Before AI, most devs didn’t roll their own unless it was simple. After AI, the range of “doable” DIY apps is much bigger — blogs, API tracking, ad engines, dashboards.
But, what about now, with AI in the picture?
I wrote about why devs should consider building more themselves:
https://coderbdev.com/blog/ai-makes-diy-dev-realistic-again
What’s the last tool you built for yourself instead of buying?
Top comments (1)
This really lands. DIY dev is getting a second life with AI, and from what I’ve built (orchestration with Claude, ScrumBuddy, etc.), I’ve got some strong feelings around what’s working and what falls apart when people lean too hard into “just try it and see.”
What thrills me about DIY + AI is how fast you can go from idea to something that actually works; dashboards, sender checks, scraped APIs, whatever and for small solo projects, that velocity is intoxicating. But here’s my pushback: just because you can build something quickly doesn’t mean it’ll scale without structure. I’ve watched many tools die slow deaths because naming conventions drift, folder structure splinters, tests are missing, and nobody remembers why something was wired the way it was. In my experience, that sloppiness creeps in when you treat dev as “fun experiment” forever.
When I was developing ScrumBuddy, I learned that pairing AI speed with deliberate scaffolding (clear task breakdowns, incremental PRs, code reviews, conventions baked in) is what turns DIY ideas into lasting products. The folks who succeed longest with DIY feel less like they “slapped something together” and more like they built something they can improve without being afraid to revisit old areas.
So yes, cheers to the range of “doable” growing bigger, I want more people building their own apps. But strong opinion: if you don’t build with structure from day one, that rapid phase becomes technical debt faster than most expect. DIY dev gets realistic again only when we accept that every side project deserves its scaffolding.