Okay, I was wrong about the options fir creating video are the same. It now has a text entry bar at the bottom of the screen for entering prompts (instead of only showing up when you press the “custom” button. It remembers and lists the prompts. And it stores up to five versions of the video. And it does that retroactively. I tend to delete “favorites” that I’m done working with to avoid clutter, but it has stored the prompts and five versions of every saved favorite I have, retroactively, to before the interface upgrade.
(Also, on the “favorites” page, instead of showing still thumbnails it plays each video (without sound) on the page one by one.)
And it looks like image generators still struggle with text. The fake Latin is nonsense, and the signatures don’t correspond to the printed names, and one of the printed names is “President: Jatr of Südents”. And the outline still looks hand-drawn.
Also, the original clearly (though illegibly) had two lines of printing below each signature, presumably a name and a title for each. The AI one only has a single line (one title and three names).
Last night I used the Grok web version for the first time (so that I could tell someone else how to use it) so I had my first access to the “spicy” button. It appears to work with only recognizably human subjects. And only shows up for a subset of those. And when it is an option a large percentage of the time the output is moderated before I see it. And the times it does make something, the output doesn’t really look any different than normal generations. So automatic “spicy” creation in my experience so far isn’t very spicy.
