Thanks for the comment, Jonathan! I’m glad you liked the article. Getting past the “visual” can definitely be a challenge. I’ve found it’s helpful to consider it from a number of angles: starting with semantics is better for building accessibility from the start, and is a requirement if you’ll be piping that content to a VUI. It also means that you can use effective visual design to _extend_ a seaworthy semantic foundation (instead of plugging holes in one that never should have left the slip).
All that said, there are definitely times where the pictures are also salient content—but even then, assuring that they have effective alt tags and support different users and use cases can help one nudge them into the structured content fold (sorry, had to lose the maritime metaphor; I was going overboard).
Please do share how your voyage into semantics goes. I think there’s a ton of good work to be done there and I love seeing how others tackle it.