Crypto security has to be able to evolve. Quantum computers and AI-accelerated math breakthroughs could eventually break today’s common signature schemes. Cryptographic agility is the ability to swap in new cryptography (new signature methods, hash functions, etc.) as threats change, without grinding the whole network to a halt. @Starknet is already built for this. Native account abstraction means accounts are programmable smart contracts. Developers and users can introduce new signature schemes (including post-quantum ones like Falcon) on existing accounts without a network-wide hard fork. STARKs themselves are hash-based, so the proving system starts from a more quantum-resistant foundation than elliptic-curve SNARKs. But as an Ethereum L2, parts of Starknet’s security (the bridge, data availability, settlement) still inherit Ethereum’s cryptography and migration timeline. Ethereum’s public target for core post-quantum infrastructure is around 2029. That’s the practical reason Starknet is exploring an independent L1: so it can control the pace of its own core security upgrades and aim for full post-quantum readiness on a faster timeline (2027 target), rather than waiting on the base layer. The recent community poll showed strong support for researching this path. It’s research, not a decision yet but it reflects a clear priority on adaptability. Features like STRK20 shadow accounts (private yield, LPing, borrowing, options, etc.) sit on top of this foundation. Privacy and agility both benefit from the same programmable, ZK-native design. Built to adapt. The question is whether waiting on someone else’s upgrade schedule is the safest long-term bet.