Muriel Médard | Co-founder at Optimum / NEC Professor of Software Science and Engineering at MIT скачать в хорошем качестве

Muriel Médard | Co-founder at Optimum / NEC Professor of Software Science and Engineering at MIT

Join Dr. Lorena Nessi from CCN in a wide-ranging conversation with Prof. Muriel Médard, Co-founder at Optimum / NEC Professor of Software Science and Engineering at MIT, on why random linear network coding (RLNC) matters for Web3, how a true memory layer can fix propagation and scalability, and what post-quantum security looks like in practice. Médard explains Optimum’s validator-first approach, updates on work across Ethereum and Solana, and why decentralization is an engineering necessity, not a philosophy. The discussion spans misconceptions on networking, AI’s need for shared on-chain data, and lightweight Flex nodes for global participation. Topics: decentralized coding, sockets for Web3 memory, gossip/sub & Turbine, bandwidth economics, survivability and censorship resistance, education across disciplines, post-quantum cryptography, RLNC efficiency, and upcoming hacknets. 00:00 – Intro and guest setup 00:31 – From information theory to Web3: why build Optimum 03:03 – Why decentralized coding scales when control traffic doesn’t 05:23 – Centralization vs gravity: engineering limits and outages 07:16 – The Web3 “memory problem” explained in simple terms 09:46 – Compute needs memory: what Web3 still lacks 11:48 – Doing it “properly”: validator-first speed and APY impact 14:03 – How Optimum accelerates propagation in practice 16:05 – Ethereum gossip/sub sidecar, permissionless integration 20:44 – Reed–Solomon vs RLNC: 1950s codes and modern scaling 22:30 – Let nodes cooperate: bandwidth, math, and efficiency 23:32 – RLNC basics: blending packets, accuracy, and anti-censorship 26:04 – Survivability under attacks; stronger hashing guarantees 28:30 – Data availability with RLNC: orders-of-magnitude gains 30:31 – A shared memory layer beats app-to-app bridges 33:33 – Analogy: atomicity, consistency, durability 37:19 – Biggest misconceptions on Web3 and networking science 40:08 – Star topologies, “dedicated links,” and real bottlenecks 43:36 – Education and cross-pollination: bringing fields together 45:29 – AI on-chain needs memory; no giant data shuffles 47:56 – Post-quantum security: Shor, fragility, and coding 50:22 – Keeping security while cutting costs 52:27 – What’s next: Flex nodes, hacknets, and participation 54:05 – Closing #XYO #DePIN #Blockchain #Web3 #Crypto #Networking #RLNC #MIT #Optimum #PostQuantum #AI