Failures

Everyone shares their wins. Few share their losses. This is my anti-portfolio — a graveyard of projects that died, ideas that flopped, and lessons learned the hard way.

“Ever tried. Ever failed. No matter. Try again. Fail again. Fail better.” — Beckett

💀 The Graveyard

CodeMentor AI (2024)

☠️ Dead — 6 months, ~$8,000 invested

What it was: An AI-powered code review tool that would automatically suggest improvements, catch bugs, and enforce style guides.

What went wrong:

1. Timing. Started building right before GitHub Copilot and similar tools exploded. By the time I had an MVP, the market was flooded.

2. Differentiation. I thought “better prompts” would be enough. It wasn’t. The big players had distribution, trust, and unlimited compute.

3. Solo founder trap. Spent 4 months building before talking to users. Classic mistake. I knew better.

What I learned:

• Validate before building. I’ve read this in 100 blog posts. Still didn’t do it.
• If you’re competing with big tech, you need a fundamentally different approach, not incremental improvements.
• Time-to-market matters more in fast-moving spaces.

Would I do it again? No. The market is saturated. But I’m glad I learned LLM orchestration deeply — it’s useful for everything else now.

FitTrack Pro (2023)

☠️ Dead — 4 months, ~$3,000 invested

What it was: A fitness tracking app specifically for natural bodybuilders with periodization planning and fatigue management.

What went wrong:

1. Niche too small. “Natural bodybuilders who want periodization tracking” is maybe 10,000 people globally. Even with 10% market share, that’s 1,000 users at $10/month = $10,000 MRR ceiling.

2. Existing solutions good enough. Strong App and similar tools are “good enough” for 95% of users. My differentiation wasn’t worth the switching cost.

3. I got bored. Fitness apps are a grind. The feedback loop is slow. I lost motivation.

What I learned:

• Calculate the TAM before building. I literally teach people to do this and didn’t do it myself.
• “I would use this” ≠ “enough people would pay for this.”
• Motivation matters. Building in a space you care about isn’t enough if the daily work drains you.

Would I do it again? Yes, but differently. ExtremeFitness.app is a second attempt with a broader target and better positioning.

The Newsletter (2022)

☠️ Dead — 8 months, 47 issues, 234 subscribers

What it was: A weekly newsletter on “the intersection of philosophy and technology.”

What went wrong:

1. No unique angle. “Philosophy meets tech” is a crowded space. I didn’t have a distinctive voice yet.

2. Commitment mismatch. Weekly was too frequent. I started resenting it. Quality dropped. Readers noticed.

3. No monetization plan. I had vague ideas about “building an audience first.” 8 months in, I had 234 subscribers and no clear path to revenue.

What I learned:

• Frequency should match your natural rhythm. I write better monthly than weekly.
• “Build an audience” is not a business model. Know how you’ll monetize before you start.
• Voice takes time to develop. I’m a better writer now than I was then.

Would I do it again? I’m writing here now instead. Blog + occasional newsletter > weekly commitment I resent.

DevOps Consultancy Pivot (2021)

☠️ Dead — 3 months of trying

What it was: Attempted to pivot from backend development to DevOps/SRE consulting. Higher rates, growing market.

What went wrong:

1. I hated it. Turns out I like building things, not maintaining infrastructure. Who knew?

2. Credential gap. Companies wanted “10 years of Kubernetes experience” (which didn’t exist). I had skills but not the resume signal.

3. Wrong motivation. I was chasing money, not interest. Never works for me.

What I learned:

• Know yourself. I’m a builder, not an operator.
• Motivation matters more than market opportunity.
• Three months is a reasonable amount of time to realize something isn’t working.

Would I do it again? No. Staying in AI/ML backend was the right call.

Startup #1 (2019)

☠️ Dead — 14 months, 3 co-founders, ~$40,000 total invested

What it was: A B2B SaaS for restaurant inventory management. (I’m being vague intentionally.)

What went wrong:

1. Co-founder conflict. Three people, three different visions. We spent more time arguing than building.

2. Wrong market. Restaurants have razor-thin margins and hate paying for software. We needed enterprise, got SMB.

3. Slow sales cycle. B2B to restaurants means convincing busy owners who don’t check email. Six-month sales cycles killed us.

4. COVID. We were just getting traction when the pandemic hit. Restaurants stopped buying everything.

What I learned:

• Co-founder alignment is everything. Skills matter less than shared vision.
• Know your market’s willingness to pay before you build.
• Sometimes timing kills you and there’s nothing you can do.

Would I do it again? Not in restaurants. But the experience taught me more than any successful project could have.

⚰️ Smaller Deaths

Projects that didn’t even get far enough for a full post-mortem:

📉 Patterns I Keep Repeating

Looking at my failures, some patterns emerge:

1. Building Before Validating

I know I should talk to users first. I still don’t. The dopamine hit of coding is too strong.

2. Overestimating Market Size

“Millions of people must want this!” No. Thousands, maybe. And most won’t pay.

3. Underestimating Competition

I consistently believe I can out-execute incumbents. I usually can’t. Distribution wins.

4. Wrong Motivations

Projects started for money (not interest) always fail. I haven’t internalized this yet.

5. Solo Founder Everything

I struggle to delegate. This limits scale but also limits complexity. Mixed blessing.

🌱 What Failure Taught Me

Failure is information. Every dead project taught me something I couldn’t learn any other way.

Speed of failure matters. The 2-week failures cost almost nothing. The 14-month failure was expensive.

Shame is useless. I used to hide my failures. Now I publish them. Much healthier.

Successful people fail more. The correlation isn’t “they’re luckier.” It’s “they try more things.”

🔮 Currently At Risk

Projects that might end up on this page:

If you’ve failed at something and want to commiserate, email me. I’ll read every story. Failure is lonely. It doesn’t have to be.

Related

See also: My Numbers | What I Believe | What I’m Doing Now