The problem
What was broken before AI
As AI-generated plans get longer, Markdown starts to break down as a collaboration format. A thousand-line plan may be technically complete, but it is hard for a human to read, judge, and edit. The result is that people stop engaging with the plan and start delegating more decisions back to the model. That weakens the human feedback loop exactly when the project needs better judgment.
What changed
What the use case made possible
Thariq began asking Claude Code to create HTML artifacts for planning and collaboration. A brainstorm becomes a visual page with cards, mockups, and risk assessments. An implementation plan becomes a single-file website with code snippets, file structure, UI references, and context. When part of the plan is hard to edit, he asks Claude to build a small custom interface just for that problem. For design systems, he creates HTML files that both people and models can read and reuse.
Why this matters
Why this use case is worth studying
Thariq’s workflow is interesting because it treats the collaboration medium as part of the AI system. The model may be able to read Markdown just fine, but the human still needs to understand what is happening and make good decisions. HTML makes the work easier to scan, question, edit, and share. That keeps the human in the loop in a more active way instead of turning them into a passive reviewer of model output.
Use this when
When this pattern applies
Use this pattern when AI is generating work that is too long, abstract, or dense to review comfortably as text. It works especially well for implementation plans, product brainstorms, design systems, decision rules, and anything where seeing the structure makes it easier to make a good decision.
