Ethereum needs a cleanup phase, Buterin says, as protocol complexity grows

In a post on X Buterin warned that chasing new features while rarely retiring old ones has produced a complex stack that undermines three pillars: trustlessness (users need “high priests” to interpret the protocol), the “walkaway test” (fresh teams could not realistically rebuild clients if incumbents disappeared), and individual self-sovereignty (even advanced users struggle to audit the system). He argued that simplicity, fewer cryptographic primitives in consensus-critical paths, and more invariants should be treated as first-class goals alongside throughput and decentralization.

Buterin framed “garbage collection” as a structured simplification pass: reduce total lines of consensus code, demote rarely used features from the core protocol into contracts, and keep gas-cost schedules tightly coupled to real resource usage. He pointed to past cleanups – most notably the move from proof-of-work to proof-of-stake and recent gas-cost reforms – as proof that large resets and rule clarifications can shrink attack surface and improve client implementability.

The comments also sharpen a philosophical split among major chains. Solana Labs CEO Anatoly Yakovenko, responding to Buterin’s recent “walkaway test” discussion, argued Solana must keep iterating rapidly to stay aligned with developer and user needs, even if no single group steers those changes. Buterin, by contrast, has pushed for Ethereum to reach a state that can operate “for decades” with minimal ongoing intervention once core principles are locked in.

Vitalik’s appeal lands as client teams balance overlapping pressures: maintaining backward-compatible pathways for long-lived applications; integrating new data-availability and execution features; and keeping multiple independent clients in sync without excessive implementation burden. A formalized “simplification” milestone would give contributors a rubric for subtractive changes – measured not only by runtime metrics, but by readability, auditability, and the feasibility of re-implementation by fresh teams—intended to preserve Ethereum’s long-term verifiability.

If adopted, a garbage-collection track would likely appear as recurring EIPs that: prune edge-case opcodes and underused features from the consensus core; replace ad hoc fee rules with resource-linked schedules; and move specialized functionality to higher layers with clearer interfaces. The immediate effect would be fewer consensus-critical branches for client authors, clearer spec boundaries for auditors, and simpler mental models for users running full nodes – changes Buterin views as essential to sustaining trustlessness at scale.