I reviewed @Starknet's implemented cryptographic changes again today, and the term "upgradeable" seems more relevant than simply "quantum resistant." Its proof layer uses hash-based STARKs, ensuring verification computations don't rely on elliptic curve cryptography—this is inherent to the architecture, not a later addition. The account layer uses native account abstractions; each account decides which signature to accept, allowing for scheme changes without a network-wide hard fork, while addresses and funds are preserved. An experimental Falcon-512 account transferred funds on the mainnet in August (not yet audited, just a demonstration), indicating this approach is viable. Some migrations have moved from planning to the mainnet. v0.14.3 changed the OS's program hash and configuration hash from Pedersen to a BLAKE2s-256-based scheme, with a clear scope and already live. Adjustments to the state tree and contract address derivation are not yet fully implemented, and migration tools for old contract storage are still under research. As an L2 layer, bridges and data availability are still affected by Ethereum's pace. The value of upgrades ultimately lies in whether the entire ecosystem can keep up smoothly when changes are needed, rather than simply changing the proof layer and considering it done.