Thanks, frankdrebin, I really appreciate the insights.
This might be (probably is) another indicator of my lack of knowledge of how AI works, but if they’re trying to fix the text/image divide, are they trying to fix the text//math/logic divide also? Seems more important.
