wide-angle photo of a transparent sculpture of a duck made out of glass in front of a painting (Dall·E 2) -- in many cases the light transport in both systems is amazing!
"Teddy bears swimming at the Olympics 400m Butterfly event" (Imagen) vs. "teddy bear swimming butterfly stroke while wearing goggles and swim cap in an Olympic pool" (Dall·E 2). Harder to get wide-angle and rapid action; using original prompt would generate actual butterflies.
Android mascot made from bamboo. Imagen looks slightly flat-textured but is clearly the logo, Dall·E 2 either trained on a more generic concept of "android" or is case-insensitive, and most results were toon-shaded.
"blue jay standing on a large basket of rainbow macarons" (Imagen best of ?? vs Dall·E 2 best of 30) most images had slightly-melty macarons, but I liked the action shots. Maybe Google feeds its AI more desserts?
An alien octopus floats through a portal reading a newspaper. (Imagen best of ?? vs Dall·E 2 best 3 of 20) I actually love the variety of outputs, random diffusion really can produce brain-tickling outcomes to trigger creativity!
"A chrome-plated duck with a golden beak arguing with an angry turtle in a forest" (Imagen) vs "chrome duck with golden beak arguing with angry turtle in a forest" (Dall·E 2), matching adjectives to correct nouns takes a bit more luck
![](data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="73" height="73"><rect fill-opacity="0"/></svg>)
