Skip to content
refs
REF-0673Source onlineBlueprint rev 01

GPT-5's Quiet Reveal and Capability Sprint

A sparse product launch film turns tiny UI changes into a reveal, then compresses varied proofs into a fast montage.

Maker OpenAIFor OpenAIOriginal post ↗
Cut timeline · 12 shots01:29
00:0000:1500:3000:4501:0001:15

Field notes / why it works

A tiny model selector expands during 0:00–0:03, delaying the new product reveal until about 0:04 and making the white space useful. The finance task at 0:18–0:29 coincides with the top replay peak at 0:18.9–0:19.8, while nine quick cuts around 0:52–1:06 turn a calm interface film into a brief capability sprint. The film returns to sparse cards for a clear access message at about 1:24.

01

Plate 01 / The format in one breath

A near empty white frame slowly reveals a familiar product control, then the new product and its promise. Realistic interface tasks establish usefulness before a short, dense montage of things people can make; the ending returns to white space and a simple availability message. The promise appears visually by about 0:04 and in a headline by about 0:08.

Why it works

  • The first three seconds build a tiny UI menu in a large white field, inviting the viewer to read the options before the new name appears around 0:04.
  • A specific finance prompt and resulting chart occupy 0:18–0:29; the top replay peak is 0:18.9–0:19.8, where that task is submitted and the response starts.
  • The edit changes gear at 0:51.7: nine fast cuts through code, footage, and visual outputs before returning to quiet interface examples at 1:05.7.
02

Plate 02 / Format card

Platform · aspect · lengthYouTube · 16:9 · 1280×720 · 30 fps · 89.0 s
Pace12 measured shots, 11 cuts, 7.4 cuts/min, 16.2 measured picture changes/min; mean shot 7.42 s, median 1.29 s. The long UI sequences skew the mean.
ScriptNo reliable narration. Use about 8 short headline cards plus concise, task specific UI text; the long prompts should read as genuine work rather than slogans.
VoiceNone required. The lone nine word speech detection at 0:25 conflicts with the visual sequence and has low confidence; treat it as an uncertain false positive.
On screenAbout 84% white UI/typography and 16% fast code, live action, and created output footage, estimated from the shot durations.
CaptionsNo speech captions; black centered headline cards and small dark UI text on white.
SoundMeasured −17.6 LUFS integrated, −2.3 dBFS true peak. Music-like rhythmic bed around 94 BPM is inferred from audio measurement; cut effects are not clearly established.
StructureReveal → two broad promises → three task demonstrations → rapid capability montage → writing and customization examples → availability and logo. CTA at 1:23.6–1:26.0.

Pro / 13 plates + recipe

The first plates are open. The full dissection is Pro.

Get the timed shots, type and sound specs, and the runnable recipe.

Open the full blueprint
  1. Plate 03 / The hook, frame by frame (first 3 seconds)Pro
  2. Plate 04 / Structure (beat sheet)Pro
  3. Plate 05 / Shot grammarPro
  4. Plate 06 / Visual style guidePro
  5. Plate 07 / Voice and deliveryPro
  6. Plate 08 / Sound designPro
  7. Plate 09 / PackagingPro
  8. Plate 10 / Script templatePro
  9. Plate 11 / Production planPro
  10. Plate 12 / PromptsPro
  11. Plate 13 / Edit recipePro
  12. Plate 14 / QA checklistPro
  13. Plate 15 / Make it yoursPro
  14. Recipe / runnable promptPro