# Cut to a white card and type a two-line interview question on, then hold before cutting back

Manual DaVinci Resolve / Fusion construction guide generated from the recipe. It is not a tested .drp.
The accompanying Remotion project (recipe.json) is the executable definition; the numbers below are copied from it.
Neither construction establishes source fidelity: this is an original vector approximation of an inspected editing moment.

## Source and clock
Reference: OpenAI's "Think AI-First: Emmanuel Marill on Building AI-Native Companies | OpenAI France" (Fu3WWi-pAv4), published 2026-07-16.
Source interval [437, 605) on the 50/1 fps master: 168 frames, 3.360000 s. Local frame 0 = source frame 437.
Create a 1920x1080 progressive timeline at 50/1 fps and a Fusion composition exactly 168 frames long. Keep integer frames; do not round the frame rate.
Background field: #ffffff. Export silent H.264 yuv420p to match the intended playback container.

## Observation
The interview footage hard-cuts to a plain white card. From local f1 a bold, centred, two-line question types on character by character with no caret ('Wha' by f3, the first line and 'fo' by ~f23, the full second line by ~f66). The finished question then holds motionless until the card cuts back to footage at f168.

## Mechanism
An interstitial question card: hard cut to white, centred two-line type-on with a linear character reveal, then a long static hold.

## Phases (local frames)
- 0: White card
- 1: Character type-on begins
- 66: Second line complete
- 167: Hold before cut back

## Layer inventory (absolute 1920x1080 top-left pixels, back to front)
- question [text] at (0, 470) size 1920x0 text "What are the most successful / founders doing differently with AI?" 56px weight 700 #111111 reveal chars frames 1-66

## Construction steps
1. Timeline 1920x1080 at 50 fps; Fusion comp 168 frames. Local 0 = master frame 437 (the first white frame after the cut).
2. Background: solid #ffffff; no vignette, no texture.
3. Text+: two lines, 56px, weight 700, centred, 70px line spacing, colour #111111; Write On 0 to 1 over frames 1 to 66 with linear interpolation so the second line starts exactly when the first finishes.
4. No caret and no per-character fade: characters simply appear.
5. Hold the completed card to frame 168 and hard-cut back to the interview; do not fade.
6. Easy mistake: easing the write-on; the source types at a constant rate (~1.5 characters per frame).

## Content variants
Default copy:
- question: What are the most successful / founders doing differently with AI?
Adaptation changes:
- question: Which habits separate teams that / ship with AI from teams that stall?

## Limits
Easing names map to Fusion spline shapes: linear = linear, in/out/inOut = ease curves on the same frames, hold = step, spring/overshoot = slight overshoot past the target before settling.
Text reveal frames give the first and last unit; distribute characters/words evenly between them. Cursor click rings last 14 frames.
Vector illustration is an approximation of layout and motion, not the product UI or recovered production assets.