A Retrospective on Dropping My Side Project After 3 Months (It's Not Just About Developing)
I'm shutting down the AI-based reading tracker I built after work. I worked hard on it for three months, so let me write down my thoughts while they're still fresh.
What Went Well
- Got full-stack experience with SvelteKit + FastAPI + PostgreSQL
- Implemented features using the LLM API directly without LangChain, which deepened my understanding of the architecture
- Built a CI/CD pipeline with GitHub Actions
What Was Disappointing
- Started with "just slap AI on it" without defining the problem first
- Kept adding features without validating user needs
- Working alone meant no code reviews and no schedule management
- In the end, only 5 users (including me), and only 1 active user
What I Learned
1. Defining the problem comes before technology
2. Side projects also need clear schedules and goals to have a definite ending
3. 10 hours a week for 3 months is enough for just one small feature
For the next project, I'm thinking of narrowing down the topic and splitting it into FE/BE. Writing a retrospective without actually doing anything is the worst, so I'm planning to start next week.
4 answers
Agreed, problem definition is the most important.
Well... we did gain some tech stack experience, didn't we? Doesn't seem like a total failure.
I also tried it for 6 months and then gave up, lol. If you pick a broad topic, you really get nowhere. The second time, I barely finished by focusing on just one tiny feature. Setting a schedule and goals is the most important thing.
Nice retrospective. I can relate—I've also built something without validation and ended up shutting it down. Three months is a long time, so next time, even if it's just interviewing people around you, that would make a huge difference. You got this!