What Kinds Of Image Requests Confuse DALL-E?

In his latest newsletter, Charlie Warzel dives into DALL-E and profiles Andy Baio’s experiments with the tool. Baio has found it rewarding to stretch DALL-E to its limits (call it boundary testing, if you like).

The engine doesn’t generate text well at all—it can generally only create short words. It’s bad at creating faces, and most trademarked material and images of famous people are blocked so that people don’t abuse the software. But Baio has also found more interesting limitations. “Anything that’s an opposite, like a horse riding a man or a hand with six fingers, is a real struggle for it,” he said. “We want to think that DALL-E has this wild imagination and that it is capable of generating wildly unusual images, but because it’s been trained on tens of thousands of images of people with 10 fingers, it really struggles.”

Rumor has it that DALL-E will be opening to a broader beta cohort in the upcoming weeks. I’d like to play around with it, but I expect my images will be fairly pedestrian. I’m not into surrealism like the image generator’s namesake.

Tech’s New Frontier Raises A “Buffet of Unwanted Questions”

Canned Dragons by Robert Rackley
Made with in North Carolina
© Canned Dragons