Three plays before lunch — interviews, cold email, and the file you're not keeping

This is a sample issue, written to show the format before the first real send. Nothing has gone out yet.

Morning.

Short one this week — I spent most of it heads-down shipping, which is the correct state for a founder and a terrible state for a newsletter writer. Three things, all of which I actually used in the last seven days.

1. One thing to try — turn your user interviews into a ranked feature list

If you’ve got interview transcripts rotting in a folder, this is the week to use them. Hand Claude all of them at once and make it count and rank the requests — with the exact quotes cited so you can trust it. I ran ours and it surfaced a dealbreaker one user mentioned in passing that I’d completely filed under “nice to have.” It wasn’t.

→ Full workflow, with the exact prompt: Turn ten user interviews into a ranked feature list

2. One thing to understand — give it the numbers, not a description of the numbers

The most common way founders get bad AI output: they describe their situation instead of handing over the actual thing. You ask for an investor update and describe your metrics in a sentence. Of course the draft is vague — you gave it a vague input. Paste the actual spreadsheet. Attach the real thread. The model is only as good as the slice of reality you give it, and most people give it a sliver.

3. One thing worth reading — the file you’re not keeping

This week it’s one of ours: what a memory file is, and why you should keep one. The setup it describes is almost aggressively simple: one folder, four plain text files, no framework. It’s a good antidote if you’ve been feeling like you need to build some elaborate agent rig before you can start. You don’t. Start with the file and the habit. The machinery comes later, and it’s smaller than you think.

That’s it. Go ship something.

Jamey

P.S. If a friend would get something out of this, forward it. That’s the whole growth strategy. Subscribe for anyone it got forwarded to.