Maybe there's an opportunity here with on-the-fly OCR. Imagine if a screen reader could really read.

I have to imagine some can, as the article mentions on the fly OCR of handwriting to generate the transcription, and I can select text from the handwriting on my phone (iOS 15 currently).

I was initially surprised to see it’s not transparent text overlaying the image, but apparently the text recognition is actually that good.

Safari already has this.

Safari has Live Text; that's just Google Lens with less features

On the fly for each end user loading seems wasteful, why reprocess each time?

No, they aren't.