Presenter
Evan Miyazono, Protocol Labs
Evan is the team lead for Research at Protocol Labs. He has lived in California his whole life, bouncing between NorCal and SoCal. He spent enough of his childhood at beaches that a sandy shore always feels familiar; the rest was spent sitting at a computer, reading books, or disassembling things. He eventually majored in Materials Science and Engineering at Stanford after unintentionally picking up minors in math and CS. He was then driven by his curiosity to a PhD in Applied Physics at Caltech and, upon completion, was lead to Protocol Labs by a desire to build systems that affect real people. Evan still loves beaches, reading, and disassembly. Figuring out how things do, should, or could work is a passion, and sharing that knowledge is a favorite hobby. He’ll go on for hours if you tell him you’re interested in his thesis, so set a time limit before starting on that particular odyssey of quantum optics and nanofabrication.
Summary:
Evan Miyazono presented a project on controllable AI and coordination systems, addressing challenges in specifying desired outcomes for AI systems. The focus is on controllability to mitigate risks associated with general intelligence by modifying constraints to bind AI behavior. The proposed Open Agency Architecture involves participants interacting with an LLM to generate outcome specifications, using reinforcement learning for controllable policies. Miyazono reminds us that collaboration among entities and research labs is crucial for building collaborative AI systems, and discusses how we could generate specifications, formal verification, and software patching. Finally, the potential of LLMs generating exploits and loopholes underscores the importance of aligning AI with the legal system.
