Run It in /
Claude Code isn't a tool, it's a different relationship with your computer.
Point it at a folder and talk to it. The right use case turned out to be all use cases. The surface area being infinite and the interface being just saying what you want — that's the product now.
Point it at a folder full of whatever — scanned receipts, half-finished slide decks, a decade of photos, raw JSON dumps, random PDFs — and just talk to it. That's it. That's the whole move.
It'll make slideshows. Write proposals that sound like you, not like a cover letter generator having a stroke. Organize the chaos into something you can actually navigate. Build little custom software to analyze the thing. Generate charts. Generate images. Lay it all out. The surface area is basically infinite and the interface is just — saying what you want.
People keep asking what the right use case is. There is no wrong use case. That's the whole point everyone keeps missing.
The mental model that finally clicked for me: Claude Code in the terminal is just a way to talk to your computer directly. Not through apps, not through menus, not through whatever abstraction layer some PM decided you needed — directly. The computer does the thing. You said the thing. That's the whole loop.
Run it in /. Give it everything. I'd trust it with my life, which is either a completely unhinged thing to say about a language model or the most sane possible response to watching it actually work.
Probably both.
Counterpoints
Push back, extend the argument, or sharpen it. New counterpoints go through review before they show up here.
No approved counterpoints yet.