- 🔭 I’m currently working on design systems, prototypes exploring AI, evidence, and human judgment.
- 🌱 I’m learning human-computer interaction (HCI), human-AI interaction (HAI), cognitive science, AI systems, and how to build with AI coding agents. I’m experimenting with OpenCode + Big Pickle in a zero-cost development environment, (Not claiming to be a design engineer).
- 🏗️ I’m looking to collaborate on interesting problems where AI can support better thinking, not just generate more stuff.
- 🐇 I like working on hard problems and have many questions, but I also make easy and simple things.
- 🌐 For the professional persona and orthodoxy, see my LinkedIn; for the ritual proof that I am a product designer, my portfolio; and my Substack, where I cosplay as an intellectual.
Pinned Loading
-
decision-router
decision-router PublicComposed LLM pipelines with typed authority boundaries: deterministic code decides, a bounded judge interprets, an LLM only writes prose. Includes an eval harness comparing composed vs. direct-LLM …
TypeScript
-
gettablox/tablox
gettablox/tablox PublicA minimal browser signal showing how much browser context is currently open.
JavaScript
-
sei-sailor
sei-sailor PublicInteractive graph of the Sei design system's token dependencies. Parses tokens.css and renders its primitive → semantic → component var() relationships, with search, layer filters and focus highlig…
TypeScript
-
epistemic-resilience
epistemic-resilience PublicA manifesto for epistemic resilience — our ability to produce information has outgrown our ability to decide what deserves believing. Ten principles, each rendered as an animated scene.
CSS
-
claimcheck
claimcheck PublicOpen-source, model-agnostic protocol for auditing claims and AI-generated critiques against a declared evidence boundary. Outputs a structured, auditable record — with null results as a first-class…
Ruby
-
radr
radr PublicA personal information retrieval system that surfaces meaningful changes in a set of topics over time. Research project testing — not assuming — whether an LLM is the best retrieval mechanism; eval…
Python
If the problem persists, check the GitHub status page or contact support.
