Vitalik Buterin: Ethereum needs a "smarter DAO," as zero-knowledge and AI will reshape decentralized governance

January 19 News, Ethereum co-founder Vitalik Buterin posted on X stating that the Ethereum ecosystem “still urgently needs more DAOs, but they must be differently designed and of higher quality.” He emphasized that DAOs were once the core inspiration for Ethereum’s birth, but the current mainstream DAO forms have deviated seriously from their original goals, and a systematic governance upgrade is urgently needed.

Vitalik Buterin said that the early concept of DAOs was to allocate resources more efficiently and robustly than traditional governments and enterprises through code and rule systems running on decentralized networks. However, most DAOs in practice have evolved into “token voting treasuries.” While feasible in form, they generally suffer from low efficiency, susceptibility to whale manipulation, and inability to address human political flaws. These issues are key reasons why external confidence in DAOs has gradually waned.

Nevertheless, he clearly stated, “We still need DAOs,” especially in several critical areas. First is oracle design. Currently, decentralized stablecoins, prediction markets, and various DeFi infrastructure rely on and are hindered by suboptimal oracle models. Token-based oracles are difficult to resist capital manipulation, while human arbitration weakens decentralization. The fundamental problem lies in design flaws, not participant moral risks. Building more reliable oracles is both a technical challenge and a social collaboration issue.

Second is on-chain dispute resolution mechanisms, which are indispensable components of advanced smart contracts like insurance; additionally, there are public infrastructure elements such as maintaining security application lists, interface standard catalogs, and token contract address libraries. Vitalik Buterin also mentioned that DAOs have irreplaceable value in project rapid launch and long-term maintenance, especially in scenarios where legal entities are hard to cover and the original team has exited.

In terms of governance methodology, he again referenced the “convexity and concavity” framework: when a DAO addresses concave problems, it should focus on medianization and attack resistance of multi-source information; in convex optimization problems, decisive leadership and decentralized supervision are not contradictory. The real challenge lies in privacy protection and decision fatigue. Lack of privacy can turn governance into social gaming, while frequent decisions can quickly exhaust community participation.

He believes that new generation technologies are opening breakthroughs. Zero-knowledge proofs will become a core tool for governance privacy, AI can help alleviate decision fatigue, and more advanced consensus communication platforms will replace inefficient discussion environments. But he also warns that AI should never be directly entrusted with full management of DAOs; AI should be introduced cautiously to enhance human judgment, not replace human will.

In his view, future DAO tech stacks must treat the communication layer as a core component. Whether it is oracle projects or governance innovations, zero-knowledge, AI, and communication mechanisms should be regarded as at least “50% of the work focus.” Only then can the decentralization and robustness of Ethereum’s underlying layer truly extend to the entire crypto world built upon it.