Minimalistisches T-Shirt-Grafik-Prompt

2x2 grid, 16:9, do this for 4 famous philosophers: INPUT: $ SUBJECT // If an image is supplied, treat the visible subject as $ SUBJECT. // ------------------------------------------------------ // SEMANTIC REDUCTION // ------------------------------------------------------ S = analyze($ SUBJECT) IDENTITY = { silhouette = S.minimum_recognizable_outline, anchors = top_3_identity_features(S), expression = simplest_readable_expression(S), appendages = simplify(S.appendages), action_logic = infer_characteristic_or_absurd_action(S) } complexity_target: semantic_information ≈ HIGH drawing_information ≈ VERY_LOW // ------------------------------------------------------ // SHAPE COMPRESSION // ------------------------------------------------------ FORM($ SUBJECT) = blob(silhouette) + primitive(anchor_1) + primitive(anchor_2) + primitive(anchor_3) + tiny_face_marks; Rules: - Convert the main body into ONE dominant flat irregular silhouette. - Remove anatomy that is not necessary for recognition. - Reduce eyes to dots, circles, slits, or deliberately mismatched primitive shapes. - Convert limbs / extensions into awkward sticks, hooks, wedges, loops, or noodles. - Preserve small irregularities rather than perfect geometry. - Slightly wrong proportions are desirable. - Make the subject look like someone understood it perfectly and then drew it in 6 seconds. // ------------------------------------------------------ // DEADPAN EVENT GENERATOR // ------------------------------------------------------ PROP = infer_unexpected_but_semantically_interesting_prop($ SUBJECT) interaction = choose( holding(PROP), staring_at(PROP), balancing(PROP), wearing(PROP), accidentally_becoming(PROP), ignoring(PROP) ) IMPORTANT: PROP should create a visual joke through contrast, not random clutter. // ------------------------------------------------------ // PALETTE // ------------------------------------------------------ BASE = black_or_near_black GROUND = warm_white / dirty_white / paper_white ACCENT_COUNT = 1 ACCENT = choose_semantically( red, orange, acid_yellow, cobalt, toxic_green ) Use accent ONLY for: one object OR one body feature OR one visual consequence; // ------------------------------------------------------ // TYPOGRAPHY // ------------------------------------------------------ OPTIONAL_TEXT = generate_1_to_3_word_deadpan_caption($ SUBJECT, interaction) Typography: crude bold uppercase slightly uneven baseline hand-cut / cheap screenprint feeling no polished branding no explanatory paragraph // ------------------------------------------------------ // RENDER // ------------------------------------------------------ STYLE: naive graphic icon black-paper-cut silhouette crude editorial cartoon zine humor imperfect ink edges tiny registration errors huge negative space BACKGROUND: nearly empty COMPOSITION: [ strange simple subject ] [tiny prop] CAPTION? Mood: confusing for 0.5 seconds then immediately funny. AVOID: polished vector mascot detailed illustration realistic anatomy gradients glossy 3D multiple accent colors decorative background kawaii overload OUTPUT: A brutally simplified, oddly charming graphic interpretation of $ SUBJECT - recognizable through silhouette and 2–3 semantic cues, with one unexpected action or object creating deadpan visual humor.