July 16, 2026
JULY 2026 – AI Workflow
I’ve been working with Claude Fable 5 since it came back online July 1st. Anthropic launched it June 9th, the US government yanked it offline three days later under an export-control directive, and it was restored at the start of the month. In the two weeks since, I’ve completely recoded four websites with custom themes and come up with a new, improved site management strategy: a central asset repository, shared by all the sites, that houses core WordPress services.
This eliminates almost all of my plugins, which in turn vastly reduces site maintenance and shrinks the attack surface. This is a huge thing. The front-ends are all different, but all my sites are harmonized on the back-end and thoroughly documented, which makes updates a snap. I am extremely impressed with how well this works.
This system has the effect of leaning out the ridiculous bloated mess that WordPress has become. I’ve now got a simple, streamlined, and elegant solution that is much faster and also much more secure. It’s just a huge structural improvement.
Although Claude is doing the actual coding, it really helps to have at least a basic knowledge of how WordPress, PHP, and CSS work so you can ask the right questions and give the right instructions. But theoretically a total noob could do it.
In terms of the AI landscape, the competitive pace right now is relentless. OpenAI pushed out its GPT-5.6 lineup within weeks of Fable’s release, and Google shipped a major Gemini update this spring. The level of competence across the board is really impressive. It’s hard to believe that this has all come into being since November 2022.
I started out in December 2022 using ChatGPT, then last year started using Claude and Gemini as well. I went back and forth between them, but I’ve now settled into the following rhythm that seems to work well for me.
ANTHROPIC
Claude: all coding tasks done here.
Gemini: all everyday common questions, image creation, integration with other Google products like Gmail and Google Calendar, vetting other AI output.
Google AI Studio: this is Gemini’s big brother, and this platform can do things like transcribe a 90-minute video in less than 10 minutes.
NotebookLM: this is Google’s research and study tool, and it’s completely crazy. Upload a PDF or a video and it can summarize it, generate a narrated video presentation, or even produce a two-host podcast on the topic. This is a fantastic tool for students and learning in general.
OPENAI
ChatGPT: image creation, deep dives on complex topics, publishing projects.
I cross-check between the three of them constantly. They’ve massively improved and simplified my workflow.
See less