Cross-Project vs. Inter-Project Task Taxonomy

A two-way split for categorizing AI work before deciding how to parallelize or isolate it:

Cross-project tasks — unrelated domains (e.g., your main job vs. a side project). These should never share context.
Inter-project tasks — isolated units of work within the same domain (e.g., different parts of one app, or separate client proposals). These are the candidates for safe parallelization via Step 2 – AI Builds in Parallel (Orchestrator Model).

The distinction answers a practical question: when climbing toward Step 2 – AI Builds in Parallel (Orchestrator Model) and running multiple agents at once, which tasks are safe to run side-by-side without their contexts bleeding into each other?

Apply: Put cross-project work into separate Claude Code projects/folders outright. Use the inter-project category to decide which same-domain tasks are safe candidates for running in parallel.

Implementation: Dedicated /projects Folder Structure

The presenter's own layout puts cross-project isolation into a single top-level /projects folder with one subfolder per domain — e.g. /buildpartner, /internal-os, /incubator-website, /clients — and selects the relevant project folder in Claude Code before starting any task.

Apply: Mirror this structure — one subfolder per distinct domain — and open the matching project folder in Claude Code before starting work, rather than relying on discipline to keep contexts separate.