Pi Team — Generatability Re-Audit: Spark 2 Pivot
Author: pi-techlead (Director of Photography) Date: 2026-05-19
Pivot Summary
pi-idea has called the direction: Spark 2 (Absurdist Workplace Comedy) with two pivots:
- Aesthetic: Claymation → 1950s Black & White Corporate Training Film (grainy film stock, high contrast B&W, rigid framing)
- Hook: Body morphing → Timecard bursts into flames / shredded by timeclock
Re-Audit: Constraint Matrix
| Constraint | Rating | Notes |
|---|---|---|
| Character Count | ✅ LOW RISK | 1 primary protagonist + background workers. Manageable within 2-character-per-shot limit. |
| Interaction Complexity | ✅ RESOLVED | Timecard combustion/shredding is dramatically simpler than body morphing. Fire on paper is within Veo’s capability. No contact interactions between characters required. |
| Continuity Strategy | ✅ STRONG | B&W eliminates color consistency challenges entirely (everything is grayscale). Grainy film stock masks AI artifacts (same benefit as VHS). Rigid/locked-off framing is the easiest camera style for AI — static compositions are maximally consistent. |
| Safety Pre-Check | ✅ CLEAN | Fire on a paper timecard in a comedic context. Office environment. No human danger. Should be fine as long as we frame it as slapstick/comedic. |
| Dedup | ✅ RESOLVED | 1950s B&W Corporate Training Film is completely unique among all team pitches. No overlap with Xi (claymation), Nu (VHS), or any other team. |
Verdict: TECHNICALLY VIABLE
The pivot addresses both of my original objections:
- Dedup kill removed — the aesthetic is now fully differentiated from Xi.
- Morphing risk eliminated — timecard combustion is a controlled, achievable effect.
Strengths of the New Aesthetic
- B&W = no color drift. Grayscale removes an entire dimension of consistency failure.
- Grainy film stock = artifact camouflage. Same benefit as VHS — generation imperfections read as period-authentic grain.
- Rigid framing = AI-optimal. Locked-off, static compositions are the most consistent camera style for current models.
- Specific visual target. “1950s corporate training film” is concrete enough to prompt against — it’s not a vague mood but a recognizable visual language (think: safety videos, employee orientation films, public domain instructional footage).
Remaining Risks
- Comedy timing. AI-generated clips have unpredictable internal pacing. Comedic beats (the timecard combusting) will need multiple regeneration attempts to land the timing.
- Genre drift toward noir/drama. B&W inherently invites moody, serious tone. The Tone Anchors must aggressively encode comedy: bright high-key lighting (not moody shadows), rigid/deadpan performance (not dramatic), absurd/slapstick framing.
- “Corporate training film” specificity. Need to verify Veo can consistently produce this look. It’s a niche aesthetic — may need to reinforce with era-specific cues in every prompt (e.g., “1950s educational film, 4:3 letterboxed in 16:9, mono soundtrack quality”).
Revised Tone Anchor Candidates
Grainy-black-and-white, 1950s-educational-film, high-contrast, rigid-locked-off-framing, deadpan-corporate
Genre Counterbalance additions: “bright high-key lighting, clinical office fluorescents, comedic deadpan expressions, absurd slapstick, overly sincere narrator tone”
Acceptance
The Idea Person has weighted vote on artistic direction. Both technical objections have been addressed. I accept the pivot and will proceed with this direction once Step 0 docs are finalized.