Technical writing · AI enablement · Research & editorial judgment
Making sense of AI, technology, and the systems we build around them.
I'm a technical writer, editor, and former cognition researcher who has spent more than a decade investigating complex systems and making them understandable.
My work spans developer documentation, knowledge architecture, and AI evaluation. I also write independently about AI, expertise, and human judgment, particularly the gap between what systems appear capable of doing and what their outputs actually establish.
This site brings together that writing, the conceptual models behind it, and practical tools for examining how AI changes work.
Choose your path
What brought you here?
I am exploring the ideas
Begin with a problem you recognize, then follow it into the writing and conceptual models that explain what is happening.
Start with a problem →I am diagnosing an organizational problem
Use the diagnostic tools to examine judgment, expertise, validation, and hidden dependencies inside an existing system.
Explore the diagnostic tools →I am evaluating Grace's work
Start with the essays, then review the experience and research background behind them.
Why this work exists
AI is making assumptions visible that organizations could previously leave implicit.
Technical systems have always depended on human interpretation, expert judgment, and contextual knowledge. As AI begins producing, retrieving, evaluating, and acting on technical information, those hidden dependencies become operational problems.
My work examines where those problems emerge, what mechanisms produce them, and what organizations need to diagnose before they can respond effectively.
Explore the body of work
Three ways into the same problem space
The writing identifies emerging patterns. The conceptual models explain why they occur. The diagnostic tools help organizations examine how those patterns are operating inside their own systems.
Writing
Essays that identify changes in expertise, trust, technical communication, and AI-assisted work.
- Why did better AI increase dependence on experts?
- Why can accurate documentation still fail?
- Why does plausible output change evaluation?
Conceptual Models
Reusable models that explain recurring mechanisms across documentation, AI adoption, governance, and organizational design.
- Expertise Relocation
- Coherence vs Correctness
- Machine-Consumed Infrastructure
Diagnostic Tools
Practical diagnostics for organizations examining where judgment, expertise, validation, and responsibility currently live.
- Locate where judgment is actually being performed
- Expose hidden dependencies and unsupported assumptions
- Identify where deeper investigation is needed
Grounded in research and practice
The ideas come from building the systems they examine.
This work combines more than a decade designing technical knowledge systems with formal research into cognition, learning, and the development of intelligence.
Documentation leadership
Built and led documentation functions supporting complex developer products, infrastructure, and APIs.
Cognition research
Earned a PhD in Comparative and Developmental Cognition studying how intelligence develops, generalizes, and becomes visible through behavior.
Knowledge architecture
Designed information architectures, developer portals, and documentation systems around how different audiences understand and use technical products.
AI-ready systems
Redesigned technical knowledge for machine retrieval and AI-assisted use, including improving an independent AI-readiness assessment from 65% to 92%.
Featured work
Start with one idea
AI, Expertise, and the Appearance of Ability
Why AI becoming better at appearing capable changes the work required to evaluate it.
Read the essay → Conceptual ModelPlausibility Anchor
How plausible output begins shaping evaluation before verification is complete.
Explore the model → Diagnostic toolJudgment Allocation Audit
Examine where judgment is being performed today and where the organization is merely assuming it exists.
Use the tool →