Daily 9/3: World Labs Atlas, Reallusion AI Studio, Comcast VideoAI
World Labs unveiled Atlas, a world model that turns a single image into controllable 3D and video. Reallusion released AI Studio 1.1 as a standalone app, while Comcast Technology Solutions introduced VideoAI workflow applications for broadcasters and streaming providers.
- [SiliconANGLE] World Labs Unveils Atlas World Model ↗
World Labs, the startup co-founded by Fei-Fei Li, unveiled Atlas, a multimodal world model for spatial intelligence. Atlas can reconstruct a 3D scene from a single image, take camera paths and geometry as native inputs, and generate up to one minute of 1440p video. It can also output 3D assets such as point clouds and 3D Gaussian splats. World Labs points to creative visual effects and game design as use cases, while also positioning Atlas as a way to turn smartphone captures of real spaces into 3D simulations for robotics training.
- [PR Newswire] Reallusion Releases AI Studio 1.1 as a Standalone App ↗
Reallusion released AI Studio 1.1 and separated the tool from its earlier embedded position inside iClone and Character Creator. The update adds 30-second video generation with Seedance 2.5 and a lower-cost Seedance 2.0 Mini tier. It also supports 4K output across image creation, image editing, video creation, and 3D-to-video workflows. The upgraded AI Actor Creator lets users upload reference images to generate structured character sheets for more consistent visual identity across scenes.
- [PR Newswire] Comcast Technology Solutions Introduces VideoAI Workflow Apps ↗
Comcast Technology Solutions introduced a suite of AI workflow applications for broadcasters, content owners, and streaming providers. The applications cover video verticalization, regional-language localization, smart chapters, metadata enrichment, automated quality control, and intelligent clip generation. CTS says the tools can create scene- and asset-level metadata for semantic and natural-language search, and can help editors find sports highlights, social clips, or trailer material inside long-form video. The functions plug into the Cloud Media Processing engine and are expected to become commercially available through 2026 and 2027.
One line: World models, 3D-guided creator tools, and broadcast workflow AI are moving video automation from generation into production operations and distribution.