Inside Aronofsky's AI "1776" Series

Aronofsky's "On This Day... 1776" launches on TIME with SAG-AFTRA actors and DeepMind visuals

Google’s Project Genie lets you explore AI-generated 3D environments in real time

Martini brings cinematography controls to AI video production

What is a LoRA? Addy explains in plain terms

Kling 3.0 announced with native 4K and audio generation

AI generates visuals, humans handle creative roles. Google DeepMind's technology creates the imagery, but SAG-AFTRA actors perform all dialogue, a traditional writers' room led by Lucas Sussman develops narratives, composer Jordan Dykstra scored the series, and a full post-production crew handles editing, sound mixing, and color grading.

TIME Studios distributes on YouTube. Episodes drop weekly throughout 2026, each timed to the 250th anniversary of the Revolutionary War event it depicts. The first two episodes premiered January 29: "The Flag" dramatizes the Grand Union Flag raising on Prospect Hill, while "Common Sense" follows Benjamin Franklin encouraging Thomas Paine to write his famous pamphlet.

This continues the DeepMind filmmaker partnership. Aronofsky's studio previously produced Ancestra, which we covered in detail when it premiered at Tribeca 2025. Google DeepMind most recently premiered an AI short at Sundance 2026.

Create any environment from a prompt or reference image. Describe your world, define your character, and choose your exploration mode (walking, flying, driving, or anything else). Nano Banana Pro lets you preview and tweak the starting image before you jump in.

Navigate in real time as the AI generates the world around you. Move through environments at 20-24 fps at 720p resolution; the world reacts to your movements and actions as you explore.

Persistence and consistency across your session. Turn around and return to an area you visited 30 seconds ago and the model remembers what was there. Paint on a wall and it stays when you look away and come back.

Remix existing worlds or start from the gallery. Branch other creators' worlds by modifying their prompts, explore curated examples for inspiration, and download videos of your explorations to share.

Three systems power it under the hood. Genie 3 is DeepMind's autoregressive world model; Nano Banana Pro handles image generation for previews; Gemini interprets your prompts.

Current limitations are real. 60-second session cap, physics inconsistencies where characters walk through walls, noticeable control lag, and no promptable events like weather changes mid-session yet.

Step into generated scenes to compose shots. Virtual camera positioning lets you place yourself in the environment and frame shots using cinematography principles like focal length, camera height, and movement, rather than describing what you want in a text prompt.

Lens selection and movement controls replace prompting. Choose your lens, set your camera moves, reframe and reshoot within existing images or video footage without re-generating from scratch.

Model-agnostic with transparent per-second pricing. Access multiple AI video generators in one interface including Veo, Kling, Sora, and Minimax; choose based on quality needs and budget.

Built for team workflows. Real-time collaboration for sharing prompts, edits, and feedback; built-in timeline for rough assemblies; XML export to move projects into Premiere, Resolve, or other professional editing software.

Add "transparent background" or "PNG with transparency" to your prompt. The model will generate the image with no background from the start.

Works best with isolated subjects. Product shots, logos, characters, and graphics elements generate cleanly; complex scenes with multiple subjects may have issues.

Output format is PNG with alpha channel preserved. Ready to drop into your compositing or design software.

Use cases include compositing elements and product photography. Also useful for graphic design assets and overlay elements for video.