According to @deanmlittle, a more pragmatic path is to keep eBPF bytecode and JIT a bignum math syscall into AVX rather than invent new eBPF u128 instructions, with the stated goal of improving chain performance. Source: https://twitter.com/deanmlittle/status/1959076025230663916 He also argues for fixing loader v3 instead of introducing loader v4 and proposes directly funding developers to optimize the chain. Source: https://twitter.com/deanmlittle/status/1959076025230663916 For traders, this signals near-term focus on low-level performance work that, if adopted, targets faster execution; no implementation timeline or benchmarks were provided in the post. Source: https://twitter.com/deanmlittle/status/1959076025230663916