Would you trust a proof of reserves that doesn’t reveal any addresses?

[TastyChillySauce00]

Well in theory Would I trust a Proof of Reserves that doesn’t reveal addresses? Yes, if they provided the ZK verifier code that is fully open-source, reproducible locally, and verifies key ownership. But it would be great if I could see the address and check it manually on blockchain explorer.

and always remember the bitcoin mantra "Dont Trust Verify"

I will only believe if zkPoH can verify to the most recent block otherwise I will prefer to verify on the address directly. The developer listed snapshot epochs under future work and that is great, but for now it's not there yet.

Although revealing reserve address is dangerous, it's still the surest way to know the money is there and it exists.

[joniboini]

I'd be willing to listen and hear arguments at least. If someone can argue and show how I can verify something without exposing too much, then I'd be willing to do it. Who knows, maybe I'll some usage for that in the future. I'll need something convincing though. At the very least the explanation have to be easy to understand and verify for the average joe.

[X-ray]

Anything that can reduce the chance of wrench attack is welcome we've gotten no shortage of crypto people getting kidnapped and we need ZK-anything related that can keep important detail hidden, I liked the ZK-KYC idea, I will like this one too.

So my suggestion: there should be hybrid system. Use zkPoH for privacy, but it must be accompanied by an annual independent audit (third-party auditor) and a transparent accounting of liabilities. "Bitcoin's motto is" "" "Don't Trust, Verify" "" "- as long as I can't fully verify myself, I won't be superstitious."

Good take, another loophole is how we can be sure the off chain snapshot taken isn't altered in anyway and the merkle root is able to be trusted?

One of many ways to solve this is also hybrid system where independent audit is able to re derive the exact merkle root.

Though i'm betting that there will be better solution in the future than this hybrid approach, because it needs everyone to be able to verify instantly and not really dependent on annual audit which is quite infrequent and requires another trust.

[d5000]

I was wondering whether using different snapshots alone would be enough to stop someone from repeatedly proving ownership over time. Would something like a nullifier or possibly a one time proof mechanism eventually be needed to prevent the same UTXO from being reused?

If I understand the concept of a nullifier properly, it would not work for this example because you can always use other private keys for the challenge. So I unfortunately think you could never allow two UTXO snapshots with a single shared UTXO, which would limit privacy a bit (because you'd have to use relatively small snapshots of at most 1000 UTXOs, so the service is able to register various users).

However, the idea is not to make "double registering" impossible, which can't be achieved with this method (neither can it be done with the "original" method where you do reveal your UTXO), but to make it more inconvenient to register several accounts in a short time. So I think if the service simply blacklists every UTXO used in the recent snapshot for let's say 100 blocks (a bit less than a day), it could reach a good equilibrium between spam protection and privacy.

A problem could however be UTXO snapshot spam: if several users want to attack the service, they can try to submit thousands of UTXO snapshots. Here the service will probably combine the system with a traditional CAPTCHA-like structure.