1/x Prior to #DevconVI, at #RollupDay, @VitalikButerin talked about “Hardening rollups with multi-proofs”. Here is a thread of brief summary.
2/x Nowadays, almost all #rollup protocols are still on “training wheels”, as proofs are either in development or can be easily overridden by simple governance.
3/x Even if we have proof systems ready say validity proof for zk-rollups, due to the complexity of the generated proofs, it’s quite tough to ensure there’s no bug in the proof. One example below, for a simple code snippet in #PSE zk-EVM impl, it has 34,469 lines of circuit code.
4/x It’s almost impossible to ensure so many lines of code are bug-free for a long long time.
5/x Option 1: to run high-threshold(the super majority of guardians or token holders) governance to override the buggy states from the single proof system.
6/x Option 2: multi provers say fraud proof from optimistic rollup and validity proof from zk rollup. Or different implements of the same proof system like WASM or MIPS for fraud proof.
7/x Option 3: two different types of provers plus governance.
8/x The advantages for option 3 are to avoid to solely trust governance and minimize the chance to have simultaneous bugs in different types of provers.
9/c To conclude, #ZK-EVMs won’t be bug-free for long time. To minimize the chance of bugs, we need multiple implementations from #op/zk rollups. The governance can be considered only when emergency happens.
12/x our friend @stonecoldpat0@infura_io shared about the various bridge approaches and how to do proper validation for them. Meanwhile, he also left a bunch of open problems for rollups.
1) With the boom of dApps including games like @AxieInfinity and NFTs like @apecoin, more app-tailored scaling solutions are urgently needed. Existing L1s and L2s can provide high throughput to some extent,
2) but due to the limited resource pool say computation shared among tons of thousands of dApps, the popular dApps like Axie and Ape with an enormous user base can easily clog or stall these protocols, exemplified by recent NFT mint events happening on @solana and @0xPolygon.
3) At Alt Research, with our expertise in sharding, interoperability and rollups, we conduct research in alternative scaling solutions that will lay the foundation for future dApps in Web3.