Thanks for some of that insight, Thomas. I agree, the implementation of DALL-E3 in ChatGPT is a little disappointing. Hopefully they'll cross-pollinate the GPT-4 modes into one. It's certainly nothing that makes me want to stop using MidJourney.

Regarding your BART data example - although DALL-E falls short, of course, Data analysis mode smashes things like this. I literally cut and paste the text from your screen shot into Data Analysis mode and here's your graph (by the way, amazing to see how the ridership has never recovered post covid!) https://imgur.com/a/opnDzTC