Hacker News Reader

100%. I keep harping on the point that "smart contract" is such a bad name for these things. A real-life contract is not the execution of the task(s) itself, it's the legal recording and reinforcement of the promise(s) made by the parties if/when things go wrong.

Smart executor does sound pretty grimy though :)

In normal contracting the paper is the source code and you need to find some real live people to be the compiler. Sometimes it's everyday people, sometimes operations professionals, developers, finance or legal personnel. Ultimately you always can fall back on the court or arbitrator as the ultimate compiler and interpreter. So, different compilers maybe for smart contracts? ....

I'm trying to roll with you on this idea, but I can't help but think that this is just a square-peg round-hole thing. I am a (mostly non-practicing) lawyer; but I teach IT full time.

Like "A.I," a lot of people dream about the idea of revolutionizing or strongly improving on law via code, and I don't see it happening, ever. It's this well meaning and seductive idea that we can leverage the superior power of computery thought and ideas to "correct" the foibles of humans -- but I think "human disputes" are too slippery to be bound by code, and the beauty and triumph of "law" as a system is that it too is slippery enough to manage it reasonably.

To me it's strongly related to the complete layman fantasy of "well, they should have just written the law more clearly and everything would be fine."

Yes, and the reverse also is an interesting thought exercise-- if this paper contract really was a piece of software, what would folks do to QA it, write it, interpret it. I mean, paper contracts are written using hopelessly buggy and ambiguous human languages theoretically capable of so many non-harmonious interpretations. The solution really is about friendly, trusted compilers approaching the contracts from a fundamentally cooperative direction. You write a contract with your trusted partner on a napkin at dinner. And when trust fails, courts interpret in an explicitly prosocial way.

Don't discount the capriciousness of the law's execution. In many cases, being the right skin color and wealthy comes across like a superpower compared to Joe Average. I'd hazard a guess that many "use code for law" types basically want to put Lady Justice's blindfold back.

That doesn't mean it's a good idea, nor that is bad. It comes from good intentions, is all.

(Very aware that those can be used for paving certain roads)

Back? :)

Sure, I get the impulse to try and that in many ways is the lovely thing about the spirit of tech and hackery... buuuut also I'm a black man whos seen e.g. entirely too much "oh they were well intentioned" when the AI thought the black folks were monkeys. So, yeaaaah, not holding my breath. :)

Not just 23x...but 230x transaction value ($23M to send $100K). Not to mention it was 7,700,000x the going rate for the other transaction in that same block.

This might be a very naive question, but would it not be the case that many contracts would be essentially similar in nature? So while weird loopholes would exist at first, over time there could be a large library of tried-and-true contracts which could be applied to the situation at hand?

You are correct and there is already a bunch of off the shelf contracts that are basically plug and play...however it takes little knowledge to actually publish said contract and a small typo can either make the contact never accessible or easily exploitable.

That is arguably tending towards centralisation, as there will then be a market of brokers that know which contracts to suggest, and then there will be other places that will only accept contracts from a known broker.

There will, of course, be the more expensive brokers that can exploit loopholes, but even as a stretch this isn't vastly different to a bank account in the Caymans.

Of course, you don't have to participate - you can hold the 'cash' equivalent or concoct your own scheme, but you're not as protected from loss as you otherwise might be (provided there's an insurance package of sorts, in the absence of regulation).

Smart contract code is open-source[1], so the centralization you're talking about shouldn't be confused with the (de)centralization of power people talk about when they talk about the decentralized web or decentralized finance. Circling around a small set of trusted open-source technologies doesn't cause centralization of power, as we can see with GNU/Linux, HTML/JS, ...

[1]: On many platforms smart contracts are stored in the form of high-level interpreted languages. On Ethereum, the blockchain stores EVM code (assembly), but contracts that haven't "verified" their source, typically by uploading the high-level code to etherscan.org, are seldom used (with some notable exceptions).

I contend that it doesn't cause centralisation of power, but centralisation and power are inevitable should the project hit the mainstream, proper:

- GNU/Linux - GNU/Linux is open source but centralised. The userspace is the part that is distributed, via operating systems, and the source control is distributed, via git. But it's all for one Linux kernel. You can also build your own kernel, but that doesn't really make linux 'decentralized'. Similarly, Linux for a lot of people means 'Ubuntu'.

- HTML/JS - this is centralised under WhatWG/W3C, etc. Arguably, these days, it's actually centralised under Chrome, because what Chrome does eventually becomes the spec. You can freely build your own implementations of HTML and JS/Ecmascript but most likely, you are using the centralised implementation via webkit, blink, or gecko.

So, the fact that smart contract is open-source doesn't really mean anything. It'll grow big, then as a matter of convenience it will start to consolidate. git, for example, is decentralised, but git forges (github, gitlab, etc.) provide centralisation as a convenience.

Imagine a scenario where a popular contract is discovered to have a serious flaw years down the line, and they can't be patched.

The key difference (which is categorical) is that if there is a latent bug in one of these contracts that goes unnoticed and then gets exploited later, you're still out the money. Intent doesn't matter.

There is essentially a 0% chance of such an outcome with a regular contract (at least in US contract law). Intent is a cornerstone of contracts in the real world. Just because someone came up with a clever "exploit" doesn't entitle them to rip you off. This is a very good thing.

On Ethereum there's an organization called OpenZepplin that does exactly that. They have common code for making contracts that are ownable, upgradeable, follow NFT / ERC20 standards etc.

Makes it much easier to write safe solidity code.

> Since code is a lot more complex than written human language, expressing intent can be very difficult.

Isn’t that the other way around? Natural language remains an unsolved problem, and formal verification of software is pretty much a requirement for compilers. Your example is evidence of that: intent in natural language is easy to distort and hard to prove, thus making it much more complex

Well let's put it this way - in natural language, we say stuff and implicitly assume we've done our best to express intent, but it may not be perfect (hence lawyers). In smart contracts you express your intent into a contract and then we pretend there was no intent. There is only the contract. And that's a massive floor.

Right. Natural language is more complex and nuanced.

Hard to say if this satirical or not ("Etherium") but will respond in good faith. 1. Same argument can be used in the opposite way, you can verify the counter party will act as expect, there is no need for costly litigation as backstop. 2. An engineer reviewing a contract once will not be more expensive than hiring a full time banker to review each transaction or a lawyer to pursue a legal battle. 3. Decentralization and regulation are not mutually exclusive.

Agree with 1/, disagree with 2 and i have a different point of view on th 3rd point.

A lawyer can review a contract, miss a loophole/backdoor, and then argue to the court that the backdoor was placed there in bad faith, recoup his client losses. You can't do that with code. If you load a backdored code, you can't argue "this is unfair" to a court of justice.

On the 3rd point, i think "good" crypto (and by that i mean non-obvious shitcoins) tend to centralize a lot.

> You can't do that with code. If you load a backdored code, you can't argue "this is unfair" to a court of justice.

Why not? Assuming the miner and you are in the same jurisdiction, what prevents you from suing them? Seems relatively straightforward.

Because of limited responsability mainly. Did anyone attack microsoft for the catastrophic 0day ~2 years ago?

This accident actually proves the exact contrary. Despite these obvious flaws, people are still using DeFi.

There must be something of value in there that you're missing. And it's not speculation, as there are way better ways to speculate in crypto than by using DeFi (shittokens, NFTs, ...).

Lottery tickets aren't a good investment, but people still buy them.

It's certainly true that there is activity in the DeFi space, and it's certainly possible that this activity is because DeFi is just so gosh darn amazing at solving people's need for basic financial products.

...at least in theory. But if you're going to make that rather extraordinary claim, I think you need to provide some extraordinary evidence.

I'm disputing the logical conclusion from the facts at hand that no real problem is solved, not providing a value judgment about DeFi.

It's not an extraordinary claim but common knowledge in the start-up investment land that when you see an ugly, flawed product get an unexplained amount of engagement, you become interested.

> Lottery tickets aren't a good investment, but people still buy them.

Lottery tickets solve the real problem of hopelessness about improving one's financial situation through work. It might not be a worthwhile investment in financial terms, but it has its use just like entertainment media or alcohol.

People inject heroin, drink liters of soda and gamble their life savings away too, but none of that is evidence that these are healthy or socially valuable activities.

My explanation of the prevalence of DeFi is that it's currently at or near the peak of the hype cycle and there's a ton of money to be made either by trying to legitimatize the ecosystem or else via run of the mill pump and dumps.

"There must be something of value in these tulip bulbs because everyone is investing in them."

Investing != using.

https://cryptofees.info/

People pay tens of millions of dollars everyday to use these on-chain platforms. They could avoid all those fees if they just wanted to speculate on token prices by using centralized exchanges, of which many have very lax KYC requirements.

I don't have a breakdown of what people are using Ethereum for, but certainly a lot of those uses are stuff like NFTs, i.e. implementing other purely speculative assets. And there are lots of crypto startups using Ethereum, whose money is sourced from a different kind of speculation (er, "investment"). So again, it's not enough to just point at activity and say "look, real value!"

I don't disagree that the vast majority of the activity derives more or less indirectly from price speculation of some sort. But the thing to note is that so much of it is done on-chain instead of on centralized exchanges which have comparatively negligible fees. At the very least this proves the technologies themselves as suitable platforms for speculation (vs simply being the target of speculation).

"Speculation platform" is already a big, legitimate and very popular use case (the importance, influence and appeal of Wall-Street in the US and the whole world is evidence of this).

But how much of the "value" of the chain "as a platform for speculation" is simply the hype around crypto? (Which is, in large part, driven by speculation-driven rising prices of cryptocurrencies.)

For example, anyone could have sold the equivalent of NFTs decades ago using a centralized ledger (perhaps using multiple trusted third parties for security). That didn't happen.

Likewise, it's pretty clear that if you have a startup idea that could be implemented with or without crypto, you get a valuation boost if you take the crypto route. (Hello Long Blockchain Corp.)

So: speculation-driven cryptocurrency prices rise -> entire crypto sector hyped to the moon -> startups move in to collect hype-driven investment -> smart contract activity observed -> this "proof" of utility drives more hype.

And of course there are other intersecting loops.

James Mickens' "Blockchains are a bad idea" talk [1] is a great explanation that discusses exactly that.

[1]: https://www.youtube.com/watch?v=15RTC22Z2xI