Whitepaper Reading Club @ SBC [02]

💡About Suffer together (in a round-table discussion), so we can also find the truth together. 12 PM–1 PM: Hyperliquid Market Microstructure - Xinmeng (Stanford) 1 PM - 2 PM: Constellation - Max Resnick (Anza) 2 PM - 3 PM: Cadence - Mussadiq Jalalzai (Monad) 3 PM - 4 PM: ZK Acceleration - Ryan Kim (Fractalyze) 4 PM - 5 PM: AI Compute Markets - Jay (Pantera) Pizza provided for lunch 📝 Papers 1. Hyperliquid Market Microstructure (Stanford) Using Hyperliquid’s fully on-chain order book—$1.9T in notional across 386,000 wallets—this session examines who wins and loses in perpetual futures, and why. The paper classifies wallets by trading behavior and traces retail’s $720M loss through fees, execution costs, voluntary trading, and forced liquidations. Its central finding: retail pays for immediacy by crossing the spread, trading into thin liquidity, and exiting with the crowd. On-chain, the purpose of a trade—opening, closing a winner, or closing a loser—reveals more than the trader’s identity. Links: [COMING] 🧠 Session Lead: Xinmeng (PHD Student at Stanford) Xinmeng is a third-year PhD student at Stanford University, advised by Ruizhe Jia. Her research focuses on DeFi, blockchain infrastructure, and market microstructure, particularly perpetual futures markets. She holds bachelor’s and master’s degrees in Mathematics from ETH Zurich. 2. Constellation: Multiple Concurrent Proposers (Solana) Redesign of Solana block production, replacing single-leader model with a 50ms multi-proposer pipeline that separates transaction collection, attestation, inclusion, and ordering using: ... [COME FOR THE SESION] ... Goal to reduce leader censorship, MEV side deals/sandwiches, and validator revenue variance. Links: Paper | Summary (coming) | Summary: Comparison 🧠 Session Lead: Max Resnick (Lead Economist at Anza) Focused on Solana fee markets, validator incentives, MCP designs and more. He previously worked on Eth MEV, PBS and L2 research. He held research roles at Special Mechanisms Group and Risk Harbor, with an academic background in economics and mathematics from MIT and the University of Michigan. 3. Cadence: Extreme Pipelining with Multiple Concurrent Proposers (Monad) Cadence redesigns BFT block production by replacing a single leader with multiple proposers running in parallel. Each slot reaches consensus independently, so it does not wait for earlier blocks or network delays. It is built on two parts: Chorus, a fast 3-round consensus protocol, and Conductor, which controls when new slots open. Cadence provides short-term censorship resistance, hides proposals until they are revealed, and keeps the same security (3f+1 validators) and low latency as single-leader systems. In simulations on Monad’s 200-validator network with 5 proposers per slot, it achieved 219 ms average finality (167 ms speculative) and 50 ms average transaction wait time with 100 ms block intervals. Links: Paper | Summary (coming) 🧠 Session Lead: Mussadiq Jalalzai (Researcher at Category Labs) Working on distributed systems, blockchain infrastructure, and mechanism design. His research focuses on building efficient markets and scalable protocols. 4. Cryptography's XLA Moment: Why a MLIR Compiler & Not Another Proof System Cryptographic engineering is where ML was in 2015: hand-written kernels, GPU lock-in, and fragmented infrastructure. This session explores a domain-specific MLIR compiler that brings the XLA-style escape hatch to crypto, compiling one Python source across GPU targets while matching or beating hand-tuned provers. Links: Summary 🧠 Session Lead: Ryan Kim (Co-founder of Fractalyze) Bringing a background in ZK compilers, zkEVM/zkRollup systems, robotics, and ML infrastructure from A41, Lightscale, NAVER Labs, and Samsung Research. 5. Compute Markets As AI infrastructure scales, compute is taking on the financial structure of a commodity market. But unlike oil, it is perishable, location-bound, hardware-specific, and configuration-sensitive. It more closely resembles electricity, with capacity contracts, real-time pricing, and regional or hardware premiums that create basis risk. This discussion will examine how compute markets may evolve from today’s bilateral reservations and OTC transactions toward reference prices, forward curves, derivatives, and clearing. We’ll explore the preconditions for compute markets to form, its likely market structure, and what the future for transacting compute may look like. Links: Compute Market | Power Market | Summary (coming) 🧠 Session Lead: Jay Yu (Junior Partner at Pantera Capital) Focused on research and investments and a Research Advisor at IC3. His work spans DeFi and DAO mechanism design, AI agents and TEE infrastructure, and is currently researching the emerging financialization of compute. 📍 Direction: School of Computing & Data From Alumni Center (10 mins walk): Map 🟢 Format Short introduction and context setting to start. Silent reading of the summary before discussion begins. 5-10 minute presentation to frame the key ideas. Moderator-led discussion with no promotions. ❤️ Thank you IC3 (for making this discussion possible and SBC Conference & Stanford) IC3 is a Cornell Tech–based research initiative spanning faculty from CMU, Cornell, EPFL, ETH Zurich, Princeton, UC Berkeley, UIUC, Yale, and others, focused on turning rigorous blockchain, smart-contract, cryptography, distributed-systems, and security research into open-source, production-ready systems. About Whitepaper Reading Club We are a community of founders, researchers, and builders across Singapore, Malaysia, San Francisco, Bangkok, New York, Lagos, Taipei, and Hong Kong. We meet in person every month to read, discuss, and pressure-test the latest blockchain papers, protocols, and technical ideas. Learn more: Website · Summaries · Calendar We create detailed, easy-to-understand summaries for each paper and have held 80+ sessions since June 2023, covering Account Abstraction, Parallel Chains, EIPs, the Bitcoin ecosystem, and AI x Crypto. We are ecosystem-agnostic, not for profit, and focused on projects with technical, product, or social innovation.