The dragonfly interception a weapons lab copied

Work requirements

Details

ASK: one short film of a hunter solving an interception. FACT (locked): dragonflies capture prey with up to 97 percent success on easy prey in a large screened outdoor enclosure, as low as 20 percent against big evasive prey. They aim where the prey WILL be; a weapons lab copied it for interceptor guidance. Source: Combes et al, 2013. THE ONE BEAT: a dragonfly locks onto a point in empty sky and the prey arrives there. ONE subject. Build tension, then push the camera IN on the decisive moment - make people WATCH, exciting not just pretty. The machine is in HOW it is shot: footage an engineer studies frame by frame. REAL: ONE single-shot text-to-video generation, 6 to 12 seconds, on a real video model (Veo, Kling, Seedance or Runway class). One beat, native aspect. We are NOT asking for a longer runtime. Deliver the raw clip plus model and prompt. Interpolated, morphed, panned or keyframe-assembled work FAILS. ACCURACY: the locked figures are the only versions; want one we have not locked, leave it out. Extra true material welcome, sourced. SOUND (judged as craft): a suspenseful score as a separate layer is welcome; native audio if you have it, else describe it. Wingbeat, air, the quiet before the strike. HEADLINE: how a person says it out loud, not how a label reads - a stranger names the insect AND the machine idea in 3 seconds. If it sounds like a caption or product feature, redo it. VOICE: one named register, in notes. CARE: the qualifier travels with the number. Never world's deadliest. DELIVER (detail in INSECTS-MAKER-GUIDE.md): the raw clip, a poster frame, 3 sampled stills, concept-note.md and sources.md. Archives will not be considered. Nothing about how it was made on the frame. JUDGED: does the beat land, is the clip gallery-wall grade, sound as craft. SURPRISE ME: one thing we did not ask for. NOT THIS: a 30 second edit; clips stitched.