25 Years of Building
Failures, Side Projects, and Why I Keep Going
Full disclosure: This post was written by a human (me), polished by an AI (it fixed my grammar and made me sound smarter), then reviewed by me again (to make sure the AI didn't make me sound too smart). Any remaining errors are 100% organic, artisanal, human-made mistakes.
The startup journey is hard. Side projects are hard. My 25 years in software is riddled with failures—or maybe you could call them small successes that didn't quite land. I've enjoyed every part of it, but it's also incredibly frustrating. It makes you question yourself constantly.
Every day you wake up thinking: Should I give up? Should I keep going? Am I wasting my time? My money? My energy? But something keeps me going. And especially in the last couple of years with AI, it's been an incredible multiplier for my capacity to build and ship ideas.
The Never-Ending List of Ideas
I'm sure many of you can relate: there's an ever-growing list of ideas you want to build. Products you'd use yourself. Problems you've personally experienced and want to solve. But the effort and time it takes to build something meaningful? It always felt impossible alongside work commitments and family responsibilities.
For years, those ideas just sat in a notes app, collecting dust. I'd occasionally revisit them, feel a spark of excitement, then reality would set in: "When would I even find the time?"
The gap between having an idea and shipping a product felt insurmountable. Not because I lacked the skills—but because I lacked the hours in the day.
Then AI changed everything.
AI as a 10x Multiplier
In the last two years, because of AI, I've been able to build almost 15 different products. Open source projects. SaaS apps. Tools I use every day. Things that would have taken months now take weeks. Things that would have taken weeks now take days.
Some highlights:
Curated Letters
AI-powered newsletter curation and summarization
Fluid Calendar
Smart calendar with AI scheduling
FamilyGPT
Safe AI assistant for families with parental controls
StoryFlow
AI-powered book writing from concept to publication
Plus open source projects that have earned 700+ stars on GitHub. Each one scratching an itch, solving a problem I personally had.
The Podcast Dream—Now Automated
Another thing I've always wanted to do: create a podcast. But podcasting is a massive time commitment. Research, scripting, recording, editing, publishing—it's easily 4+ hours per episode for a daily show.
Thanks to AI, I built a system that automates the entire pipeline. Now I run three daily podcasts, all published on Spotify, Apple Podcasts, and YouTube:
That's 33+ podcast episodes so far. I actually listen to all of them daily because I built them to solve my own problem: staying up to date with the news without scrolling through countless headlines and articles. They keep me informed, and hopefully others find them useful too.
Coming soon: AutoPodcaster.io — I'm turning this into a SaaS product. You'll be able to create your own podcasts on any topic, put them on autopilot, or customize every aspect: write your own scripts, use AI to generate them, record your own voice, or use AI voices. Whatever level of automation you're comfortable with.
Current Focus: StoryFlow
Right now, my main focus is StoryFlow—the AI-powered book writing app. I think it's really cool, and I'm using it myself to write stories for my kids. I've already published two books using it (see my previous post about why I built it).
But I'm struggling with how to monetize it. I've been thinking about several approaches:
Option 1: One-Time Purchase
Sell the software as a one-time fee. Users bring their own Gemini API key and have full control.
Option 2: Free + Platform Fees
Give the software away for free, charge to publish on the StoryFlow Bookstore, or take a transaction fee from book sales.
Option 3: Subscription
Monthly subscription that includes AI credits, premium features, and publishing benefits.
There are multiple paths forward. I'm genuinely curious to hear what you think—if you've used or considered using a tool like StoryFlow, how would you prefer to pay for it?
Why Do I Keep Going?
Honestly? I don't have a clean answer. Maybe it's stubbornness. Maybe it's the dopamine hit of shipping something new. Maybe it's the hope that one of these projects will finally break through.
What I do know is that AI has fundamentally changed the equation. The time and effort required to go from idea to shipped product has collapsed. Ideas that would have stayed in my notes forever are now live, being used, generating feedback.
25 years of building. Countless failures. Some small wins. And I'm still here, still shipping, still questioning whether I should keep going.
I guess that's the answer: I keep going because I can't imagine doing anything else.
What About You?
If you're on a similar journey—building side projects, questioning yourself daily, but still pushing forward—I'd love to hear from you. What keeps you going? What are you building? How has AI changed your process?
And if you have thoughts on how to monetize StoryFlow, I'm genuinely all ears. Sometimes the best insights come from the community, not from staring at spreadsheets.
Check out what I've been building
Related Articles
Why I Built StoryFlow: From Published Author to AI-Powered Book Writing
After 25 years in software and publishing a book the hard way, I built StoryFlow to transform how stories are written. Here's my journey from struggling author to building an AI-powered writing tool.
Vibe Coding vs Vibe Engineering: The Difference Between Slop and Software Excellence
Explore the critical distinction between vibe coding—which produces superficially working applications—and vibe engineering, which creates fantastic software backed by real engineering principles, foundational work, and performance optimization.
The Pieter Levels Model: One Developer, Simple Tools, Millions in Revenue
PHP, jQuery, SQLite powering million-dollar businesses. Why 'boring' technology wins, and what solo developer success stories teach us about the myth that you need modern frameworks to build real products.