Subtitle: How I body-slammed a stubborn Java app with the help of artificial intelligence (and a little stubbornness of my own)
I’ve been doing database programming since the mid-1980s—back when floppy disks were actually floppy and 64K of memory felt like an embarrassment of riches. My first project was an internship log database on a Commodore 64 using a variant of the Clarion database system.
I built it to log my medical internship activities—typed, printed, and turned in every month on time (well, mostly). And just like that, I got hooked on database development.
While my friends were busy building spreadsheet monsters, I knew the real power came from the one-to-many relationship. With that kind of setup, I could summon reports, analyze data, and look smarter than I actually was. It was glorious.
I came up through the DAO with Microsoft, evolved into the ADO, and eventually started doing relational databases for websites with PHP and SQL. That setup worked great: data in the cloud, accessible from any browser, and conveniently safe from my dog accidentally sitting on the power button.
But there was always this nagging fascination with Java. The siren song of “write once, run anywhere” was too strong to ignore. I learned enough Java to build small applications that could run on Windows, macOS, or Linux—pretty impressive for something written between coffee refills.
Then I tried to build a Java app that talked to a database.
And the result? Let’s just say it made professional wrestling look polite. No matter how many YouTube tutorials I watched, how many Stack Overflow posts I read, or how many cups of coffee I consumed, I couldn’t make the database behave. It just wouldn’t connect, no matter what driver, library, or ritual sacrifice I tried.
Enter ChatGPT
After retiring, I found myself even less willing to pay the American Osteopathic Association just to track my Continuing Medical Education credits. I wanted to build my own system—cross-platform, database-driven, and 100% under my control.
That’s when ChatGPT stepped into the ring.
Using VS Code (well, technically the open-source sibling Code-OSS), I built a simple app framework and asked ChatGPT to help with the database part—the portion that had thrown me for years.
And just like that, it worked.
The AI wrote clean Java code, created a SQLite database with two related tables, and showed me exactly how to wire it all together. It took seconds to do what I’d been struggling with for… let’s call it “a respectable amount of time.”
Sure, I wasn’t working on it full-time (retirement has its perks—like naps and barbecue), but still, the contrast was humbling. ChatGPT not only fixed the problem, it explained why it worked. For the first time, my Java database actually ran on all platforms—without the smoke, cursing, and forehead-shaped dents in my desk.
So… Will AI Replace Programmers?
I don’t think so.
AI can generate code faster than I can find my reading glasses, but it can’t innovate. It can’t decide what should be built—only how to build what’s already been done before. It’s a phenomenal assistant, not a visionary.
That said, it sure made me look good.
With ChatGPT’s help, I finally created the Java database application I’d been wrestling for years. It worked perfectly, cross-platform and clean. After years of failed takedowns, I finally pinned the problem to the mat.
And this time, I didn’t even need a Commodore 64 to do it.