Hacker News Reader

Comment by rawtxapp | Hacker News Reader

The flip side of that is any smart contract that stood the test of time should be rock solid. For example, there are huge incentives to go ahead and hack a big contract like maker, compound, uniswap or aave, so you can bet that there's highly qualified people out there trying to hack them as we speak, yet after all this time, they are still working as intended.

I have a lot more trust in that kind of product than in a product where contracts are subject to interpretations by humans which may or may not be reliable. It's one of the reasons why people incorporate their companies in Delaware, there are well known case law, so you know in advance what to expect from the justice system. Predictability is an important part of a contract, I trust a well inspected and battle tested smart contract much more than a human enforced contract. That said, not everything is suited to smart contracts, but many financial applications are.

View on HN

Considering there are still new bugs found in chip designs, operating systems, and compilers that have been around 4 times as long as any cryptocurrency I have literally no idea why you would feel safe in “smart contracts.” At least if my bank account is hacked, I have a solid legal standing for compensation and I have good reason to believe centralized financial institutions will keep extensive documentation and logs. With decentralized finance, all I have is some Medium dot com post on the “Forward Plan” and an ominous message that “SAFEDOLLAR IS UNDER ATTACK.”

The bugs found in smart contracts equal large sums of money right away unlike 0-days in OSes or other systems where the link is much more indirect. For example, if you find a 0-day on Windows, sure you might sell it for 7-8 figures on the black market, if you found a bug on maker (or any of the big contracts), you could walk away with billions of dollars in a practically untraceable way.

https://en.wikipedia.org/wiki/Market_for_zero-day_exploits#M... The wikipedia entry on this is a bit misleading or misinformed. As there are multiple above ground buyers for 0-days that have no interest in making it difficult for people to sell to them.

There's still a middle-man to deal with in that process. In this case, you extract a decent amount of 'money' directly.

Would you rather win $100 and have to make a phone call to claim it, or find a $100 bill on the street?

> Considering there are still new bugs found in chip designs, operating systems, and compilers that have been around 4 times as long

Not quite the same. Those codebases are much larger, often include legacy spaghetti code, or are closed source. This makes it more difficult to find exploits, not to mention there's not as clear of path of financial incentive.

As someone that has experience in compilers and smart contracts, I can tell you they are an order of magnitude or more apart in terms of difficulty. Most smart contracts are not that complicated.

I'm not sure this follows.

It's a fair point that any easy-to-find bugs in large, time-tested contracts will have been found. Any medium- or hard-to-find bugs also will probably have been found. But there might still be a very-hard-to-find bug lurking; and given the value of finding such a bug, people might look hard enough to find it.

In other words, the same scale that ensures there is no low-hanging fruit, also provides the incentive to pick high-hanging fruit.

> I have a lot more trust in that kind of product than in a product where contracts are subject to interpretations by humans which may or may not be reliable.

This is a valid concern, but it's also extremely well-understood at this point; we have centuries of global-scale experience with traditional financial and legal systems. They're certainly not flawless, but it's rare for gigantic new flaws to emerge. An important point is the existence of mechanisms for rolling back bad transactions and challenging / appealing flawed decisions.

> it's rare for gigantic new flaws to emerge

There are countless instances of rouge traders, high level financial crime and corruption.

Wirecard is one recent example costing billions. The Libor scandal another that comes to mind.

Who were the victims with wirecard?

They faked the numbers, but they didn't manage people's money.

I can't tell if you're sarcastic or not. In case you're serious, you don't see how a company with a market cap of 28B$ going belly up causes damage to investors, the markets, their partners, clients and employees?

The stock going belly up is the consequence of the fraud, but not the actual fraud. Just like the other people you've mentioned.

Eg. Their clients switched payment processor.

When a btc exchange frauds or a coin hacked or a smart contract is bugged, they get away with all your money. Which is the actual fraud.

In some cases they post a postmortem on medium and call it a day.

I don't think wirecard's fraud is comparable with all the crap in crypto town.

How many major bugs in software and especially cryptosystems went undisclosed for decades?

The core issue is the inherent asymmetry where 1 person finding 1 bug can destabilize giant systems. Even if these systems where hundreds of years old that doesn’t actually mean much.

By this logic most e-commerce is a mistake as well, since a serious flaw in the linux kernel could take down most web servers.

Nah, that's governed by contracts, law and enforced by the police. That money tends to be traceable and recoverable.

Credit card transactions are reversible, that means such flaws can be considered a cost of doing business rather than permanently destroying anything.

With ecommerce the worst you can do as a seller is take out as much money as the credit limit on the card. Which for most people is in the low thousands.

With smart contracts the liability upside is orders of magnitude higher.

How many bugs had a direct financial return in the same way? If you found a bug in maker (or any of the other big contracts) today, you can walk away with billions of dollars worth of coins, that's a huge sum.

Major bugs in smart contracts can cause the system to stop working rather than handing anyone billions. As such people can discover such issues without disclosing them in much the way infrastructure can be vulnerable without people damaging it.

The Parity wallet bug from 2017 is the classic example so far, where someone who was just messing around with the smart contract 'accidentally' locked up $280m in assets[1]. I've linked to the GitHub thread below...this person didn't get anything from the exploit, and it wasn't for street cred (they deleted their account).

https://github.com/openethereum/parity-ethereum/issues/6995#...

https://www.cnbc.com/2017/11/08/accidental-bug-may-have-froz...

Even being able to grind a smart contract to a halt would be very profitable if you shorted it right beforehand.

Even then you are limited things to people who are willing to break the law and have a giant sign pointing them as a likely culprit.

I mean people short crypto all the time, averaged over a long enough period of time long/short ratios are roughly ~50/50 [1]. It's normal market behavior. And if you're shorting one of the big ones, you can most likely do it on defi without even providing a name.

1: https://www.bybt.com/LongShortRatio

Let’s assume something like 100 people have very large* short positions on a system that goes down. Dropping the possibility from X,000,000,000 to 100 is quite a big deal.

*very large being required for significant money to be at stake, making say 10k illegally from a short is hardly going to inspire a lot of effort.

Yeah, the classic example is the original binary search algorithm. It had an overflow bug that went undetected for two decades: https://thebittheories.com/the-curious-case-of-binary-search...

It's an incredibly simple algorithm and still got missed. Smart contracts have no chance.

There was not $Billions to be paid out for finding that - it's a matter of incentives.

Not every way to break a smart contract results in a major payout.

How often was that bug actually hit in production? How many people are allocating arrays of size 2*31? A bug is less likely to be caught if it's not exploitable.

This sounds like a fallacy of the inverse. Those contracts not being hacked yet is no proof that they are resistant to hacks.

It's not clear what fallacy you're thinking about because the fallacy of the inverse is not it.

From https://en.wikipedia.org/wiki/Confusion_of_the_inverse:

> Confusion of the inverse, also called the conditional probability fallacy or the inverse fallacy, is a logical fallacy whereupon a conditional probability is equated with its inverse; that is, given two events A and B, the probability of A happening given that B has happened is assumed to be about the same as the probability of B given A, when there is actually no evidence for this assumption.[1][2] More formally, P(A|B) is assumed to be approximately equal to P(B|A).

The inverse of "any contract that has been hacked was insecure" is "any contract that hasn't been hacked must be secure".

I think this is what OP meant. If something has been around for a long time it does probably mean it's less likely there is a really obvious security flaw, but it doesn't necessarily mean it is 'rock-solid' as plenty of things that have been seen to be 'rock-solid' in the past have turned out to be insecure.

yes this is what I was implying. I am curious how else the previous commenter would interpret my statement.

Let's formalize this to make it easier to discuss.

  𝐏(𝐀)  = 𝐏(contract has been hacked)
  𝐏(¬𝐀) = 𝐏(contract has not been hacked)
  𝐏(𝐁)  = 𝐏(contract is hack resistant)

Relevant conditional probabilities:

  𝐏(𝐁|¬𝐀) =
    𝐏(contract is hack-resistant given that it has not been hacked)
  𝐏(¬𝐀|𝐁) =
    𝐏(contract has not been hacked given that it is hack-resistant)

The fallacy of the inverse would be assuming that:

> The probability that a contract is hack-resistant, given that it has not been hacked, is approximately equal to the probability that it has not been hacked, given that it's hack resistant.

More succinctly:

  𝐏(𝐁|¬𝐀) ≅ 𝐏(¬𝐀|𝐁)

In https://news.ycombinator.com/item?id=27666484, the fallacy of the inverse was presented as "those contracts not being hacked yet is no proof that they are resistant to hacks" or "NOT (NOT A implies B)", i.e.:

  ¬(¬𝐀 → 𝐁)

In summary:

  𝐏(𝐁|¬𝐀) ≅ 𝐏(¬𝐀|𝐁) => the fallacy of the inverse

    and

  ¬(¬𝐀 → 𝐁)         => statement in comment

These two statements are fundamentally different.

Note that the first statement is a comparison of probabilities, and the second is not. They're not the same. There might be another fallacy at play here, but it's not the fallacy of the inverse.

To be honest, I don’t fully follow the above, and I’m not sure if this formalises the same problem. Copying the definition from Wikipedia:

> If P, then Q.

> Therefore, if not P, then not Q.

In this case, we could plug in the below values:

P = Contract Hacked Q = Insecure

I.e. if the contract has been hacked the code was insecure. Therefore if the contract has not been hacked, it is not insecure.

It is pretty good evidence that they are resistant to hacks though, and speaking about interesting evidence that doesn't pass the bar of proof isn't some kind of logical fallacy. That's a pretty high bar to set for what you're willing to integrate information-wise about the world into your model of it.

I get where you are coming from but this is a highly adversarial environment where huge amounts of other people's money is at stake with a history of very clever hacks. This looks like precisely the sorts of systems where you ought to raise the bar as high as possible.

Certain systems may have proven themselves somewhat over time but for this I say there is always the possibility for some extremely sophisticated attack and for any other new system they by definition wont have even this proven track record.

> The flip side of that is any smart contract that stood the test of time should be rock solid.

They all are until they aren't.

That's no guarantee. Someone might have a zero-day stashed away somewhere.

The game theory comes into play, if you found a bug that could empty maker's vaults for example, you better do it right now, otherwise, someone else will before you and you wasted all your time/energy.

Exactly. This is precisely why there aren't (or at least, in a rational world, should not be) any undisclosed zero-days in crypto. If you're aware of an exploit, you're just going to exploit it immediately unless you're willing to let someone else exploit it in your place.

There is a large amount of money at stake with these exploits and it doesn't make economic sense to let one sit around.

I would trust things with formal proofs a lot more

Reply on news.ycombinator.com