One confusing point is - "After integrating ZK proofs in the future, we can achieve 'infinite' gas (validators only look at the proof, no need to rerun), and the problem is solved." What is the essential difference between this second phase Native Rollup and the current ZK-Rollup? The only differences I can think of are one verifies ZK proofs using contracts on L1, while the other directly verifies ZK proofs within the L1 client, but fundamentally it seems there’s no difference... Which technical expert can clarify this?