Image ChatGPT Image 2.5
Sliced Asian street food reveals layered cross-section on clean grid
A 2x2 grid showing four underrated Asian street foods cut open, each layer labeled with real thickness and honest blade behavior.
Prompt
2x2 grid, 16:9, do this for 4 underrated street foods in Asia: INPUT = [ PRODUCT ]
Run CROSS_SECTION_STRATA_PAGE:
INFERENCE_CASCADE := derive the real LAYER STACK (3–7) → each layer's actual THICKNESS
→ its MATERIAL and job → the ONE layer that is the selling point → whether the
category cuts clean (blade) or tears (foam, bread)
HERO_DNA := one unit sliced on a SINGLE FLAT PLANE, cut face square to camera,
intact exterior still visible at the edge so you read object then interior
STRATA_LAW := layer thickness in the render is PROPORTIONAL to the real spec — a 2mm
layer cannot look the same as a 40mm layer. Order is fixed and matches
the labels. Every cut face shows interior structure DIFFERENT from that
material's exterior (open cells, crumb, weave, grain).
CUT_HONESTY := the cut behaves like the material — foam compresses at the blade, bread
tears, metal shows a burr, chocolate fractures. A laser-clean cut through
a soft material is a failure.
LABEL_GEOMETRY := leader lines land on the exact layer named, no crossings, all labels on
one side, thickness in real units beside each
DEPTH_CUE := the cut plane is lit slightly brighter than the receding interior
COLOR_DISCIPLINE := neutral ground + ONE signal color reserved for the hero layer and CTA
IMPERFECTION := one layer slightly compressed or displaced where the blade passed
FRAME := 16:10; commerce chrome present — nav, price, ADD TO CART
MOOD := "so that's what's inside"
Render: one screen where the interior sells harder than the exterior ever could.
Forbidden — objects: invented layers, decorative gradient bands standing in for material
Forbidden — look: uniform layer thicknesses, labels pointing between layers, identical
texture on cut face and outer surface, multiple contradictory cut planes, a soft material
cut like glass
Published: October 16, 2026 by Gadgetify