#dusk $DUSK @Dusk
The part of @Dusk_Foundation's design I'm most genuinely uncertain about is proof-generation cost at scale.
Generating a zk-SNARK proof isn't free — computational cost scales with circuit complexity, and across zk systems generally, proving times for complex circuits can run from sub-second to tens of seconds depending on hardware and constraint count.
Confidential smart contracts with compliance logic baked in are, almost by definition, more complex circuits than a plain transfer. That has real implications: who bears the proving cost (the user's device, a delegated prover, the node), and does that create centralization pressure if only well-resourced provers can generate proofs quickly enough to be useful?
Dusk has iterated on this (PLONK, then PlonKup with lookup tables to speed up complex proofs), which shows they're aware of the bottleneck. Whether it's fully solved for institutional-grade transaction volume is an open, empirical question, not something I'd take on faith.
$DUSK #dusk
The part of @Dusk_Foundation's design I'm most genuinely uncertain about is proof-generation cost at scale.
Generating a zk-SNARK proof isn't free — computational cost scales with circuit complexity, and across zk systems generally, proving times for complex circuits can run from sub-second to tens of seconds depending on hardware and constraint count.
Confidential smart contracts with compliance logic baked in are, almost by definition, more complex circuits than a plain transfer. That has real implications: who bears the proving cost (the user's device, a delegated prover, the node), and does that create centralization pressure if only well-resourced provers can generate proofs quickly enough to be useful?
Dusk has iterated on this (PLONK, then PlonKup with lookup tables to speed up complex proofs), which shows they're aware of the bottleneck. Whether it's fully solved for institutional-grade transaction volume is an open, empirical question, not something I'd take on faith.
$DUSK #dusk
