Abstract:
According to Bloomberg, ByteDance is preparing to launch an AI model for real-time spatial video generation to compete with Meta and Google parent company Alphabet in the fields of robotics and autonomous system applications. ByteDance founder Zhang Yiming is personally overseeing the development of the new model, which could be released as soon as next month, according to people familiar with the matter. People familiar with the matter said that Zhang Yiming has been coordinating the relevant work of the company's various business departments and allocating AI resources and computing power for this project. The model's release date is uncertain and plans are subject to change.
ByteDance
Informed sources said that the model is built on Bytedance’s existing movie-level video generation AI model Seedance and will allow users to create interactive virtual worlds for live broadcasts, short plays and games.
Zhang Yiming hopes that this latest AI model will allow the company to enter the rapidly developing world model field with its top professional capabilities in the video field. He will join Fei-Fei Li, Yann LeCun and others to explore the visual path of AI, a direction considered critical to robotics, gaming and self-driving cars.
ByteDance’s latest model aims to generate virtual worlds that respond to the Pico headset user’s voice or movements. A person familiar with the matter revealed that the model can generate on-demand video at a rate of 20 frames per second with a latency of approximately 0.05 seconds, and can serve spatial computing environments and interactive interfaces.
The Chinese model aims to reduce the upfront costs of VR adoption by moving the computationally intensive process of generating spatial content to the cloud. This would reduce the processing power required for the headset, opening a path to lower-priced devices with simpler hardware configurations. Rapid content generation may also help shift the battleground for users from hardware specifications to AI models, computing infrastructure and content distribution platforms.
If successful, this new model may open up a new battlefield for ByteDance to compete with Meta and Apple. Both Meta and Apple have invested heavily in virtual reality and mixed reality equipment. Meta has been trying to build a mass-market ecosystem around the Quest headset, while Apple has positioned Vision Pro as a high-end spatial computing platform. But neither company has been able to bring such products to the mainstream market.
As of press time, a ByteDance spokesperson has not commented on this.
Comments