Skip to content

Protocolized Modeling and Verification of Cyber Physical Systems - Giovanni Merlino

Lecture Giovanni Merlino

Giovanni Merlino, professor of computer science engineering at the University of Messina, presents the redesigned final module of his cyber-physical systems (CPS) course, framing formal verification techniques—linear temporal logic (LTL), finite state machine equivalence/refinement, and model checking—as a form of 'protocolization' that makes hidden assumptions, tradeoffs, and human-machine interactions explicit. He walks through concepts like safety vs. liveness properties, simulation vs. bisimulation, the CEGAR (counter-example guided abstraction refinement) technique, and Büchi automata, ultimately proposing a layered model (robot-level, swarm-level, smart-city-level) where engineering tensions like hyper-safety vs. liveness or pedestrian priority vs. system efficiency require explicit governance levers.

Related resources

Lecture

Designing Digital Worlds Communities, Protocols, AI, and Power - Andrés Monroy Hernández

Andrés Monroy-Hernández and Soyeon (So-young) outline a Princeton course on 'designing digital worlds,' tracing a lineage from Elinor Ostrom's commons governance and Yochai Benkler's commons-based peer production to today's decentralized social media protocols (ActivityPub, AT Protocol, Nostr, Farcaster), arguing each architecture encodes distinct political trade-offs around identity, curation, and infrastructure power. So presents her research on 'governance frictions' between federated communities and proposes design concepts like 'governance nutrition facts,' 'trust bubbles,' and an 'issue bulletin' to help fediverse moderators coordinate across incompatible norms.

commons-based-peer-production-vs-common-pool-resource-governance governance governance-frictions-and-trust-bubbles

Andrés Monroy-Hernández

Lecture

Beyond Consensus - Protocols as Digital Institutions - Martin Harrigan

Martin Harrigan, a lecturer and teaching fellow at SETU in Ireland, presents a method for using LLMs to study protocols: first building an LLM context file by transcribing and clustering ideas from the entire Summer of Protocols YouTube channel, then using that context plus a 10-property framework from 'The Unreasonable Sufficiency of Protocols' to roleplay protocols (like his 'Kevin the dinner host' example) and test properties such as defensibility, mortality, and legitimacy. He walks through building prompts, meta-prompting, and steering LLM conversations to demonstrate these properties, while noting weaknesses like circular reasoning and the difficulty of generalizing from a single instantiation.

hardness-joshua-stark llm-generated-protocol-studies-context meta-prompting-and-conversation-steering

Martin Harrigan

Lecture

Designing Trust Protocols, Society, and Web 3.0 - Helena Rong

Helena Rong, assistant professor at NYU Shanghai, previews her spring 2026 course on trust protocols for the agentic web, introducing 'trust experience design' as a framework with three layers—evidence (cryptographic attestations), primitives (verifiable credentials, zero-knowledge proofs), and rituals/experience (signals that make trust felt). She traces trust's evolution from value-based/social forms (handshakes, institutions) to proof-based/technical forms (blockchains, smart contracts), and discusses decentralized AI agents (DeAgents) like Truth Terminal and Sporef.fun as cases of emerging AI self-sovereignty that outpace existing trust protocols like A2A and ERC-8004. The session concludes with a group exercise where participants design trust protocols for near-future scenarios like autonomous mobility, smart grids, and robotic surgery.

agentic-web-trust-protocols-a2a-erc-8004-know-your-agent decentralized-ai-agents-deagents-and-self-sovereignty fiction

Helena Rong