I learned the hard way that scarce energy should go into core work, not into tools or infrastructure you don’t really need.
That matters because the real cost isn’t money—it’s time. When you’re working on a side project in the margins of your day, every hour spent wrestling with setup or learning a new workflow is an hour not spent building the product itself.
Where most people stumble is chasing what looks “cool” instead of what’s useful:
- Trying to be the hacker who sets up their own VPS and knows every config file.
- Switching to a shiny new project management tool because everyone raves about its UX.
- Burning hours tinkering with setups that don’t actually bring you closer to shipping.
Cool tools won’t ship your project.
Here’s the process I use now to keep myself focused:
- Check if the task is core to moving your project forward
- Prefer familiar tools that get you to the same result faster
- Delegate or outsource when possible
- Use AI to cut through setup and research
- Only detour if you’re genuinely curious and willing to invest the time
Check if the task is core to moving your project forward
I was worrying about cloud bills before I even had a single client for my side project. The better move was to ship faster on free tiers with Vercel, Neon, and Sentry—tools I already knew. Instead of paying for a VPS and sinking hours into setup, I could validate my idea with no upfront cost.
Prefer familiar tools that get you to the same result faster
I already use GitHub every day at work. With only 1–2 hours a day to spend on side projects, I don’t have the bandwidth to learn a new tool that does the same thing. Linear may be excellent, but GitHub Issues and Projects were plenty for me as a solo builder.
Delegate or outsource when possible
In my case, there’s no one to delegate to—it’s just me. But if you’re on a small team, that might be different. Handing off tasks to someone with bandwidth frees you to focus on the part only you can drive.
Use AI to cut through setup and research
ChatGPT saved me hours by comparing providers in a structured way. The key is giving a detailed prompt:
- What is the current situation?
- What is the problem?
- What is the outcome you want?
- What are you familiar with?
- What are you trying to avoid?
I asked GPT to produce comparison tables of free vs. paid tiers, list pros/cons, and recommend a winner and runner-up. Then I read the report. That was much faster than combing through docs myself.
Only detour if you’re genuinely curious and willing to invest the time
That’s the only exception. If you want to learn sysadmin or test-drive a new project management tool out of curiosity, go for it. Just be honest with yourself that it’s a detour. Exploration is fine, but don’t mistake it for progress.
