The Box and the Module
Source: FitzyHistory Substack
Author: FitzyHistory
Original source: https://fitzyhistory.substack.com/p/the-box-and-the-module
Published: 2026-03-16
Source type: essay
Private backup: the full article text is archived in the private repository at archives/articles/fitzyhistory-substack-com-the-box-and-the-module.source.md. It is not published on the public Quartz site.
Summary
FitzyHistory argues that agentic AI shifts teacher AI use from asking for generated content toward designing functional learning experiences. Using the example of building an interactive Constitutional Convention module with Claude, the author contrasts one-shot chatbot use, which can enable cognitive offloading, with design-based AI work that requires upfront pedagogical judgment, audience awareness, sequencing, organization, and iterative evaluation. The article frames AI as an emerging design medium: non-coding teachers can now build bespoke educational tools, but the quality of those tools still depends on the quality of the teacher’s instructional thinking.
Pull quotes
Teaching knowledge drives design
“None of these decisions required coding knowledge. All of them required teaching knowledge.”
Design demands expertise
“The tool needs design specifications that reflect real expertise: what the activity is for, how information should be organized, what the user needs at each stage, and what problem the whole thing is solving.”
Quality depends on thinking
“The quality of what comes out still depends entirely on the quality of the thinking that goes in.”
Big ideas
- AI is changing what knowledge work asks people to do
- Students need to bring the purpose; AI should not supply it for them
- Learning still needs some struggle, even when AI can make things easier
- AI tools should be judged by the work they will actually do
- Schools should start with learning values before choosing AI tools
Claims
- Agentic AI can preserve thinking when students have to design the work
- AI can make a poorly framed problem worse
- AI changes how people come to know things, not just how fast they work
- Prompting AI is a literacy practice, not just a technical trick
Key evidence and examples
- The author compares making a wooden box in shop class with building an AI-assisted digital module.
- The Constitutional Convention simulation includes delegate tabs, issue positions, side-by-side comparison, embedded sources, and a live voting tracker.
- The author argues AI execution was only useful because the teacher specified classroom flow, student needs, bottlenecks, and pedagogical purpose.
- A kitchen renovation analogy illustrates why complex AI-assisted builds reward careful upfront planning and punish sloppy specification.
Education relevance
Highly relevant for teacher professional learning, AI-assisted curriculum design, agentic AI in classrooms, instructional design, assessment redesign, and non-programmers building bespoke learning tools.
Durability note
The tools named here will change quickly, but the core distinction between loose prompting and expert design work is a durable lens for teacher-facing AI.