The App Store Prints. Your Rental Doesn't — Amplifye Signal #003

👋 WELCOME BACK TO THE SIGNAL

Issue 003. Two things I stand by. First — the math nobody talks about: why a few hundred app subscribers beats a cash-flowing rental property. Second — a full breakdown of vibe coding, the tools to use at every level, and the exact prompts to turn an idea into a working prototype. This one's actionable. You should be building something by the weekend.

Plus — we tried writing this newsletter with AI agents. Here's what actually worked and what flopped.

No fluff. Just signal.

⏱️ THIS WEEK AT A GLANCE

Read time: 6 minutes

I own real estate. I collect rent. Which means I've also gotten the calls — "the basement is flooded" and the very typical "the heat isn't working." Idk who glamorized landlording, because they left out that one semi-bad repair costs you 2–3 months of cash flow. I still think investing in a secondary market makes sense where the numbers are smaller and the risk is manageable. I can't imagine taking on the risk in a big city where one vacancy is thousands a month — like what my current landlord would be dealing with if I moved. That would be terrifying.

So what does a modest deal actually look like? You're not buying in Miami or LA or NYC. You're buying in one of those middle-America cities nobody really talks about — solid 3 bed, 2 bath for $100–150K. Put 20–30K down. Mortgage around $640–$660/month at today's rates. Tenant comes in at $1,400–$1,600. After everything — mortgage, insurance, taxes, maintenance — you're clearing maybe $400–$600/month.

Want to go bigger? A duplex or triplex in a similar market, $300–500K. Two units at $1,400 each is $2,800 gross. After the mortgage and carrying costs you're netting $500–$800 — similar margins, but now you're building equity on a bigger asset. It's scalable, you just need capital.

There is a way to get in at a lower entry point. You can house-hack with an FHA loan — 3.5% down on up to a 4-unit, live in one, rent the rest. Lowest barrier entry in real estate. But two things that'll bite you if you're not paying attention. One — anything under 20% down means PMI on top of your mortgage payment. On a conventional loan that's 0.5–1.5% of the loan per year — on a $300K property you're adding $125–$375/month, which eats the cash flow. Two — FHA ties you to that property as your primary residence for at least a year. Do your own research on that. Personally I went the seller financing route, but that takes hustle to find deals.

Regardless, real estate cash flow is real. I'm not here to talk you out of it. But it's slow, it comes with phone calls you don't want, and you can only move as fast as the market allows. I got tired of waiting for rates to come down. So I started launching apps.

Here's the math that changed my thinking. If I could get 6 apps to $500/month MRR (monthly recurring revenue) — that's $3,000/month. No tenants. No repair bills. No waiting on a rate cut. And I can ship another app in 3 months. Try buying a property every 3 months.

At $9.99/month, the App Store takes 15% — about $1.50. Server costs and RevenueCat overhead run maybe $0.20. You're keeping $8.30 per subscriber — 83 cents on the dollar. 100 subscribers is $830/month. Same as that rental, without the risk.

The rental has a ceiling. One house. One tenant.

The app has billions of potential users. Getting to 1,000 subscribers means convincing 0.000001% of the planet to spend less than a coffee a month on something you built. At 10,000 subs you're clearing $83,000/month — and by then you're operating a machine, and these days one viral post and you're there. Which is why I don't understand why more influencers don't use this as their main monetization vehicle. Convert 10% of your following for 3–4 months and you do the math.

I ran my first ads, went to sleep, and woke up to 10–20 free trial notifications. Not paid — free trials. Some converted, some didn't. But I'll tell you: I've got a new found love for people who forget to cancel their subscriptions.

My buddy Khalil is proof of this same model. He's the founder of Deduct AI — an expense tracker that helps business owners find write-offs. We're in a group chat together called "Build & Ship." When we were in New York in late 2025 he was still deep in the grind — ads, retention, all of it, still figuring out the how. Not to say he's got it all figured out now — but recently he shared screenshots with a screen full of RevenueCat notifications. New subscriber. Free trial started. Yearly sub locked in. In a year and a some change he's managed to get on pace to pass $10–11K MRR. We'll give Deduct its own full spotlight soon — but that number says everything.

How many properties do you need to gross $11K a month? How much capital? How many tenants? You do the math.

You build it once. You refine it. And when it's working — there's no phone call. Just notifications and deposits.

What is vibe coding?

You describe what you want to build in plain English — an app, a tool, a dashboard — and AI builds it. You don't write code. You direct it, review it, describe what needs to change. In early 2025, Andrej Karpathy (OpenAI co-founder, ex-Tesla AI) posted that he'd been building apps by fully giving in to AI without reading a single line of code. He called it vibe coding. By end of 2025 Collins Dictionary named it word of the year. 25% of YC's Winter 2025 batch had codebases that were 95% AI-generated. This is real.

Why it matters:

For non-technical founders, this is the unlock — your imagination can finally run ahead of your resources. And something interesting happens when you actually vibe code a prototype: you get it. You understand the tradeoffs. You feel where the hard parts are. You become a way better collaborator when it's time to bring in real technical support. For technical founders — solving the problem is one thing. Typing thousands of lines of code to implement it is another. AI handles the typing. The thinking is still yours.

The How : Before you open any tool — validate first.

Step 1 — Flush out the idea. Dump it into Claude or ChatGPT. Use the prompt from Issue 002: "I'm about to build [this thing]. What are all the things I need to know, watch out for, and make sure I've covered?" You'll be surprised how much gray area clears up.

Step 2 — Define features. Get the LLM to prioritize everything — Must Have (MVP), Should Have (v1.1), Nice to Have (future). By the time you're done you have a product spec and you've probably killed 3 bad ideas and sharpened 1 good one.

Step 3 — Build the scaffold. Open a tool below and give it that spec. You're not saying "build me an app" — you're saying "here's exactly what I need." Iterate feature by feature from there.

That's the loop: idea → validate → define → scaffold → prototype.

The 4 tools you need to know:

Lovable (Free + $25/mo Pro) — Start here if you know nothing. Describe what you want, it generates a full-stack web app — frontend, backend, database, auth — and deploys it. Founders have gone from idea to live product in under 4 hours. Best for: landing pages, SaaS dashboards, MVPs. Limitation: web only, gets tricky with complex logic.

Cursor (Free + $20/mo Pro) — AI-native code editor, VS Code with AI woven into everything. What engineers use to go faster. Honest take: you need to be able to read code. Can't? Start with Lovable. Once you can navigate a codebase, Cursor is incredibly powerful.

Replit (Free + $25/mo Core) — Browser-based, no downloads, no setup. Everything in one tab — IDE, database, hosting, deployment. 40 million users. Their Agent builds, tests, and deploys autonomously from a natural language prompt. Watch billing — heavy Agent use can creep past $100/month fast.

Firebase Studio (Free — firebase.studio) — Google's play, live now. Gemini 2.5 Pro built in, no API key, browser-based. I've used it — genuinely impressive, especially for AI-native apps inside the Google/Firebase ecosystem. The real differentiator: backends, frontends, AND mobile apps with built-in Android emulators so you can preview on a real phone layout without leaving the browser. Lovable and Replit are web-first — if you're thinking mobile, Firebase Studio is the move. Requires slightly more technical comfort, but it's free and powerful.

Know nothing → start with Lovable. Ready to go deeper → Cursor or Replit.

⚡ THE POWER COMBO FOR SERIOUS BUILDERS

Once you've outgrown point-and-click — Cursor + Claude Code (or OpenAI Codex if you're in that ecosystem). Cursor handles the IDE and multi-file edits. Claude Code runs from the terminal as an autonomous agent — reads your entire repo, writes and executes code, runs tests, fixes its own errors. You're directing an agent that understands your whole project. Codex from OpenAI works the same principle.

The difference from Lovable or Replit is control and depth. Lovable builds fast and clean. Cursor + Claude Code builds anything, at production quality, at a pace that would've required a full dev team two years ago. Navigate a terminal, understand your code at a high level — the ceiling is basically gone.

The challenge for this weekend: Pick one idea. One problem in your own life or business. Open Lovable, describe it in plain English, spend a few hours. Send us what you built — reply to this email, DM us on IG, drop it in the comments. We want to see it.

Padel is the fastest growing sport in the world. 100,000 active U.S. players today, projected 15 million by 2030. Qatar Sports Investments, Mubadala, Ronaldo, Messi, Nadal — the capital is already moving. Miami is the epicenter.

I play padel. The real problem isn't finding courts — it's coordination. You get 5 texts on a day you can't play, the thread dies, someone scrambles for a fourth, and the match doesn't happen. Every time.

PadelMatch fixes the underlying issue. Mark your free windows once. Friends with overlapping availability surface automatically. One tap creates the match. No back-and-forth. No dead threads. Just padel.

The club math is simple. Miami: 12 clubs, 108 courts, 30% utilization. One extra booking per court per day is a $6M annual revenue lift. No new courts. No new capital. Solve the player problem and you unlock the club revenue problem automatically.

Founder Peter Menakis built a 150+ player Miami padel community before he built the product. MVP is live and getting real user feedback now. If you play — grab your spot or take the quick survey.

If you're an investor — [email protected].

We told you in Issue 002 we'd try it. Here's the truth.

Gave the AI the topics, some context, said: research this, write the sections. What came back was statistically sound, well-organized, completely soulless. Every stat cited. Every paragraph technically correct. But none of it — none of the flooded basement call, the 20 free trial notifications, Khalil's RevenueCat screenshots, the "idk who glamorized landlording" — none of the stuff that makes this worth reading actually came from the AI. It couldn't have. It doesn't know what any of that felt like.

That's not a knock on AI. It can research, structure, and draft. It cannot replace lived experience.

What DID work: talking it out. Voice recorder on, brain dump, feed the transcript to an LLM and say "shape this into something structured." Human experience goes in first. AI helps form and edit. That's the workflow — and honestly, that's how most of this issue got made.

The newsletter stays human. The tooling is real — and we'll keep experimenting.

If this was useful — forward it to one person who should be reading it. That's how we grow.

Have a deal, a pitch, or a story worth featuring? → [email protected]

And seriously — vibe code something this weekend and let us know what you built.

See you next week. ✌️

— Team Amplifye