Martin Scorsese Joins Black Forest Labs: Storyboarding with FLUX
Melvin Vivas · X video post · 2026-06-03 · 0:30 · 46 views · Open on X
Topics: Industry Trends & Job Market, Prompt & Context Engineering · Level: beginner
Summary
Melvin Vivas reposts Black Forest Labs' announcement that director Martin Scorsese is now an advisor to the company. The 30-second clip shows a storyboarding session in which Scorsese describes a scene out loud and FLUX, Black Forest Labs' image model, generates frames from it. It is mostly industry news. The one practical lesson is how a director phrases scene prompts: setting, mood, specific details and camera angle.
Key points
- Black Forest Labs, the maker of the FLUX image models, has named Martin Scorsese as an advisor. They say they want human taste and craft at the center of 'visual intelligence'.
- Use case shown: using a text-to-image model (FLUX) to make storyboard frames for film pre-production.
- Scorsese's prompt pattern: give the setting precisely ('a town, not a village, not a city'), then the mood ('doesn't feel modern', 'almost medieval'), then concrete details ('narrow cobblestone streets', 'twisting main road'), then the camera direction ('put the camera higher, looking down').
- Scorsese compares AI storyboards to Cecil B. DeMille having production designers paint oil paintings of scenes.
- Trend: generative image companies are bringing in well-known creative professionals as advisors to shape their products.
- Note: the machine-generated transcript after 0:30 is a repeated-phrase glitch and has no content.
Resources mentioned
- FLUX · tool · github.com · check price
Black Forest Labs' family of text-to-image generation models. Some versions have open weights and some are only available through a paid API.
Also in: Evolution of Image LoRAs: Stable Diffusion → Flux → Krea 2 (Melvin Vivas on X · notes), Martin Scorsese Storyboards a Scene with Black Forest Labs' FLUX (Melvin Vivas on X · notes) - Black Forest Labs · website · bfl.ai · check price
The lab behind the FLUX image models; the speaker's team, which published Self-Flow.
Also in: Martin Scorsese Storyboards a Scene with Black Forest Labs' FLUX (Melvin Vivas on X · notes), Self-Flow: Training Multimodal Generative Models Without an External Encoder (Melvin Vivas on X · notes)
More in Industry Trends & Job Market
- SpaceX acquires Cursor
- Dreamina Seedance 2.0 Mini: Faster, Cheaper AI Video Generation
- Ideogram 4.0 Released as an Open-Weights Image Model
- Martin Scorsese Storyboards a Scene with Black Forest Labs' FLUX
- NVIDIA's LocateAnything: Vision-Language Detection Model (CVPR 2026)
- Opinion: GPT Image 2 Is the Current SOTA Image Model