On: Imagen, a text-to-image diffusion model

[Tehdasi]

Still has the issue with screwing up mechanical objects. In their demo checkout the wheels on the skateboards, all over the place.

[sdenton4 → Tehdasi]

For comparison, most humans can't draw a bicycle:

https://www.wired.com/2016/04/can-draw-bikes-memory-definite...

[dclowd9901 → sdenton4]

I blame it on the surprisingly structural cleverness of a bicycle. Opposing triangles probably isn’t the first thing most people think of when they think of a bicycle (vs two wheels and some handlebars)

[gwern → dclowd9901]

They also can't draw pennies, the letter 'g' with the loop, and so on (https://www.gwern.net/docs/psychology/illusion-of-depth/inde...). Bicycles may be clever, but the shallowness of mental representation is real.

[gpt5 → Tehdasi]

I only see the problem for the paintings. If you choose a photo it's good. Could be a problem in the source data (i.e. paintings of mechanical objects are imperfect).