After reading an article about the new AI tools (such as Chat-GPT) in the New York Times this morning, I used DALL-E 2 to generate some images based on textual descriptions. Here is what was generated for the following descriptions I provided:
“Monument Valley in Utah filled with trees, flowers, and exotic animals including dinosaurs” :
What?!? Where are the critters? This is not much different than the actual place.
“Yosemite Combined with Yellowstone”:
That’s more like it!
“Bora Bora with the mountains of Switzerland in the background”:
Not bad! It’s “funny” that the Bora Bora huts are rocks in the images.
Inspired by Napoleon Dynamite’s Liger, “A cross between a lion and a Bengal tiger” :
OK.
Combining my near-obsession with the Ornithorhynchus and the Baltimore Orioles (Oh, for the days of the four 20-game winners (Palmer, Dobson, Cuellar, and McNally), the Robinson “brothers,” Brooks and Frank, and the Ripken brothers and dad), “A duckbilled platypus playing baseball at Camden Yards” :
This is the best of four images DALL-E 2 generated for that description; I’m not overly impressed with this one.
All in all, a mixed bag, but still impressive.
I will later put ChatGPT through its paces.