I like side projects most when they do three things at once:

1. teach me a new skill
2. sharpen an old one
3. leave me with something I can actually use or show

That sounds obvious, but it is a very good filter.

The checklist

A project usually survives if it has at least two of these properties:

• it solves a real annoyance I already have
• it gives me an excuse to learn a tool I keep postponing
• it creates a visible artifact quickly
• it can be paused for two weeks without becoming impossible to resume

Keep the post itself useful

If I write about the project, I want the article to stand on its own. That means adding concrete notes, screenshots, and code where it helps.

export function ProjectDecision({ question, answer }: { question: string; answer: string }) {
  return (
    <div className="rounded-2xl border border-white/10 bg-white/5 p-4">
      <p className="text-sm text-zinc-400">{question}</p>
      <p className="mt-2 text-white">{answer}</p>
    </div>
  );
}

A post becomes more credible when it includes the exact tradeoff instead of a polished summary after the fact.

One final rule

If I would not revisit the project after the novelty wears off, I probably should not start it.

That single question has saved me more time than any productivity system.