Encrypted mempools promise to conceal transaction content until ordering is final, addressing front-running and censorship on Ethereum and its Layer 2s. This talk traces the evolution of encrypted mempools from early schemes like Shutter to the emergence of primitives like Batched Decryption and Silent Setup, and examines the community proposals now aiming for protocol enshrinement. We argue that none of these designs can be economically viable and decentralised without programmable privacy to enable MEV extraction.

Recent years have seen an explosion in decentralized financial infrastructure. Within it, automated market makers (AMMs) implement continuously operating algorithmic markets that enable traders to exchange one asset for another programmatically. In this talk, we will take a principled approach towards a general framework for exchange design wherein liquidity providers (LPs) who deposit the assets traded on these markets specify non-increasing demand curves representing the quantity of risky assets to be held at each price. We will formalize the trade-offs such designs need to make, and connect this setting to a generalization of Myerson's optimal auction theory, placing for the first time on common ground the (so far distinct) literatures of mechanism design and market microstructure.

Liquid democracy is a transitive vote delegation mechanism over voting graphs. A central question in the sphere of delegated voting is when does liquid democracy outperform direct voting. In this talk, I will examine this question from two perspectives. In the first, I will look at liquid democracy in a setting with a ground truth voting outcome. Previously, Kahng et al. showed that it is impossible for liquid democracy to always outperform direct voting in general voting graphs in this setting. Specifically, there are graph topologies that result in every voter delegating their voters to a "dictator", which can have a potential negative impact on the final voting result. In this talk, I will show positive results for liquid democracy for specific graph topologies. A common feature of these graph topologies is that they preserve a certain amount of variance in the delegation process, which has implications on fairness in voting graphs. In the second perspective, I will examine liquid democracy when there is no ground truth voting outcome, and instead two opposing factions. I will show an interesting result that vote delegation enables the weaker faction to win the election with higher probability compared to direct voting, implying that delegation can be fairer than direct voting for the weaker faction.

Inequality measures such as the Gini coefficient are widely used to motivate public policy, and are increasingly being applied in blockchain settings to design fairer mechanisms. Several challenges impair the reliability of inequality measures. First, the pseudonymous nature and the fact that actors may create multiple sybil accounts, distort asset distribution. Second, there are complex token-holding patterns, where an entity’s assets may be stored across various smart contracts. In this work, we propose methods to improve the accuracy of inequality measurements in blockchain networks. We empirically test these methods on Ethereum and assess policy impacts. Our results suggest that while inequality in Ethereum appeared to decline over time, and several previous protocol changes such as the transition from PoW to PoS indeed reduced inequality, this trend has recently reversed.

Zero is a unified blockchain architecture built around a single, coherent security assumption, guided by two primary design goals. First, decentralization is only meaningful if home stakers can participate in consensus by running validators. Zero separates execution from verification and keeps the settlement layer lightweight. Validators verify succinct proofs and participate in consensus without executing full application workloads, lowering hardware requirements and enabling broad, permissionless participation. Second, Zero addresses the tension between security and upgradeability. Many blockchains rely on Layer 2 solutions for scalability while claiming to inherit Layer 1 security, yet retain unilateral control over execution logic through small security councils. Upgrade authority becomes a systemic risk when controlled by entities that do not share the Layer 1 trust assumptions. In Zero, protocol upgrades are governed on-chain by all stakeholders under the same security assumption as the settlement layer, eliminating additional trust dependencies. Architecturally, Zero is a unified multi-shard settlement layer where independent execution environments operate under a common security framework, enabling scalability without fragmenting its security model.

"RANDAO is Ethereum's randomness beacon for selecting block proposers. Unfortunately, economically rational validators can bias it to claim more than their fair share of blocks - and the transaction fees and rewards that come with them. In this talk, we attempt to answer the following two research questions. Given a rational adversary controlling a fraction α of the deposited stake, what is the theoretical upper bound on the proportion of blocks they can propose in expectation? How could we detect RANDAO manipulations in practice? Do we observe any RANDAO manipulations on Ethereum today? In "Forking the RANDAO", we searched for statistical evidence of RANDAO manipulations, and solved the theoretical research question with Markov Decision Processes. In our ongoing research, we showed that RANDAO is even more biasable than demonstrated previously. Perhaps surprisingly, an attacker with 35% of the staked ETH can propose 50% of blocks, outperforming an attacker on Bitcoin with 35% of the hash rate utilizing selfish mining. "