I noticed that Vitalik recently shared important details about the future roadmap for the execution layer in Ethereum, and the topic is very worth paying attention to.

The first point concerns the state tree – Vitalik supports transitioning from the current hexagonal structure to a more efficient binary structure. The idea here is to use advanced hash functions like Blake3 or Poseidon instead of the current method, which will significantly reduce the length of Merkle branches and verification costs. Additionally, the plan includes consolidating storage cells into pages to reduce access costs, while maintaining certain metadata to enable future state expiration features.

Regarding the virtual machine, the discussion is more bold – it proposes replacing the EVM entirely in the long term, possibly using a RISC-V architecture. The new VM should offer higher execution efficiency, better proof capabilities, support for generating ZK proofs from the client side, and simplify code execution processes.

The implementation won't happen overnight – the plan begins with replacing pre-compiled contracts first, then allowing new contracts to be deployed on the updated virtual machine, and finally achieving backward compatibility and a gradual transition. This approach reflects the team’s careful effort to avoid breaking the current system.

Interestingly, these improvements will directly impact network efficiency and scalability – a topic of great interest to everyone in the community.