Vitalik questions the L2 scaling path, Arbitrum, Optimism, and Base collectively respond towards the direction of de-homogenization
Ethereum co-founder Vitalik Buterin recently stated that the initial idea of using L2 as the main scaling engine for Ethereum is no longer reasonable, and he called for second-layer networks to evolve towards stronger specialization. This statement quickly prompted responses from mainstream L2 teams such as Arbitrum, Optimism, and Base.Vitalik pointed out that many L2s still rely on multi-signature bridges and have not fully inherited Ethereum's security; at the same time, with the increase in gas limits and the advancement of native Rollup solutions, the throughput capacity of the Ethereum mainnet itself is strengthening. Against this backdrop, the positioning of L2 needs to be rethought.Karl Floersch, co-founder of the Optimism Foundation, stated that he supports building a modular L2 that covers the entire decentralized spectrum, but also admitted that there are still real challenges such as long withdrawal periods, immature Stage 2 proof mechanisms, and insufficient cross-chain tools.Steven Goldfeder, co-founder of Offchain Labs, the developer of Arbitrum, emphasized that scaling remains the core value of L2. He believes that even if the Ethereum mainnet's capabilities improve, it will be difficult to replace L2's role in handling thousands of TPS during peak times, and warned that if Ethereum is unfriendly to Rollups, institutions may choose independent Layer 1s instead.Jesse Pollak, head of Base, stated that scaling Ethereum L1 is beneficial for the entire ecosystem, while agreeing that L2 cannot just be a cheaper version of Ethereum. He pointed out that Base is forming its own characteristics through application layer differentiation, account abstraction, and privacy features.Additionally, StarkWare CEO Eli Ben-Sasson hinted on social media that ZK-native L2s (such as Starknet) have, to some extent, aligned with the specialization direction proposed by Vitalik. Overall, L2 builders generally accept the trend of de-homogenization and repositioning, but there are still divergences on whether scaling remains a core mission.