Very well written article, This feels like very obvious bottleneck in the runtime especially given how things like NFT mints can clog up the network, do you think it's possible to modify Block STM over time to support true multi lane parallelization like Solana?