I support your call about SDXL. Been using DALL:E 2/3 and Midjourney since v3 for production of actual design collages for actual events, which get actually pri

I support your call about SDXL. Been using DALL:E 2/3 and Midjourney since v3 for production of actual design collages for actual events, which get actually printed. SDXL is nowhere near this all, unless you need a simple image of female-looking manga character with a dreamy eyes. I suspect the perceived superior quality of said providers is potentially due to their experience in wiring multi-model (multi-modal) systems and having more than one agent work on the image.