In that previous post about writing DUSK by XSC, someone in the comments asked: “For every transaction, do we really use zero-knowledge proofs? If we run it, won’t it be so slow that it can’t be used?” The question isn’t random—I’m taking it seriously and answering it properly.
The first half of an XSC transaction happens locally. The amount, the counterparty, and the execution process are all computed entirely on your own machine. A proof is generated, and only the proof itself is submitted on-chain. The part about queuing and waiting for broadcast pressure isn’t even on the privacy-computation path. What DUSK’s on-chain nodes need to do is verify that proof—not rerun the whole transaction. Verification is much lighter than proving. Once you see that difference, it’s obvious. $ETH
Confidential smart contracts go down the same road. In scenarios like confidential DeFi and digital share registry systems, the bulk of the computation is also done locally. The Piecrust layer (now the core execution environment of DuskVM) is responsible for making execution fast, while the chain only records the verification result. I looked up the Gas accounting in the documentation: it charges for the cost of verification and execution. According to the latest docs, privacy itself doesn’t add an extra fee layer. $BNB
I went through the proving and verification paths against the documentation myself, and confirmed that re-computation truly happens locally, with the chain only doing lightweight verification. That question last time was asked the right way, and my answer after verifying it can stand up to the scrutiny too: it’s not that privacy is slow—it’s that there’s no chain that’s designed to handle privacy in the first place. #dusk $DUSK @Dusk
The first half of an XSC transaction happens locally. The amount, the counterparty, and the execution process are all computed entirely on your own machine. A proof is generated, and only the proof itself is submitted on-chain. The part about queuing and waiting for broadcast pressure isn’t even on the privacy-computation path. What DUSK’s on-chain nodes need to do is verify that proof—not rerun the whole transaction. Verification is much lighter than proving. Once you see that difference, it’s obvious. $ETH
Confidential smart contracts go down the same road. In scenarios like confidential DeFi and digital share registry systems, the bulk of the computation is also done locally. The Piecrust layer (now the core execution environment of DuskVM) is responsible for making execution fast, while the chain only records the verification result. I looked up the Gas accounting in the documentation: it charges for the cost of verification and execution. According to the latest docs, privacy itself doesn’t add an extra fee layer. $BNB
I went through the proving and verification paths against the documentation myself, and confirmed that re-computation truly happens locally, with the chain only doing lightweight verification. That question last time was asked the right way, and my answer after verifying it can stand up to the scrutiny too: it’s not that privacy is slow—it’s that there’s no chain that’s designed to handle privacy in the first place. #dusk $DUSK @Dusk