Optical illusion city portrait: pattern-fusion state machine prompt

2x2 grid, 16:9, do this for famous cities: STATES: S0: input($ city, $ pattern_family, $ illusion_behavior, $ aspect, $ mood) S1: infer_city_dna -> extract structural signals (verticality, density, waterfront, etc.) S2: load_all_pools -> if user didn't supply pattern/illusion, randomly pick 2 from each of the 10 categories (DNA drivers, pattern families, illusions, emergence, figure-ground, density, composition, color, surface, recognition) S3: fuse_selections -> for each category, create a combined descriptor (e.g., "Moiré Wave Bands fused with Vertical Bar Vibration") S4: construct_base_field -> using the fused pattern family and density logic, generate a precise geometric matrix (e.g., stripes, grids, dots) across the canvas S5: apply_city_dna_modulation -> warp the field according to city drivers: compress vertically for tower clusters, introduce river-curve negative space, place monument voids, etc. S6: inject_illusion_behaviors -> overlay moiré, figure-ground reversal, hidden depth by shifting pattern spacing or contrast locally S7: enforce_composition -> align with selected composition framework (centered mass, radial, etc.) S8: set_color_system -> apply stark two-color palette (maybe third accent), pure flat, no gradients S9: apply_surface_discipline -> crisp vector edges, mechanical repetition, no texture S10: recognition_layer -> ensure city emerges only through pattern density and shift; verify delayed recognition via recognition strategy S11: render_final -> output pure graphic op-art image in specified aspect ratio, no text, no illustration, negative constraints enforced