[simonw]

Honestly it never crossed my mind to waste some artist's time with this, but now that the joke "benchmark" has somehow reached orbital velocity maybe I should be thinking about it!

I've run the prompt through dozens of dedicated image generation models so I've seen many versions of this that are better attempts than a text model spitting out SVG - here's gpt-image-2 as a recent example: https://chatgpt.com/share/69ea21ab-8738-83e8-a4d7-67374d84e0...

[tripzilch]

I believe that if you pay them for their time, it's not really "wasted", at least not nearly as "wasted" as when the next person would pay them to design some vapid advertisement.

In addition to that 1) it's for science and 2) maybe you owe it to yourself to have a really nice framed picture of a pelican riding a bicycle on the wall :D

About the dedicated image generation results, I still would have made the bicycle smaller, but it starts to depend on how motivated the artist is to make both the bike and pelican accurate. Which is fine, but if you want to have a benchmark, it's important to have at least one "known good" example, I think.