Skip to content
Blockworks

Debate How Should Ethereum Scale | PMLS 2025 | Day 2 | Blockchain Design

Macauley Peterson, Dan Robinson, mteam.eth, Max Resnick

0:000:00

Key takeaways

  • L1 First, but “try everything”: Dan Robinson argues that Ethereum must pursue every credible scaling path, yet prioritise raising Layer-1 bandwidth and cutting latency even if that lifts validator hardware requirements.
  • Fragmented Vision Hurts Progress: Max Resnick warns that splitting resources between a rollup-centric roadmap and L1 upgrades leaves both unfinished; Ethereum needs a decisive strategic commitment.
  • Solo-Staker Ideology Is a Bottleneck: All panelists agree the current obsession with home staking constrains block size, slot times and client design; raising the minimum stake and formalising validator specs are likely.
  • Governance Process Is Broken: Repeated last-minute U-turns (e.g., EVM Object Format) expose unclear inclusion rules and an overloaded Core-Dev call; written, client-weighted voting and top-down prioritisation were proposed.
  • Easy Wins on Latency Exist: Resnick claims slot time can be cut from 12 s to 6 s simply by changing constants and improving the peer-to-peer layer with techniques like Turbine or Random Linear Network Coding.
  • Skepticism of ePBS: Enshrining Proposer-Builder Separation with an auction that maximises MEV is viewed as the single most dangerous pending change; the panel would separate asynchronous execution from the MEV auction logic.
  • L2s Still Capture Value—For Now: Robinson maintains that chains like Base keep ETH as money, so “declaring war” on rollups would backfire. Resnick counters that today’s centralized-sequencer L2s can walk away at any time.

On L1 vs. L2 Scaling Strategy

Macauley Peterson opens with the core dispute: should Ethereum primarily scale its base chain or lean on rollups. Dan Robinson takes an inclusive stance, supporting gas-limit hikes, lower slot times and heavier validators while still welcoming optimistic and ZK L2s. Max Resnick draws an analogy with developing economies: Ethereum is experimenting with exotic tech while competitors like Solana already deliver 100× throughput using straightforward engineering. He urges Ethereum to “fire under-performing employees,” meaning stop donating Core-Dev attention to L2s that extract more value than they return. mteam.eth tries to reconcile the two views: prioritize L1 sequencing bandwidth first, let hyperscaled centralized-sequencer L2s live for now, and push based rollups only once the base layer can handle them.

On Governance, Roadmap and EOF

All three highlight that the multi-client, consensus-by-Zoom approach stalls progress. The late removal of the EVM Object Format after extensive engineering is held up as proof. mteam.eth calls for a separate prioritisation committee led by the Ethereum Foundation’s executive arm. Resnick wants each client team to submit written votes, weighted by real network share, replacing verbal filibusters. Robinson adds that the slowest-to-ship client currently sets the tempo; tough deadlines or fewer clients may be necessary.

On Validator Counts, Solo Stakers and Node Specs

Spectra’s Max Effective Balance tweak has reduced validators by barely two percent. According to mteam.eth, large operators were unprepared and sometimes earn less yield after consolidation. Robinson sees the bigger issue in design philosophy: the network has allowed solo-staker sentiment to veto hardware or bandwidth increases, even when empirical data says otherwise. Resnick notes that without a formal spec, discussions devolve into ideology; once requirements are codified (EIP-7870), the community can debate numbers rather than principles. Feather-staking or separate “proposer” and “attester” tiers could square higher performance with home verification.

On Latency, Gas and Networking

Raising the gas limit to 60 million remains politically stuck, yet the panel thinks throughput can double without hurting decentralisation. More striking is slot time. Resnick insists half the delay is arbitrary padding; better block-gossip protocols could drop to three seconds with no protocol change, citing Solana’s Turbine as a template. mteam.eth supports formalising higher node bandwidth then voting in gas-limit hikes. Robinson emphasises that the cultural shift—accepting heavier validators—is as important as the engineering.

On ePBS and Asynchronous Execution

Asked which proposal they would kill, all three point to enshrined PBS. Robinson likens it to “making sandwiches mandatory”—locking in maximal MEV extraction. Resnick would salvage the underlying asynchronous execution work but strip out the builder auction. mteam.eth notes there are cleaner EIPs that deliver delayed execution without MEV enshrinement and urges focusing on those.

Key actionable insights

  • Track discussions on raising the gas limit and halving slot times; any consensus here signals real near-term throughput gains.
  • Monitor EIP-7870 and related efforts to formalise validator specifications; agreement would unlock bolder hardware requirements.
  • Watch whether enshrined PBS is decoupled from asynchronous execution; separation would indicate growing resistance to MEV-maximising designs.