Skip to content
refs
REF-3507Source onlineBlueprint rev 01

SIMA 2 Moves Between Virtual Worlds

A launch film that reveals an AI game operator, proves its reasoning through native interfaces, and expands to cross-world tasks.

Maker @GoogleDeepMindFor Google DeepMindOriginal post ↗
Cut timeline · 33 shots01:36
00:0000:1500:3000:4501:0001:1501:30

Field notes / why it works

The 9.72-second opening lets viewers inspect gameplay before the agent identity and giant title appear, turning the reveal into an answer. From 0:24 to 0:44, native task UI, a prompt, and a reasoning card make the capability claim inspectable; the faster seven-cut world montage at 1:20–1:26 then shows breadth before a slower real-world close.

01

Plate 01 / The format in one breath

Open on what looks like ordinary gameplay or product use. Reveal that {{PRODUCT}} is operating it, then climb from a single example to harder tasks, varied inputs, independent learning and transfer across settings. End by connecting the demonstrated behavior to a larger future use case.

Why it works

  • The first 9.7 seconds let the viewer inspect an action before the SIMA 2 title and explanation; narration starts at 2.94 seconds, so the reveal answers a visual question.
  • A claim is paired with visible proof: prompts and an inventory around 0:25–0:44; action across several games around 0:56–1:25.
  • The final 10.36-second shot moves from virtual footage toward household and robot imagery, expanding the stakes after the capability has been shown.
02

Plate 02 / Format card

Platform · aspect · lengthX · 16:9 · 1280×720 · 24.99 fps · 96.4 s
Pace33 measured shots; 32 cuts; 19.9–29.9 possible cuts/min; 20.5 definite picture changes/min; mean shot 2.92 s, median 1.89 s
Script160 spoken words; 107 wpm across the full film, 156 wpm during speech; seven long sentences, no questions
VoiceLow-register, assured explainer voice; animated pitch; measured pauses between claims
On screenPredominantly interactive-world footage; brief dark composite opening, title, prompt/reasoning cards and closing real-world inserts; no talking head
CaptionsSmall white sentence subtitles on some game shots; UI text and black reasoning cards carry the key proof
SoundContinuous low tonal bed; no clear cut-triggered effects; −19.0 LUFS integrated, −5.9 dBFS true peak
Structure6 beats: mystery → reveal → task breadth → reasoning → learning/transfer → larger implication; no explicit CTA

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