Throughout my MBA, I've tried every organizational method I could find. Apple Notes, Notion, various folder structures. After eight months and three quarters, I settled on something almost embarrassingly simple: a plain directory tree synced via iCloud.
School/
├── 01 Fall 2023/
│ ├── Class 1/
│ │ ├── Week 1/
│ │ ├── Week 2/
│ │ └── ...
│ └── Class 2/
│ ├── Week 1/
│ ├── Week 2/
│ └── ...
├── 02 Winter 2024/
│ └── [Similar structure]
└── 99 Admin/
├── Scholarships/
├── Financing/
└── Career Dev/No proprietary formats, no learning curve. Just folders and files.
The Context Problem
I use AI constantly for coursework. Tutoring, writing feedback, working through concepts. But there was a persistent friction: every time I started a session, I had to rebuild the context. Open the chat, find the files, copy the relevant sections, explain what class this was for, remind it what we covered last week. Any single instance was trivial. Multiplied across hundreds of sessions, it became a real tax.
The Connection
On the side, I'd been teaching myself to code. I went through the whole tooling journey: V0, Replit, Copilot, Cursor, Windsurf, Claude Code. I landed on Claude Code because it has full visibility into the project directory. It reads the file structure, understands how things relate to each other, and carries context across sessions. That visibility is what made it useful for coding.
Three quarters in, I noticed something: my MBA folder structure was already organized the same way a codebase is. A hierarchy of directories, files of different types, a consistent naming convention. The only difference was that I wasn't opening it in a tool designed to work with structured directories.
So I opened my School folder in VSCode and ran Claude Code against it.
What Changed
The context problem went away. When I needed help with a private equity case, Claude could reference the LBO model from Week 3, the valuation framework from a previous class, and the industry analysis from Strategy. No copy-pasting. No re-explaining.
VSCode also turned out to handle non-code files better than I expected. PDFs, Word docs, PowerPoints, Excel files. Instead of switching between four applications and a browser, I had one workspace with everything in it.
But the part I keep thinking about is the cross-referencing. When concepts from Corporate Finance connected to something from Operations Management, the AI noticed it because it could see both course directories. I'm not sure I would have made those connections on my own, or at least not as quickly. The structured directory made the knowledge searchable in a way that a pile of files in different apps never was.
The Question I'm Sitting With
I keep coming back to this: the tools developers built for managing codebases turn out to work for managing any structured collection of knowledge. Version control, directory-based organization, AI assistants that read your file tree. None of this was designed for education, but it fits.
That raises a question I don't have a clean answer to. Is the distinction between "technical" and "non-technical" work actually meaningful when the underlying problem is the same? Organizing information, navigating complexity, making connections across a large body of material. Developers have been doing this for decades with tools specifically built for it. The rest of us have been using folders and keyword search.
I'm not claiming this is a revolution. It's a folder opened in a different program. But the change in how I work with that folder has been large enough to make me wonder what else I've been approaching with the wrong tools simply because the right ones were labeled for a different audience.
— Lo