I Vibe Coded 10+ Apps Used Almost a Million Times. Then I Had To Stop for 90 Days.

I deeply love vibe coding. I genuinely do.

Since summer 2025, I’ve shipped 10+ production apps using Replit and Claude Code in particular. Those apps have been used close to a million times. Pitch deck analyzers. Startup valuation tools. SaaStr.ai, which hit 500,000 users in its first 45 days and now processes hundreds of thousands of valuations every month.

That’s not a hobby. That’s a real product surface built by someone who cannot write a line of Python from scratch.

And then I shipped Founderscape.ai, a game that simulates running a startup, just for fun, and I stopped. Cold. 90 days ago.

I haven’t vibe coded a single thing since.

Why I Stopped

Not because the tools got worse. The opposite, actually.

Replit and Claude Code have become insanely capable especially since Opus 4.5 launched in late 2025. The gap between “what I can imagine” and “what I can actually ship” has collapsed in ways that felt impossible 18 months ago. The tools are better. My instincts for what to build and how to describe it have gotten sharper.

So why the break?

Honestly? Because it’s exhausting.

Two hours a day, every day, shipping production software as a non-developer while also running an eight-figure events and AI business, managing a venture fund, and trying to be present for the people who matter to me… it catches up with you.

Vibe coding is not passive. It’s not “describe something and watch it appear.” It’s relentless iteration, debugging sessions that spiral, product decisions made at 11pm, and a constant low-grade anxiety about whether the thing you shipped actually works. For someone without a traditional engineering background, every session requires full cognitive presence. You can’t coast.

I needed a break. So I took one.

What Happened While I Was Gone

Amelia, our Chief AI Officer at SaaStr, picked up the slack and then some. She shipped two apps that are genuinely among the most impactful things we’ve built.

The first is our AI VP of Marketing, an internal tool that now runs almost all of SaaStr’s marketing operations. Not assists. Runs. That’s not a casual claim. We went from a full marketing team to essentially AI-first execution, and it works.

The second is our Sponsor Portal, which manages the full operational relationship with our sponsors across what is now a $10M+ sponsorship business. Contracts, deliverables, communications, tracking, all of it flowing through a custom-built tool that didn’t exist six months ago.

Both of those would have taken a traditional development team months and real budget. Amelia shipped them while I was on my vibe coding sabbatical.

The lesson there is that vibe coding scales when more than one person on your team can do it. We have two builders now instead of one. That matters enormously.

The Real Question for Non-Developer Builders

I’m good at vibe coding. Better than I expected to be, and probably better than most non-developers who’ve tried it.

But I’m no longer confident I can sustain 2 hours a day, 365 days a year. Not anymore.

Part of that is capacity. I have a day job that demands a lot. Part of it is that the intensity of building production software, even with AI doing most of the actual writing, takes something out of you that doesn’t automatically replenish overnight.

There’s a version of this that’s genuinely addictive. You get a product shipped and used by real people, you feel the feedback loop, and you want to build the next thing immediately. That cycle is seductive. It’s also not sustainable at full speed indefinitely, at least not for me.

What I’ve landed on is this: vibe coding is not a daily habit for most non-developers. It’s a sprint capability. You can go deep for weeks, ship something meaningful, and then you need to recover. The developers I know who do this for a living have 10,000 hours of pattern recognition that makes the cognitive load lower. I don’t have that. Every build still costs me more energy than it would cost a real engineer.

That’s okay. The ROI is still extraordinary. But I want to be honest about the cost.

I’ll Be Back. I Almost Am

The 90-day break is coming to an end. I can feel it.

There are three or four things I want to build that I haven’t been able to stop thinking about. The ideas have been queuing up. That’s usually the signal.

When I come back, I’ll probably structure it differently. Fewer daily sessions, more focused two-week sprints with clear shipping goals. Less ambient tinkering, more intentional builds.

The tools will be even better by then. That’s the other thing about this space: every time you come back, the ceiling has moved up. What felt hard in October feels easy by March. That compounding improvement in the underlying platforms is the most underrated part of this whole story.

Vibe coding as a non-developer is one of the genuinely new capabilities that 2025 and 2026 have made available to founders and operators. It’s not a toy. The apps Amelia and I have shipped have meaningfully changed how SaaStr operates and how we serve our community.

But it will burn you out if you treat it like an infinite tap. Build hard, ship real things, and when you need a break, take one without guilt.

The tools will be waiting. They always get better while you’re gone.