In the latter half of this year I felt my momentum on side project slow down to a crawl. The only one I was regualrly doing was this blog, but all my other projects fell into the same pattern: The idea was grand, but vague. No actionable steps that are small enough, so these projects start to feel a bit overwhelming, and I seek out another dopamine source instead (such as video games).

So I started a side project to help with this! To learn about streamlit, langchain, and sqlmodels, I built a little LLM-powered project planner (sidestep) that uses a description of your skills and the project you’re working on to ask you questions and generate a shecklist of small and actionable steps. This allows me to pick up projects whenever a have a spare moment, but not feel I’m overcommitting myself.

But first we should ask: why even do side projects? If there’s no point to it then why do it? I boiled down my answer to these main points:

• To learn new skills that I wouldn’t otherwise learn (and fail safely)
• To solve a problem (usually small ones)
• To express myself creatively

Not all side projects are important — it’s just another hobby. Moreover, not all side projects need to be finished. Sometimes you just need to engage with a project long enough to find where the enjoyment is and just leave it at that.

This small adventure was a fun way to explore how these technologies work together, and helped me understand the nature of why I do things like side projects.

Sometimes the best way to learn is to scratch your own itch.

Check out the codebase if you’re curious!