. So since the flag from third party is being used to decide which transaction is relayed, it is as good as handing over the transactions to the third party to relay them. Majority consensus can still be done by using multiple third-parties.

Yes, the problem is about not what happens on one machine, but N machines globally. That is a system that scales linearly and does not use proof-of-work. One needs to solve the Byzantine General Problem in a more general way than Bitcoin does.

The example contract is very vague and seemingly pointless. The stated problem is 'if the buyer pays before 1st April, then they own the website'. Fair enough, but usually you'd just deal with this without any kind of contract in place. We don't need contracts for simple cases of 'buyer pays money, gets purchase.'

A contract would only be interesting in the example if it was structured as an option: The buyer has promised to pay for the item by 1st April, and in return, you have promised that you won't sell the item to someone else before then.

How would this be handled? How would be example contract react if you set 'WEBSITE_ADMIN' to someone else prior to the purchase? Does the example contract mean that WEBSITE_ADMIN cannot be altered by anyone else? What are the access controls to this storage slot WEBSITE_ADMIN? Does the seller lose write-access to it after the deal completes? Does the buyer gain read access? If not, what's to stop anyone changing this value and breaking the deal?

The data for WEBSITE_ADMIN is set inside the Ethereum network. I think the term 'website' is tripping people up here. It's actually better defined as 'site with data' which can be accessed via the Ethereum APIs. If you are interested in playing around with the API calls, there is a Docker image for it. I put together a hosted version of this as well, so you don't have install anything to play with it: https://github.com/giantswarm/swarm-ethereum

Are you saying that Ethereum is a web hosting platform?

Ethereum is not a web hosting platform. It is not similar to Amazon's EC2 or anything like that. Conceptually, it is a single global computer that anyone can access. It's not the same kind of computer you're familiar with, though.

Programs on this computer can't be changed, and nothing can write to the program's storage except for the program itself. You don't have to trust a service provider to maintain their guarantees: it is impossible to break a guarantee.

These characteristics are useful for writing programs that coordinate interactions between humans. They're also useful for letting devices publish and access information about the world long after their manufacturers go out of business.

Entirely new kinds of software can be built to coordinate social and financial systems.

No, it's not a web hosting platform. It's a new compute/storage/network platform though. Primary feature is a consensus mechanism and built in identity management.

I think Ethereum is trying to replace the web (though it may [or may not; not sure] still run HTML/CSS/JS on top).

Here is a framework that allows you to easily try this in practice: https://github.com/iurimatias/embark-framework

The example is pseudo-code and assumes a future with smart-property domain registrars. A better example might have been a smart contract trust which is 100% doable today.

There are other examples under the Samples menu at etherscripter.com. Also see ethereum.org.

"The example is pseudo-code and assumes a future with smart-property domain registrars."

Exactly. That's the problem. There has to be a much larger infrastructure which handles ownership of real-world objects for this to be useful. There are already smart-property domain registrars. MERS [1] and DTC [2] already do that, for mortgages and stocks respectively. But they don't need Etherium.

[1] https://www.mersinc.org/ [2] http://www.dtcc.com/about/businesses-and-subsidiaries/dtc.as...

"You might even imagine a nation defined not by geography but by the rules and benefits of citizenship. Imagine that people of the future could opt-in to the virtual nation that suits them best." . I wonder if we will ever get there . But its eye opening just to think about it.

It may sound good but this doesn't make sense in the real word. Who decides what happens when two persons of two nations have conflicting rights?

Some extreme examples, you may also combine them:


Nation A: Copyright is forever. (Because 1000 years is not enough for Mickey Mouse.)

Nation B: No copyright at all.

Nation C: Only Copyleft software.


Nation M: No taxes

Nation N: Basic Income for everyone


Nation X: Death squads (but only for "bad" people)

Nation Y: No incarcerations, only voluntary "rehab" camps

Nation Z: No incarcerations, only "voluntary" rehab camps


The current solution is that each nation has a territory and tanks. With some simplifications, inside each nation there are clear rules and when there is a conflict between nations the one that has more tanks is right. Perhaps this is not the best method, but a virtual nation world will need a realistic method to solve conflicts.

I have proposed this idea in the past as a solution to the problem of stateless prople. If it ever does happen, I want it to be called the Peoples Republic of Hackistan :-)

I am sure I have brought it up here before, but I will again. Here it goes.

My father is an intellectual property attorney. Among colorful clients was a very active entrepeneur/inventor.

One year he insisted the future was a digital nation. You see how people drool or get excited even with 1/4-of-the-way-there e-Estonia system? He was talking about this years ago, when I was finishing high school and starting college.

My father even drafted letters to the UN on his behalf pitching the notion of offciail recognition, and he was laughed off completely.

Mark my words, if govts still antagonize their people and an alternative of this form exists, a vocal minority will popularize the notion.

As an American, I cannot wait.

How would a distributed "digital nation" (which is, rather conveniently, very poorly defined) exercise force to enforce its "rules" on the physical world?

You'd probably have to limit the use of force to things almost everyone agrees on, like detaining people to prevent imminent harm. For other things, sanctions are probably effective enough. Given universal, stable, cryptographic identities, you as an individual would "enforce" your rules and desired contributions to public goods by consulting your chosen sanction list before trading with someone.

Ah the slippery slope of imminent harm.

I'm not sure I understand why I need a cryptographic identity to boycott certain vendors? I can already do that today.

You need cryptographic identities for automated boycotts. Boycotts aren't very effective today because they're hard to organize, and some vendors have relative monopolies. I think Ethereum enables both effective boycott coordination and monopoly elimination.

Also, these sanctions apply to both sellers and buyers. I think economic isolation will be an effective deterrent to undesirable behavior. It's basically prison, just without force.

Imminent harm may be a slippery slope, but it's really all up to the individuals who support the enforcers. If someone detains someone who shouldn't have been detained, stop funding their activities and sanction them. Fund a different enforcement organization instead.

I can see the boycott coordination angle (except, the offensive vendor can just change its identity).

By what mechanism does Ethereum enable monopoly elimination? (sorry, I don't know too much about Ethereum)

The problem with detainment, collective detainment, sanctions, etc. is that the local conditions when a detainment occurs may be far from the theoretical situation. What if everyone around you believes that you should be detained, but you actually shouldn't (according to the theoretical 'rule')? You just threaten to "stop funding" them and they let you go? What is to stop them from escalating the violence? Nothing.

"I can see the boycott coordination angle (except, the offensive vendor can just change its identity)."

If the vendor changes its identity but the public is still aware of the connection, the new identity goes on the sanction list. If the new identity can't be linked, then it has absolutely no reputation and has to start from scratch to convince people to buy from them. That is an effective deterrent.

"What if everyone around you believes that you should be detained, but you actually shouldn't (according to the theoretical 'rule')?"

There are no "rules". What I'm describing is a society organized via voluntary actions, but by following prescribed rules. If everyone around you thinks you should be detained, you get detained. That's how society works.

Sanctions don't prevent bad behavior, they punish bad behavior as a deterrent. Separately, a police force exists only as long as people fund it, and as long as people approve of their uses of force.

I'm not saying this is all bulletproof, but I think it's going to be fun to explore as we build the foundations for these sorts of systems.

Did he ever published any of that? I'd love to read it.

They are already called corporations.

This is one of the clearest and understandable-without-prior-knowledge descriptions of the project I've seen so far, can anyone familiar with the details weigh in on how accurate it is?

EDIT: not-so-happy sibling comment: https://news.ycombinator.com/item?id=9987351

Presumably smart contracts would very often depend on input external to the system - for example, imagine a will that says if Person A dies, Person B inherits the millions of dollars worth of Ether at a particular address that Person A owns.

Somebody (or some other system) has to be in charge of the input to the Ethereum system to tell it that Person A has died. This seems like a huge single point of failure that doesn't exist with current contracts, if it gets hacked then the money gets transferred irreversibly and that's that.

Trying to build a smart contract with any kind of "real world" dependency like this seems equivalent to using a crypto system with an extra key that the government controls, there's a single point of failure that could compromise the whole system.

While I generally agree with your point, it seems to me like that specific scenario could be solved with a dead man's switch contract. Not ideal, but still better than other cases.

I got the impression that the answer you'd get from the Ethereum team is: at the worst case, this is just as bad/good as the current system - i.e. real world effects have to be mediated by an executing authority, same as a real world contract. They would add (and I agree) that smart contracts still have advantages, even in this case.

You're right, the dead man's switch is interesting here. I wasn't thinking about the ability to reference a fairly precise point in the future via block number instead of an external input to the system.

Don't get me wrong, I'm pretty excited about smart contracts. It just seems like there are a lot of things that we don't think much about currently because our current system is generally pretty forgiving in its fuzziness, but that could have big consequences (and introduce new incentives to act badly) in the world of script-controlled, irreversible contracts.

You can decentralize the reporting of an individual's death using Schelling point consensus, which is what we use to decide Augur's prediction markets on Ethereum. In brief, we distribute a token to a "reporters." At the appropriate time, reporters submit their decisions about events. Reporters who deviate from the consensus lose some of their token balance, which is redistributed to the honest reporters. If an event is ambiguous, reporters report that fact itself instead of guessing at what everyone else will report.

Oracles will be fine for some use cases. Schelling consensus works for others. There are many ways to get real world information into Ethereum, and many ways to reflect how well real world actors adhere to their commitments.

So it's a currency, a contract system, a new form of law, a distributed computing system... I think I need to lie down.

These explainers I keep reading are not helping.

Can anyone tell me whether this is some giant boondoggle circlejerk, or something that is really going to be used widely and perhaps lead to things changing?

Is this going to be widely used? No. It's early technology.

Will a successor of this idea change things? Yes.

Ether is law, but the example contract says nothing about websites, nor has any provisions to enforce the transfer of said site.

I think it's more appropriate to say that ethereum is naive

If we are being fair here neither does a traditional contract. The enforcement of a that contract falls to person (i.e Policeperson, Bailiff).

The benefit of etherscript in theory at least is that the wording of the contract is unambiguous, with the caveat that there may still be some areas where ambiguity is unavoidable. Which means that in the case of a website a digital bailiff could enforce the transfer. This is a clear benefit in some sense at least. How many times have you read about a court case and been bewildered about how someone could possibly be arguing over something that should be clear cut.

For contracts like a website domain transfer this sounds at least initially doable. What etherscript will have trouble handling is establishing what actually happened. It will undoubtedly be very good at specifying what the outcome of a sequence of events should be but as soon as it hits the real world the problem of establishing what that sequence of events was will still require some form of investigation and enforcement mechanism composed of people and thus there will still be an unavoidable amount of ambiguity there.

The debate in law is usually about what the words mean rather than what words appear in the contract. (Of course there are always issues where contracts have been lost, etc)

Yeah what if the website isn't finished? Does the seller still get paid? I really wonder how they plan on solving conditional logic that depends on programmatic sources.

What happens if you want to change the contract? How do you revoke the previous one? I suppose this would have to be done in a manner that involves both parties, some sort of two part signing process.

What if there's a bug in the contract? The contract then becomes ambiguous and you're left with the same problem.

Not without further integrations, but there are proposals for blockchain-based decentral name resolution etc. Nothing you could directly use right now without a lot of effort around it, but I'm glad they are exploring that space.

I have not seen any clear explanation of how these networks of interacting contracts will be verified, so you know they will work the way you intend. Anyone with any knowledge of the formal verification of software written in Turing-complete languages knows that it is difficult, and it is something that not even many programmers know how to do. Until I see a plausible answer to this question, I have to regard the sort of future presented here as a flight of fantasy.

In the samples there is an unbreakable marriage contract where divorce is only possible if both parties agree. Scary.

This will never be unbreakable because no complex system is unbreakable.

The reason contracts are messy is because they're a system that has vulnerabilities. Lawyers are just hackers with different tools and training.

So moving to another system, equally complex or dynamic, will not solve anything in the long run.

A system that must support contracts between humans would be dynamic enough to allow for vulnerabilities that will eventually be discovered.

In fact I find it laughable that these guys are trying to replace a legal system with a computer system. When computer systems get hacked all the time.

Skilled programmers will be hired to write these contracts in a complex and verbose way such that the other party is unable to understand it and may not be able to afford to hire programmers to analyze it. They will then be taken advantage of.

So pretty much the same as the current legal and financial system.

Maybe. Nobody is going to sign a contract he can't understand if he get the same goods and services from somebody else with a contract he can understand.

People do exactly that all the time.

It's very cool that you know exactly what is going to happen. Or, at least, you could simulate exactly what should happen. And, if there was some repository of contracts, you could simulate the effects of the system. Such higher-order possibilities are so opaque most of the time.

But, I'm not sure ambiguity is a bad thing. Or, at least, there is a dose-specific effect. A lot of ambiguity is bad, but so is none. The flexible Goldilocks Zone is the least fragile -- to society that is, not to law.

I'd be a lot more interested to see statistics on adoption, and actual use cases by real people.

Stats: https://stats.ethdev.com/

Transactions go live on Monday, if I remember correctly.

Smart contracts are one of the worst ideas ever conceived.

Nobody actually wants smart contracts, for excellent and obvious reasons. Nerds think they're a good idea because they'll sort out all the annoying grey areas in social human interaction.

Dr. Strangelove is literally about an unstoppable smart contract going wrong.

No business actually wants smart contracts - they always, always want the option of lawyering out of a bad deal. Or stopping it if the government gets pissed off and says it will penalise you to a crisp if you proceed with that smart-arse contract you just put on the blockchain.

No consumer wants smart contracts, for the same reason "no chargebacks" is a loser. You always want a human to be able to fix the problem.

The only use I can think of for smart contracts is for businesses to screw over customers who have no other option, c.f. mandatory arbitration clauses. Note how the standard example of a smart contract is a car that stops working if your payment fails. Or the Internet fails. Unstoppably. Immune to human intercession, circumstances, etc. (Such locked cars already exist, only bought by people who have no other choice.)

This explanation subtly forgets to explain which agent is responsible for executing the contract-programs. Obviously you don't want the contract to be processed (and the money spent) two or more times at its deadline. Who is doing the processing and ensuring the execution only occurs once?

As far as I understand, it's a distributed consensus system. Scripts don't perform their effects in the "real world"; instead, they append values to a blockchain. External systems can inspect the most recent values in the blockchain and act accordingly. This is how bank accounts work, conceptually.

Many programming courses use bank accounts as examples of race conditions in concurrent systems, eg.

    User withdraws £10 from ATM           User receives £1000 salary           Balance
    ---------------------------           --------------------------           -------
    ATM looks up balance (£100)           Bank looks up balance (£100)         £100
    ATM subtracts £10 from value (£90)    Bank adds £1000 to value (£1100)     £100
    ATM dispenses £10 note                Bank sets balance to value (£1100)   £1100
    ATM sets balance to value (£90)                                            £90
This race condition has blown away the user's salary! However, this is a terrible example, since it is nothing like the actual operation of a bank account!

An account is exactly that: an account (log) of what has happened:

    Debits              Credits
    ------              -------
    ATM withdraws £10   Salary adds £1000
The balance is exactly that: the balance of (difference between) the credits and debits. The balance is derived from the account as needed, taking new entries into account as they get added.

Blockchains are a way to keep such accounts, in a consistent way without giving any one entity control over updates.

For example, the article's Web site ownership system might keep an account with entries like this:

    User A transfers example.com to User B
    User B sets example.com IP to
When contracts are executed, they can inspect the latest values on the blockchain and append alterations. For example, if user B sets up a contract to transfer ownership of example.com to user C, one of the pre-conditions is that user B owns the domain. Hence after execution we get:

    User A transfers example.com to User B
    User B sets example.com IP to
    User B transfers example.com to User C
Anyone can execute the same contract on their local copy of the blockchain to get the same result, hence everyone can reach a consensus about what the new blockchain should contain. The contract also can't be run twice in a row, because the resulting blockchain says that User C owns the domain, and hence user B doesn't have permission to transfer it. The same thing happens to any currency which B and C might transfer in exchange for the domain.

To actually use such a value, we can imagine a DNS-like system which resolves names by looking up the latest IP address from the blockchain; in the same way that an ATM displays our current balance by looking at the credits/debits in the account. All of this can also be cached, of course; provided we either trust the cache to provide the right value, or that we don't care enough about the value to verify it ourselves (eg. transactions between third parties which do not affect us).

Looks like someone is trying to bring about the Diamond Age.

Perhaps I'm wrong, but the simplest explanation of Ethereum appears to be that it's a way of specifying and sharing contracts.

I think the whole digital currency thing muddies the waters. People, myself included, hear the word currency and think of something we earn and spend, but then you hear about all the other uses of Ethereum and wonder what that has to do with money.

What I've just realised by reading this article is that contracts are what Ethereum is all about, the currency aspect is just one of the many use cases. Laws are another example of a contract. The base class is contracts, everything else is a subclass on top of that.

What's a contract? (Playing devil's advocate a little)

(edit: for the sake of pedantry, assume I did indeed mean contract)

Assuming you meant contract, I'd say it was a formal agreement between two or more parties. Will that do?

Not really, no.

I actually have a decent understanding of what's involved regarding Ethereum, but assume for a moment that I don't.

Statements like "a formal agreement between two or more parties" are of little use to someone starting from ground zero.

I can try to put it in plain English if you think it'll be confusing to a newcomer.

Before I try, let's explore how Ethereum is described by those who develop it. The home page seems as good a place as any.


Here's what the first paragraph on the homepage describes Ethereum as...

"Ethereum is a decentralized platform that runs smart contracts: applications that run exactly as programmed without any possibility of downtime, censorship, fraud or third party interference."

So even the Ethereum developers describe the core use as contracts. It doesn't strike me as a controversial idea.

How would you describe Ethereum?

It comes down to prior restraint.

With most laws, there is the possibility of breaking them and maybe you have to face the consequences.

With this stuff, you won't be able to use the stuff for what it wasn't designed for. Stronger guarantees. It's like a car that slows down in a construction zone, and is unable to go any faster, instead of cops giving tickets. Like the 90 year old engineer guy behind the Venus project says in his interviews.

That can be useful. Not for everything, but for a whole lot. The current DRM stuff works like this.

For a less instrumental account, my article with Bill Maurer might be of interest: http://kingsreview.co.uk/magazine/blog/2015/06/23/ledgers-an...

It interrogates the intersections between law and contracts as it pertains to the shift from Bitcoin to Ethereum.

Such a system needs integration with subjective variable sources. For example let's say the website also needs to meet some basic code quality level. They could agree to give a third party exclusive access to set some "quality check" variable.

If that's implemented it can be a great way to codify all sorts of legal docs and laws since a lot of legal stuff requires human interpretation at the time of execution.

The event calculus could be used to specify the behavior of the agents instantiated by ethereum smart contracts. This formalized description of the smart contract (e.g. a legal system) could then be used to verify its actual behavior in the wild.

I wonder if it is possible to pass an actual law which would render digital 'contracts' illegal. So all lawyers become coders/analysts? I find the idea interesting, but in reality this is never going to go further than an academic or research project ...

This seems like a neat idea and seems quite ambitious. I guess I'm jaded from all the other crypto-currency and I just can't get excited about this. I fear that this will be the only time I ever hear of Ethereum again, this moment when I read this article.

I wonder how such things as "the tyranny of the majority" will be handled, because if you make a poll at a gang rape, 9 out of ten will vote to continue the rape.

Who is behind this? I'm interested in talking to them.

Might work in some areas, but not the ones mentioned.

The author falls into many traps of the law vs code debate. There is a big difference between wet and dry code, as Nick Szabo, the original inventor of 'smart contracts' put it.

Admittedly, it is very stimulating to entertain thoughts in the direction the author takes his description. But in the end it is not a fruitful approach to understand ethereum and other smart contract systems.

Law has an open texture. It is deliberately broad and ambiguous to allow specific real life events to be subsumed under the conditions of a given regulation. Legal provisions do not have to be formulated in a natural language. They can be expressed in pictures or even in (pseudo) programming code. What is important is who (or what) parses and interprets the code. Wet code, i.e. legal provisions, are interpreted by humans. Nothing else can (currently) understand the "actual meaning" of a legal provision. Dry code is the opposite in that it is not interpreted by humans but by machines. Smart contracts - the official (but historically inaccurate) name of those programms that are hosted and executed on the ethereum distributed virtual machine - are written in dry code.

Ethereum is a 'rich and trusted' execution environment for software. 'Rich' because the ethereum virtual machine has access to a whole bunch of data that carries deep meaning to specific individuals and human society in general (or rather might some day). Trusted because the code execution is performed in a way that is extremely reliable. Whatever the competing interests behind the outcome of a specific computation may be, influence over the network may only be gained through computational resources (and at a later point probably stake/ether) which - in the real world - means economic resources. The achilles heel of a network such as ethereum (and bitcoin) is that "the crypto-economy" needs to stay causally disconnected from the other economy, that uses blockchains as a tool to run trusted fiduciary code (currency, insurance, identity, reputation).

"Law vs Code" is a very interesting talk to be had but there is no room for a "doomsday argument". Both are separate but intertwined domains. A fruitful approach to understanding the relationship between code and law may be Luhmann's system theory.

[further reading] https://en.wikipedia.org/wiki/Niklas_Luhmann https://medium.com/@heckerhut/smart-contracts-platforms-and-...

What is really interesting is when parties sue each other in court over Ethereum contracts. Courts will always be involved.

Definitely interesting, but how does it help with this?

> even if you "win", you have the problem of collecting.

I think because the actual procedure is included in the script, so "collection" takes place automatically in the code when it's processed.

It will be hard enough to get widespread social adoption of either blockchain currency or contracts written in programming languages without making them tied to each other.

A better choice for an initial compile target for EtherScript would simply be English.

English is ridiculously colonized by obfuscators already. Think about the Clinton I debates on what the meaning of 'is' is.

I think those same obfuscators will be happy to emigrate to any new linguistic country in which their skills can be put to use, but I think the Etherium folks are fantasizing that they can make contracts non-ambiguous. Good luck trying.

I think the problem comes in that the contracts must eventually have impact in human minds. However you slice it, there's still an interpretation step. Efforts to make those interpretations unambiguous will be opposed.

This convinced me to not waste another second on Ethereum. It serves a self-defeating purpose, and I think it's diametric to the hacker ethos. Bitcoin is neat because it can't be controlled, nobody can tax it or pose fees on it. It's free market at it's very best. The bitcoin protocol is weak, but luckily developers are strengthening it all the time.

Laws as a general concept are diametric to human nature, and they should be avoided whenever possible, especially in regards to social realms like bartering and contracting and the like.

Applications are open for YC Winter 2023

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact