How fast can we get this to $100 MRR?

Howdy guys, Jaryd from How They Grow here!

Welcome to my first commit message on this new writing project, The-Diff.

This is a build-in-public log where I’m going to share real commits, real decisions, real mistakes and real lessons as I try to build products from scratch as a non-technical solo builder using AI tools to do things I couldn't do a year ago.

It’s bonkers how much the game has changed and what a nano budget can do for you these days in terms of getting a thing from 0 to something.

If you're here from How They Grow, all my past deep dives taught me a lot about how great products get built. Now I want to actually do it for myself for something I own, and write about it as I go. Less deep analysis, more action and musings. Same product and growth lens obviously, but think about it more as personal little letters from a foot solider in the trenches vs a General writing some big report on how the war is going as an abstract from a safe office.

You get it.

Anyway, here's how I want this to go: I pick a product idea, give myself a tight window to ship it, and try to hit a specific revenue target. Along the way I'll share everything. The stack I went with, the tools giving me superpowers and a false (but lovely) sense of technical power, the product changes, the growth ideas, the numbers, the things that don't work. The things that do…we'll see!

The first self-inflicted challenge: $100 MRR in 15 days.

You don’t need me to tell you that’s not life-changing money. But it’s enough to prove something works and to learn whether an idea has legs. If it doesn't hit, I'll humbly share why, what I’d do different, and move on to the next one. These are mini products with mini timelines and mini goals.

No precious attachment. Build, break, debug, ship, learn, share, repeat.

Let's get into it.

I’m calling it Dateful.

For couples who love date nights but hate planning them, Dateful is your tasteful personal date night assistant that learns what you and your partner love, finds amazing things to do in your city, and plans everything for you—so you just show up.

I want to use Claude as the brain, running an agent-based date night planner that works over text and Whatsapp. You tell it your preferences as a couple, what you like, where you live, budget, dietary stuff, etc—and it does the rest. A few days before your date night it texts you a tailored plan. Restaurant picked, activity suggested, or whatever suits that couple. View and approve. Logistics handled. It checks in afterward to learn what you liked so it gets better over time.

The core insight: Julia and I want more date nights. She often says I go through periods of being super planny and finding stuff, then lulls. The pressure of finding restaurants or things to do makes be procrastinate too, I hate browsing to find stuff.

I don't lack the desire, but I do lack the planning energy sometimes. By the time we're both free and trying to figure out where to go, we're already tired and the default wins. Couch. Takeout. Netflix. The friction is the logistics of making it happen.

Can’t we just show up dammit!

Dateful removes (will remove?) planning from the equation entirely. It's like having a concierge who gets you both and texts you saying "here's what you're doing Friday."

This is not my best idea. This is not my second best idea. Nor third?

But it’s an idea. Not one powered by some calculated market opportunity. I haven't done extensive competitive analysis and discovered a gap. I have a personal problem, I want to build something for myself and my wife, and I want to try my hand at two things I've never built before: an agent-driven product and something that lives in messaging instead of an app.

Different type of channel. Different type of brain.

That's it. Sometimes the best reason to build something is that you want it to exist and can learn something from trying to make it so.

The goal for Day 0 was simple: shape the idea and get the boring-but-essential scaffolding done so I can start building the actual product on day 1.

1.Used Claude as a product partner. Before writing any code (in English 😈) , I spent time with Claude shaping the idea—working through the product spec, user flows, what the MVP needs vs what can wait. I also used it as a technical co-founder of sorts, thinking through stack and architecture and types of tradeoffs. This is a phase in vibe coding I think people skip too quickly. I think this because it’s a phase I’ve skipped completely…keyboard happy just machine gunning tokens while drunk on power.

Not smart.

Having a real conversation with AI about what to build before how to build it is one of the most useful things I've learned in the last few months.

Spend time here working to get a MD file, it’s the biggest ROI to give whatever model is going to code for you vital context.

~45 minutes

2.Bought the domain. dateful.chat — $9. The hallmark start of every successful product starts with GoDaddy, obviously….and every cobwebbed idea 😄

~5 minutes

3.Created the GitHub repo and connected to Cursor. github.com/jaryd-hermann/dateful. It's public, so you can follow along if you want. Pulled it to local and connected to Cursor which I already have installed, which is my favorite tool right now. Agent chat, IDE, inline terminal, parallel agents— all making idea happen while you type in English typos.

Seeing it believing if you haven’t tried it.

~10 minutes

4.Set up Supabase. Paid $10 for a micro instance on my existing account. Postgres database and edge functions. For all my solo projects, Supabase is my go-to backend. She’s fast to set up, does 90% of what you need out of the box, has auth systems, and it just lovely. This was just getting it ready and my API keys.

~5 minutes

5.Shipped a dead simple landing page. Nothing fancy. Just enough that if you click dateful.chat right now, you can join the waitlist. There's something motivating about having a real URL doing a real thing on day zero even if the thing is just collecting only my own email. 💰

Also, this was the first commit where I had Cursor make my landing page, connect it to Supabase, and get pushed to my repo.

Something now lives in the place I want it to exist.

~20 minutes

6.Connected domain to Vercel. Dev is running on localhost, landing page live on prod pointing to my domain. Two environments, both working. This kind of setup used to take me a full day to understand.

~3 minutes

7.Made a quick logo. Because I always have fun with branding stuff 🙂 Just used GPT to generate something based on some Google search inspo. Not permanent, but enough to make the landing page feel real instead of a template.

~20 minutes (least economical use of time lol)

The start isn't about the product. It’s about the start.

It's about removing every excuse not to start building the product you want tomorrow. The goal is to wake up and open your compuer with zero friction between you and the work.

My Day 0 checklist for any new project:

If you've been sitting on an idea…start?

Two hours. Under 20 bucks. Find time, and you might go to bed tonight with a live URL and a repo ready to go the next day.

See you after some more progress.
— Jaryd