Skip to content
refs
REF-1117Source onlineBlueprint rev 01

Gemini 2.0: Cinematic Claims to Agent Demos

A chaptered AI launch film that turns an abstract promise into live phone, browser-agent, and game demonstrations.

Maker @GoogleFor Google GeminiOriginal post ↗
Cut timeline · 28 shots02:52
00:0000:3001:0001:3002:0002:30

Field notes / why it works

The title is readable by 0:00.75, then 18 seconds of restrained blue motion gives the claim room before narration. Real-world phone proof begins at 0:44, and the browser task shifts from a legible prompt into 12 cuts in the 1:40–1:50 bin; that handoff contains the strongest replay peak at 1:40.3–1:42.1.

01

Plate 01 / The format in one breath

A widescreen product launch film opens with a quiet, abstract brand promise, then proves it through three increasingly concrete environments: a person using the product in the world, an agent taking steps in a browser, and a specialized use case. Chapter cards connect the examples, while narration explains the capability and authentic task audio carries the proof.

Why it works

  • The headline arrives within 0.75 s, yet narration waits until 0:18; moving blue filaments and a spare title let the promise register before the explanation.
  • The first real-world demonstration starts at 0:44.2. The viewer sees a phone aimed at a distinctive sculpture, then hears questions and answers tied to that scene.
  • The most-replayed region, 1:40.3–1:42.1, sits at the handoff from a task prompt to rapid website actions; visible work validates the narrated claim.
02

Plate 02 / Format card

Platform · aspect · lengthYouTube · 16:9 · 2:52.0 · analysed copy 640×360 at 23.98 fps
Pace28 detected shots; 27 cuts, or 9.4/min. Allowing uncertain changes: 9.4–12.9 cuts/min. 34 changes within shots; 21.3 total picture changes/min. Mean shot 6.14 s; median 2.22 s.
Script328 transcribed words; 132 wpm across the full film, about 197 wpm during speech; 21 sentences averaging 15.6 words.
VoiceMeasured medium-pitch, animated narrator plus conversational question-and-answer voices during demonstrations.
On screenApproximate time mix from sampled frames: 35% abstract motion/title cards, 22% location footage and phone close-ups, 19% browser/UI capture, 24% game/other application proof. The last two shares are approximate because long shots contain internal scene changes.
CaptionsNo persistent social captions. Small white dialogue subtitles appear on live-action and game proof; large white sentence-case chapter titles appear on navy motion cards.
SoundRestrained tonal cinematic bed about 15 dB below voice; approximate 80 BPM opening pulse (low-confidence estimate). −14.2 LUFS integrated, −1.9 dBFS true peak.
StructurePromise → platform capability → live assistant demo → browser agent demo → applications → end card and URL. Proof begins 0:44; URL at about 2:45.

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