Spent yesterday improving a new WordPress plugin. 10x Raybo does exist, here’s what I did.
About the 10x
I feel like there’s a lot of controversy about this topic, but I really felt like I was able to put together a vibe coded project that I would have never done because of the volume of work. Even things like updating the docs, prepping the changelog or release docs was made ridiculously easy by the AI.
Working with Codex
My main new tool was Codex from OpenAI. I started by asking it to go through my code and find potential improvements.
It surprised me with 10 possible improvements. Noticing that I had updated the license in some places, but I missed it in others, some of the automation scripts needed to be updated since I had made modifications, there were classes that I needed to remove since I had done some other work, clean up my version numbers since I had done a small cleanup release after I ran the plugin verifier.
Those are the sort of small things that I just didn’t notice were missing. Codex was golden at finding those sort of things easily, then creating issues for taking care of them, fixing them, then pull requests.
Adding Guttenberg Blocks with AI
One of the suggestions was to generate Guttenberg blocks, which I hadn’t done before. That ended up being my first real upgrade from my original idea. I ended up doing this inside Copilot in VSCode just to be able to see what I was doing better. A total of 12 issues, solved and resolved mostly by AI with a lot of input.
I’m not sure how many X this is, but I feel like it’s at least 10x ing me. I don’t think it would 10x everyone, but if you count the amount of fixes it was able to locate and take care of that I would have never found, I think it was worth it.
One thing I might do in the future is take this project, and run a Claude Init command to generate a document for the next time I want to do a plugin. Seriously, that command is worth the Claude Code download by itself.
