Solve Your Own Problems First (and Only Your Own)

More than a year into vibe coding, I still haven’t quite figured out how to relate to this new kind of productivity. Whenever the hype flared up, I’d jump in with great excitement and hack away for a bit. But pretty quickly, the lack of practical utility and any real sense of accomplishment would deflate me. It’s been an endless cycle of fizzling out and starting over.

Recently, while building a few small tools to fix UX annoyances in my daily workflow, I realized this may well be the use case that makes the most sense for regular folks vibe coding: solve your own problems, and only your own. A few key traits make this approach work:

• Low complexity: AI can knock it out in no time.
• Small scope: but because you keep using and tweaking that one specific thing, it gradually gets smoother and better to use over time.
• Productized: packaging it into an actual tool makes day-to-day use that much more convenient.

In the end, you only ship the ones you’re genuinely satisfied with. That sidesteps the hollow, aimless feeling of vibe coding and building products without a clear goal in sight.

As a longtime Raycast user, I’ve often wondered: where on earth do they find so many tiny needs to solve? I suspect they did the exact same thing: solve their own problems first, not the users’.

Once you adopt this mindset, you don’t decide whether to build something based on whether it’s already been done before, or whether anyone else will use it. If existing solutions aren’t elegant enough, then I guess I’ll have to build it myself; and because it’s solving my own problem, I couldn’t care less whether anyone else ever uses it.

Perhaps this is the vibe coding playbook that actually works best for regular people.

Another truth that hits home the more I code: someone without real engineering practice will find it very hard to vibe-code a solution that combines taste, solid engineering, and widespread appeal.

The barrier to entry has dropped, but the effort required is still substantial. It’s no place for a half-hearted coder.