Latter-day Saint Where’s Waldo and Stephen Biesty Cross-Sections

•

March 6, 2026

Google recently released Nano Banana 2, the most advanced image generator for small detail kind of stuff. (Still not the more advanced for artistic purposes, that’s still Midjourney–when people still say that all AI image generation is slop I point them to online fora of Midjourney artists).

When I was a kid I loved those Stephen Biesty cross-section books, the kind where you folded out the image of the Titanic, castle, or whatever and you were able to see the insides. I also liked Where’s Waldo, so I used the new Nano Banana 2 model to do Latter-day Saint versions. (Noting that Pat Bagley at the Salt Lake Tribune already produced a Nephite “Where’s Waldo”).

Of course there are inaccuracies, and if I wasn’t preparing for an international business trip I’d clean it up, develop it more and add more funny aside scenes like the real Where’s Waldo (these are all one-prompt, simple requests), but the basics are there.

Comments

2 responses to “Latter-day Saint Where’s Waldo and Stephen Biesty Cross-Sections”

Pontius Python

March 6, 2026

On the one hand, this is AI slop. Yes, there is art (“artifice”) and skill involved in crafting the AI prompts to get what you want, and there can be value in using AI tools as part of a larger creative process. But this is still slop.

On the other hand, I am highly amused by all the hallucinations, confabulations, and conflations in the Kirtland Temple cross-section. Clearly, this AI tool does not have enough domain-specific knowledge either in the Kirtland Temple or in cross-section illustrations to come up with a remotely convincing model of the building’s interior and grounds.
Coffinberry

March 6, 2026

As someone who has spent the past couple years deeply understanding downtown Salt Lake City in the period 1882 – 1915, that ZCMI marquee intrinsically annoys me!