I got laid off in September. My fiancée and I had just bought a house. We had a wedding to plan.
Instead of job hunting, I started building software.
I'd spent the previous year learning to code with AI while working full time — slowly enough that I wasn't sure it was going anywhere, then suddenly fast enough that I couldn't stop. By the time the layoff happened, I had already built a working web app and knew this was the next chapter.
Before that I'd spent six years as a product manager. I loved most of it: talking to customers, discovering what actually hurt, collaborating with engineers to build something real. What I didn't love was waiting. Waiting on sprint ceremonies, ticket refining, retros. Waiting on engineers to build the thing I'd already developed in my head. I'd always assumed that if I ever wanted to build software on my own, I'd need a technical co-founder. AI threw that assumption out the window.
My first app after the layoff took five months. Not because the code was hard — because everything around the code was hard. Picking a name, then picking a new name after a trademark conflict. Setting up developer accounts. Building marketing pages. Getting rejected four times before MEMO (a relationship manager for real estate agents) finally shipped.
The second app was supposed to be a distraction I couldn't afford. I built the first version of Loop (a recurring reminders app) in a weekend, shipped it to friends on TestFlight, and told myself I'd come back to it after MEMO shipped. I did come back, eventually. By then I'd started building a THIRD app, this time in Swift instead of React Native, and the experience was different enough that I scrapped the Loop codebase and rewrote the whole thing from scratch, plus added widgets, lock screen complications, and a task library. The rewrite took a few sessions. Last week I put it on the App Store for $2.99.
I stopped thinking of myself as "the guy building MEMO" and started thinking about what it means to just keep building things. Small, focused apps that do one thing well, shipped by one person who actually uses them.
I've been posting about my journey on LinkedIn for five months. Solid engagement, zero follower growth — the audience was my existing network, not new people finding me. The build-in-public community I actually want to be part of lives here.
I'm about to ship Temp, my third app. I built my wife a personalized workout app for Valentine's Day that'll probably be fourth. A year ago I was a PM who'd never shipped anything. Now I'm shipping things I couldn't have imagined building six months ago. Your baseline quietly relocates while you're not looking.
I'll write about what I'm building, what broke, and what I'm figuring out as I go. Probably some Chiefs takes too.
@asmartbear For whatever reason I'm more receptive to tough feedback from Claude (less ego about how the other person perceives me, I guess), so I have my blindspots and the instructions to call them out baked right into my claude md
YSK: You can update your App Store screenshots without submitting a new version and skip the review process.
Create a product page optimization test, upload your new screenshots, launch the test, then once it starts running, click the name of the test, scroll to your new treatment, and then click 'Apply Treatment to Current Product Page.'
yw
@VirtualMaestro@felixrieseberg I wrote up a quick start guide, get your project folder set up and then have CC move existing docs from your repo into it for you.
Most people will bookmark this and never build it. If you're a solo founder, there's a version of this that solves a different problem, and if you're already paying for Claude, you have it. Setup takes five minutes.
The Setup
/your-working-folder
CLAUDE.md
context/
business.md ← who you are, what you're building
priorities.md ← this week, parking lot, today's focus
CLAUDE.md does one job: tell Claude what to read before every session and what to update after.
`Every session: read context/business.md and context/priorities.md before responding.
End of any session with real progress: update priorities.md. Note significant decisions in business.md.`
Start there, then let Cowork build it out. The business.md and priorities.md docs don't have to be fully fleshed out; in your first session, you can prompt it to ask you questions about your business, current priorities, etc. Mine now has sections on how to work with me specifically: my working style, what to flag, how to structure sessions. Cowork wrote and maintains all of it.
Daily routine
Every morning, open Cowork, say "let's plan today." It reads your files, checks your calendar, tells you where to point your energy. At the end of the day, tell it to update the docs.
The sessions stack. Context builds on itself. Every conversation starts from where the last one ended. Over time it becomes the brain of your business and self-updating LLM wiki/knowledge base. No more stuffing your code repos with your marketing and business docs for CC to wade through.
I run this across a portfolio of 3 apps. Each has its own subfolder with context, plans, specs, images, whatever's relevant. Any time I have an idea I want to capture some context about while it's fresh gets a folder. No more random Apple notes with half-baked ideas. I throw reference articles, design inspo, anything I'll want during the actual specing and scope shaping phase before hading over to Claude Code.
My working folder has grown a lot of branches, all created by Cowork, but it all started with those three files.
If you're already paying for Claude, it's worth trying. I've been doing this for about two weeks (since they released projects) and it's been a revelation. As my app portfolio grows, these so much mental bandwidth I've been able to free up having an agent that hold the entire context of my business and help me decide how to prioritize my time and plan my day.