I've been building a student registration system for homeschool co-ops to use for tracking students and courses and organizing rooms, basic stuff. The impetus for this was personal because my kids are in a couple of co-ops and they all handle this need in a range of ways, spreadsheets, home rolled websites, group chats. Everyone basically gets by with what they can, but shared passwords are the norm so I started building something to fill the gap.
At first the goal was to build an open source solution that could be checked out and deployed by anyone, I would be user 0 and deploy it for a couple of the co-ops that we're in, maybe it would get a few stars on github? Maybe people who liked it would reach out for consulting work, I don't know, I had a lot of ideas kicking around when I pushed the initial commit back in July 2025.
I spent 6 months on it, on and off, nearly all vibe coded using co-pilot + whatever my favorite agent of the moment happens to be (tied between Codex 5.2 and Claude Opus 4.3) and I finally got it to a point where it was feature complete and ready to deploy. It's an MVP but it covers the basics of what I wanted to do. On a whim I decided to take things to the next level and turn it into a SaaS offering.
I guess I shouldn't really say it was on a whim, part of the thought process that influenced my decision came from seeing a real product take shape over time, and knowing that it fills a real need. The other part comes from pushing the envelope with LLMs and coding agents and starting to see the writing on the wall. Coding agents are going to enable a lot of people to become entrepreneurs in their spare time, and with the way the economy is looking a lot of people might have to become entrepreneurs, so I may as well get ahead of the curve.
I know I'm not the only person to see this sea change, but I do think I'm pretty close to the leading edge as far as adoption and use goes. And while I'm not a doomer here to tell you that the programming profession is dead, there are a lot of my peers out there clinging to "pure programming" and memorizing manuals and documentation, and I am here to tell you that those folks are dinosaurs, and coding agents are a meteor.
The first time I tried to build a blog I was 19 years old, penny-arcade was wildly popular and it seemed like anyone with a modicum of talent could make money by building a blog and a comic strip. I taught myself php, learned to draw, learned photoshop, the ins and outs of self hosting, and eventually petered out, while never really getting good enough at all of the skills I would need to make anything worthwhile.
Now, with 25 years of experience under my belt I can easily accomplish the kinds of builds I struggled with in my early days, I could probably get something really solid to production in a day or two. But for perspective, an LLM can do the same thing in about 20 minutes. If that math holds then equipped with an LLM I can do a solid weeks worth of work in about an hour. This isn't hyperbole, this is the first time productivity gains like this have been put directly into the hands of the people.
Before agentic programming turning a weekend hobby project into a SaaS offering would have been a major decision, not something done on a whim, only if you have a really good idea. Because the thinking back then was that if your side project got popular you would have to make a hard choice about your future. Do you really want to commit a huge chunk of your life to this project? Is it even worth it? Now it has been trivialized to the point where one could feasibly run several SaaS side businesses without quitting their full time job, what a time to be alive.