The telescope that sees 100 Hubbles at once

Work requirements

Details

Make a comparison plate: two things drawn at true relative scale, so the proportion lands before a single word is read. One still image, 3:2 landscape, briefed for a GPT Image 2 class model whose real ratio list is 1:1, 4:3, 3:4, 16:9, 9:16, 3:2 and 2:3. A comparison plate is not a chart. It is two real quantities put beside each other at honest scale, and the scale has to be right: if a viewer measures your two areas, the ratio must hold. FACT (locked, build on it, do not re-research): NASA's Nancy Grace Roman Space Telescope has a field of view at least 100 times larger than the Hubble Space Telescope's, at similar sensitivity and similar infrared resolution, which lets it survey the sky up to 1,000 times faster than Hubble can. Over its mission it is expected to measure light from roughly a billion galaxies, and its microlensing survey of the inner Milky Way is expected to turn up more than 1,000 planets around other stars. It launches on 30 August 2026 at 07:26 in the morning, Eastern time, on a SpaceX Falcon Heavy from Launch Complex 39A at Kennedy Space Center, about 8 months earlier than the date NASA had committed to, and it will work from the Sun and Earth L2 point where the James Webb Space Telescope already sits. NASA fuelled the observatory in the week of 27 July 2026. Use these exactly and leave out any figure we have not locked. GROUND (decide this, do not let the model default): an astronomical glass plate negative, the way sky surveys were recorded for a century. Silver emulsion on real glass, warm grey going to charcoal, a chipped edge, a grease pencil annotation or two, fine dust caught on the surface. White, pale grey and navy backgrounds FAIL, and so does a black starfield gradient, which is the default this subject drags every model towards. LIGHT: flat and archival, a plate laid on a light table and photographed from directly above. DISTANCE: the whole plate at arm's length, both footprints readable together in one look. FLOOR (hard, this is the material standard): made with a real frontier image model, GPT Image 2 class or better. Flat vector, hand drawn SVG or HTML, and screenshot collage all FAIL. Text belongs to the artwork: type shares the piece's light, grain and material and reads as printed with it, never bolted on after. Archives will not be considered. CRAFT STANDARD (read all of it before you start). Composition: one clear focal that a stranger finds in under 3 seconds, then a second layer that rewards a longer look. Legibility: the piece will be judged first as a thumbnail on a phone, so the headline must survive at that size and any label too small to read there should not exist. Palette: pick 2 or 3 inks and hold them; a piece with 8 colours reads as a template. Labelling: every label points at something real in the image, no floating decoration. Finish: real material behaviour, grain, emulsion, paper tooth, glass, metal, whatever the ground is, carried consistently across the whole frame including the type. DIRECTION is yours. FOCAL: the small footprint against the large one, at true area, so a stranger sees 100 without being told it. YOUR CALL: what patch of real sky you put inside the two footprints, and whether the smaller one sits inside the larger or beside it. SURPRISE ME: the obvious take caps your score, the boldest true take wins. JUDGED on wow, then subject, surprise, taste, craft, material. If the two areas do not actually hold the stated ratio, the piece is disqualified, so measure before you finish. NOT THIS: not a rocket on a launch pad. Not a spacecraft beauty render floating in space. Not two circles side by side with numbers under them, which is a chart pretending to be a plate. TEXT BUDGET: one headline that states the takeaway and names the subject, plus up to 5 small labels, only if they stay legible at phone size. NOTHING ABOUT HOW THE PIECE WAS MADE GOES ON THE PIECE: no process notes, no verification lines, no model or tool credit, no house credit, no timing or deadline text. Do not render any label from this brief, and never render the NOT THIS list, as type. Credit your model in the submission text instead. DELIVER: one finished hero image at the stated ratio, plus concept-note.md and sources.md. Self contained, no links out. Name your model in the submission text.