I’m still down here in the trenches, determined to finish Data Machine and release it to the world. The project has taken me much longer than I could have expected. Last month, I was in upstate NY at a friend’s bachelor party, and on my way back I thought I was ready to release the plugin the following week.

Since then, the system has been completely overhauled and rebuilt from the inside out. I’ve learned a lot about WordPress filters and actions and have put that knowledge to work in this system. As a result, it has gotten much more powerful. Data Machine has become a magical WordPress automation engine, and it’s almost completely functional.

I have been using Claude Code at the $100 per month level and I’ve found it to be the best AI coding tool I’ve tried thus far. It started with basic ChatGPT stuff, and since then evolved through Cursor, and Roo Code, and finally, now Claude Code.

I know I’m getting better at this because I’ve noticed more instances where I catch the AI making mistakes or not understanding my established frameworks. I’ve learned the importance of structure and architecture, and carefully planning file placement and responsibilities. I’ve found that aspect to be even more important than the code itself.

When gone unchecked, the AI will add a bunch of slop to your code. It’s just like when you use it for writing. It adds a bunch of fancy over-engineered crap that you don’t need. It also has a very hard time seeing itself as part of the system, so it often ignores established pathways and invents its own solution.

These are things that you might not notice when you first start coding. It takes a while to understand your role as an orchestrator. Your success lies in your ability to see the forest and the trees.

I’m definitely not at a totally pro level yet though. I’m good at it for sure, especially when it comes to WordPress, but there are certain things that give me trouble. For example, my AI Bot for bbPress plugin, I’ve had a few users report errors with blank bot responses, which I have not been able to replicate on my end.

On the forum, too, I had a bug report that I was not able to replicate, but eventually figured it out. User-reported bugs are something new for me, as I’ve always been testing things myself. It’s hard to notice everything when you’re just one guy in your bedroom.

However, this problem solving process has made me infinitely better at what I’m trying to do. I may not be a traditional developer, in that I never actually learned how to write code from scratch, but I don’t think that truly matters in today’s world. And if it does, it won’t matter for much longer, as the AI continues to advance.

I recently thought about how I wished I had majored in computer science instead of English. The more I think about it, the more I understand that my English degree may have been the most valuable pursuit I could have taken. Being skilled in written English communication is something that not many can say, and I’ve got a wrap on it after so much writing. Not just in college, but through music journalism.

I’m feeling pretty optimistic about my future right now. The first half of 2025 was tough for me. I went broke, and have been climbing out of that hole. But recent wins, not limited to Data Machine but also the fact that I’ve got a decent job as a tour guide right now.

Traffic on Sarai Chinwag is also trending upward, and I created a little plugin for the site called Sell My Images which provides upscaled, hi-res versions of all site content images for download. This is making me money right now.

Also, traffic on Extra Chill has seen a significant spike in the latest Google algorithm update. I don’t think that it will last forever, but I’m happy to see that it’s here. It gives me a bit of financial ease that had been seriously lacking in my life before.

Overall, I think I’m on the right path. I just have to knock everything out, one step at a time, and the chips will fall into place.