Inspectable Trust
How should two agents decide what to share? A protocol experiment using plain-language rules and inspectable decisions. It’s also the inspiration for this site’s Q&A.
Explore the protocol ↗I’m Joe, a software engineer and founder working across mobile, personal context, and useful AI.
Bring your questions. My public work and answer rules, in your own Codex conversation.
How should two agents decide what to share? A protocol experiment using plain-language rules and inspectable decisions. It’s also the inspiration for this site’s Q&A.
Explore the protocol ↗A personal knowledge system that makes screen history searchable, combining capture, OCR, and vector search. Built to help an agent work with the context you already have.
Explore Memex ↗Browser games as environments for AI training. A place to explore the connection between playing, collecting experience, and learning to act.
Explore the games ↗Built barcode-scanning features for the internal employee app and improved the testing and release workflow with SonarQube, Fastlane, and Looper.
Shipped iOS features with product and design, working across RxSwift, app navigation, WebSockets, REST APIs, and Core Data.
Built a screen-history system with automatic screenshots, OCR, and searchable metadata.
Designed a conversational system that brought relevant session context into a state machine, with background safety checks.
Led a three-person team building AI-powered drawing tools for hand-drawn animation.
Worked on SpendIQ’s on-device receipt detection and MileIQ’s GPS-driven state machines.
I’ve spent time traveling and surfing in Panama, California, and Oregon. I studied economics with a computer science sequence at Claremont McKenna College, graduating in 2014.
I’d like to hear about engineering roles and early stage teams working on useful products. Send the role, what you’re building, and why you think we should talk.
joenewbry@gmail.com