Last month, I shipped two macOS apps. Two actual products that people are using, not two more half-finished repos collecting dust in ~/projects.
Six months ago? I was the king of unfinished side projects. I had 14 repos with names like cool-idea-v2 and definitely-shipping-this-one. None of them ever saw a user.
What changed wasn't motivation, discipline, or some productivity framework. It was recognizing that my biggest enemies were invisible: the feeds I scrolled and the money I wasted.
Here's what I learned.
The Two Problems Nobody Talks About
Problem 1: Your Feed is Eating Your Focus
I used to think I had a discipline problem. Turns out, I had a design problem — the apps on my Mac were literally engineered to pull me back.
I'd open Twitter "for a second" to share a build update. 45 minutes later, I'm deep in a thread about whether Rust is better than Go. My code editor is still open behind the browser. My context is destroyed.
The real insight: you don't need to delete social media. You need to remove the algorithmic feed while keeping the useful parts.
I started using Monk Mode to block feeds at the content level — not the whole app, just the infinite scroll portions. I can still post, DM, and search. I just can't doomscroll. The difference was immediate: I went from 3-4 hours of "productive distraction" per day to maybe 20 minutes.
The key: traditional website blockers are too blunt. Block Twitter entirely and you can't share your launch. Block nothing and you'll lose two hours to engagement bait. Feed-level blocking is the sweet spot.
Problem 2: You're Bleeding Money on AI Without Knowing It
The second thing killing my side projects was more subtle. I was spending way more on AI coding tools than I realized, and the anxiety about costs was making me hesitant to use them effectively.
I use Claude Code and Cursor daily. Some days I'm mostly on Sonnet, some days I need Opus for complex architecture decisions. But I had no idea which sessions were expensive and which were cheap.
Once I started tracking per-request token costs with TokenBar sitting in my menu bar, everything changed:
- I discovered that 80% of my spend came from 20% of my requests
- Simple refactoring tasks I was throwing at Opus worked just fine on Sonnet (at 1/15th the cost)
- My "cheap" coding days were actually expensive because I was using inefficient prompts that required 3-4 follow-ups
The result: same output quality, 40% lower monthly AI spend. That $200+ I was burning became $120 for the same work.
The Shipping Stack That Actually Works
Here's my actual daily setup as a solo dev:
Morning (Focus Block)
- Feed blocker ON — No algorithmic content until after lunch
- Token counter visible — Keeps me cost-conscious without being anxious
- One task — Pick the single most important thing. Do that.
Afternoon (Build & Ship)
- AI coding with awareness — Use the right model for the right task
- Ship something — Even if it's small. A bug fix. A docs update. Something.
- Feed blocker stays on — The afternoon scroll is even more dangerous than the morning one
Evening (Engage)
- Feeds unlock — Now I can scroll, but after a productive day it feels earned
- Share what I built — Post the update, respond to users
- Review costs — Quick glance at daily AI spend
The Numbers
Before this system:
- Side projects shipped in 6 months: 0
- Average daily screen time on feeds: 3.2 hours
- Monthly AI spend: $220+ with unclear ROI
After:
- Side projects shipped in the last month: 2
- Average daily screen time on feeds: 35 minutes
- Monthly AI spend: $125 with clear understanding of where every dollar goes
What I Wish Someone Had Told Me
You don't lack discipline — you lack friction. Add friction to distractions, remove friction from work.
Invisible costs are the worst costs. Whether it's time lost to feeds or money lost to inefficient AI usage, you can't fix what you can't see.
The tools exist. Feed-level blockers like Monk Mode and real-time cost trackers like TokenBar aren't "nice to have" — they're the difference between shipping and not shipping.
Start ugly. My first shipped app looked terrible. It still got paying users on day one. Your 15th unfinished beautiful app has zero users.
Track everything. Time, money, focus. The data doesn't lie, even when your motivation does.
What's your biggest blocker when it comes to actually shipping side projects? Drop it in the comments — I bet it's one of these two problems.
Top comments (0)