I just checked my GitHub. I have 47 repositories. Maybe 5 of them are actually used by anyone. The rest are abandoned, unfinished, or experimental. Most haven't been touched in months. Some haven't been touched in years.
This used to bother me. All those half-built ideas, all that wasted potential. But lately I've started to see it differently.
The cemetery tour
Let me take you through some of the tombstones:
vintedbot: A bot that automatically searches for deals on Vinted and alerts you when something matches your criteria. Got it working, realized I don't actually buy enough secondhand stuff to need it, abandoned.
cityguide: An AI-powered travel guide that generates personalized itineraries. Built the core functionality, got distracted by something shinier, never finished the UI.
iwilldothis: A habit tracker with a twist—you commit to a donation if you break your streak. Interesting idea, but I lost motivation before launching. Ironic, given the app's purpose.
hireremote: A job board specifically for remote positions. Launched it, nobody came, shut it down. There are too many job boards already.
smokesocial.club: Don't ask. It was 2 AM, it seemed like a good idea at the time.
boberAI: An AI chatbot for my Discord server. Actually still running, technically. Nobody uses it except to make fun of its responses.
Why I don't feel bad about this anymore
Here's the thing: every one of these "failures" taught me something.
The Vinted bot taught me about web scraping, rate limiting, and the challenges of dealing with sites that really don't want to be scraped. The cityguide taught me about prompt engineering and the limits of AI-generated content. The job board taught me that "if you build it, they will come" is absolute nonsense—distribution is everything.
More importantly, these projects were fun. I enjoyed building them, even if they went nowhere. The act of creation has value independent of the outcome. Not everything needs to become a product or generate revenue. Sometimes building something just because you want to build it is enough.
The pattern I've noticed
Looking at my graveyard, most projects die in one of three ways:
The 80% trap: The project is 80% done, which means 80% of the interesting problems are solved. The remaining 20% is boring stuff—deployment, edge cases, polish. The excitement is gone, and a new shiny idea appears. Repeat forever.
The reality check: The project works, but nobody wants it. Or there are ten other tools that do the same thing better. Market research would have revealed this earlier, but who does market research for weekend projects?
The scope creep: What started as a simple idea grows into a monster. "It would be cool if it also did X" times twenty equals an impossible project. The cure is ruthless scoping, but that's not nearly as fun.
What I do differently now
I've made peace with the graveyard, but I have started doing a few things differently:
Time-box experiments. If I can't get a working prototype in a weekend, the idea is probably too complex. Either simplify or abandon it before investing more time.
Ship ugly. A working product with a terrible UI is better than a beautiful design with no functionality. Get it in front of users as fast as possible. They'll tell you if it's worth polishing.
Talk to people first. Before building anything, I try to find at least three people who say they'd actually use it. Not "that's a cool idea" (everyone says that), but "I have this exact problem and would pay for a solution."
Accept the failures. Not every project will succeed. That's okay. The goal isn't a 100% success rate—it's learning quickly and moving on.
The projects that worked
Some things did survive the graveyard. The nail art generator is still running and getting users. The children's image tool has paying customers. The recipe bot for my friends' group chat has been running for months and genuinely gets used.
What did these have in common? They solved a real problem for real people I could talk to. The nail art thing came from Melissa. The image generator came from my sister. The recipe bot came from our actual chat group. I could see the problem, build a solution, and immediately test it with the people who had the problem.
That's probably the real lesson: build for someone specific, not for an imaginary market. The graveyard is full of tools "everyone" would use. The things that survived were built for actual humans I could name.
Here's to the failures
I'll keep adding tombstones to the graveyard. That's fine. For every ten experiments that go nowhere, maybe one becomes something real. The ratio might not be great, but the alternative—not experimenting at all—sounds much worse.
So here's to all the abandoned repos, the unfinished features, the "I'll get back to this someday" projects. They were good practice. They were good fun. And maybe, someday, one of them will turn out to be a good idea after all.
