$HYPE
StableSnapshot Window: 2026-07-15 15:40 UTC · ← Back to Crypto Overview
Social Momentum Summary
Total Engagement - Comments: 1, Retweets: 0, Likes: 19, Impressions: 1593
Verbatim Community Citations & Social Evidence 2 source posts analyzed
This deal will be good for market
. @zkdragon explains formal verification. 00:00 Proofs vs. tests 00:27 The 3 properties of sound money 02:26 Why ZK proofs create unique risk 03:33 The Orchard bug was the only one [Spoken audio]: So, for verification is when I run a program, I have an expectation of what I want out. And, okay, today you just trust that, hey, some developers could define what did they want to happen, and then they wrote code that makes that happen in a bunch of tests. For verification gives you a mathematical proof that machines will check that, hey, here's the inputs, here's the outputs, and here's all the properties you want to hold. This is satisfied. So for Zcash, the end state of this looks, okay, if you do a payment, we need three high-level security properties. These, you actually own the money it's being spent, the money's not double spent, and value is preserved. If I do a payment, value in equals value out. If you have these three things, you basically have sound money, and we can have a formal proof that computer checks that says, yeah, this transaction satisfies these, or I would say this program satisfies those rules. As long as XYZ cryptographic assumptions hold, and now we out of proof know that these cryptographic assumptions hold up to true to the minus 128 soundness rules. So, for verification, it gives you the property that we have a mathematical proof that is checking this. I think the word math proof or mathematical proof does hide what's going on a little bit. If you haven't done those logic puzzles of all animals like parks. Bob does not like parks, but Bob is not an animal. Something like this, this is a simple implication that you can go prove in some of the logical rules. Formarification is really just that to very extreme layering. And so you can imagine if you have 50 of these clauses, yeah, you could go prove extensively. If you ever did a logic class, it is the most boring homework assignments known to man and it feels contrived. large applications doing that at scale in an automated way. That makes sense, yeah. So you're basically really mathematically guaranteeing that the code of the implementation does exactly what the specification does. So the only risk remaining is in the specification. Yeah, and so basically reducing this to two, we're doing, we're formally verifying two parts of things for Ironwood, which is actually the two scary parts. They covers everything for undetectable sound as bugs and a little bit more. So what does that mean? Why does Zcash have any net new risk? One way to frame in your head is privacy. But I think that's hiding it a little bit. What's really causing the risk is that with a zero knowledge proof that says, hey, there is this program, here's a proof and some input and output of that program. The input and output is commitments for the payment. And we require the proof to hold, which means that the program with a payment as input equals this payment as output. The proof compresses every step we need to check that. So that compression is what's introducing a risk that there could be an undetectable bug, that what if something got inflated, we wouldn't see it. Whereas you could have an inflation bug in kind of every chain, and you've had one in Bitcoin and elsewhere you can always imagine database errors do something. The unique thing is in other chains you can detect it immediately. You know with certainty did it happen. And then in Zcash, This compression from the ZK proof hides this. There could be a bug and we have no idea if it got exploited. So with forward verification, we fix this because we have this logical proof that this compression step did not do anything. And that compression step works with two parts. There is the program, and there is this zero knowledge proof verifier. So we formally proved both sides of that that, look, the program satisfies everything we need. So this orchard bug was basically one edge case was not satisfied in one part. It was perfect clockwork of the circuit. But now what we have is that component was, hey, let's do this thing inside of elliptic curves called elliptic curve scalar multiplications. So we say this program will only pass in this elliptic curve scalar multiplication if the input and output are provably a curve relation, which is cool. So now that this would have caught this bug and crazily there was only one bug We got all this for verification. We have mathematical proof. There's no others Yeah, so we basically the only remaining one. Yeah, and now we did this whole formal verification process Just to figure out that yeah, it was actually the only one left. Yeah, that makes sense
