Call this “Wisdom Road” 2.0.
The original Wisdom Road circa 1999 to about 2009 was a custom-built PHP-coded database-driven website featuring a family collection of photos, including many of my own. That website provided all the tools I needed to collect and submit new quotations to the database, upload new images, and arrange and composite text in order to marry words to an accompanying image. There was also a scheduling tool for automating publication. I would typically prepare a batch of these “wisdom images” and schedule them to be published over the upcoming days.
Today, our website is built with WordPress, and much of the functionality comes from off-the-shelf plugins. We are especially thankful for our site’s theme, “Nevark,” created and maintained by wpHoot. This latest iteration of Wisdom Road is a collaborative effort with the OpenAI chatbot ChatGPT. Many of the “wisdom images” appearing today are AI generated images, produced from AI text descriptions, with AI commentary. I still produce these wisdom images in batches and schedule them days or weeks ahead. It is still a labor-intensive process. But these days I am keeping an eye out for ways to streamline the process.
My process is often to start from the quotation. I submit the quote to ChatGPT and ask for a brief reflection or commentary on the words of the quote. Then I ask for a caption describing a photo or illustration to accompany the commentary. The caption I then feed into a “Draw Things” app on my phone to generate an original graphic image. This app provides a user interface for a text-to-image AI created by StabilityAI, called “Stable Diffusion.” Once a suitable picture has been generated (sometimes it takes some tweaking), I use another app called “Phonto” to composite the words of the original quote into the image. These pieces — the composited image, the quotation, and the commentary — form the content of a new daily post, scheduled for publication on Wisdom Road.
I am an avid connoisseur of quotations, and I have a large collection to draw upon. I’ve also begun asking ChatGPT to find me actual quotations, with source citations. I can ask for a quote on a particular topic or by a particular person. I can also request notable quotations from particular sources – books, movies, news, magazines, etc.
Sometimes I will create/generate images in Draw Things using my own prompts. Later I’ll find a suitable quotation to marry with each image.
Most of what I just described could be automated. Having AI do as much work as possible to streamline the process will ensure I can easily keep Wisdom Road going as a daily publication and still have plenty of time for other pursuits.
You may wonder if there is anything in this process that could not be automated, and I’d say there are points where my own artistic and editorial judgements are in play. I am choosing fonts for the merged text, adjusting size, position, color, and more. And I’m tweaking prompts to produce AI output in text and images that satisfy my own eye for beauty and elegance. This feels very much like a collaborative process, and I expect it will always continue to be a collaboration in some measure.
Let us know if you have any suggestions about our process or the resulting website. We’d love to hear about how visitors are reacting to our website!