Odyssey, an synthetic intelligence (AI) startup based final 12 months, shared particulars about its first AI product on Monday. The agency revealed that it’s constructing an AI video mannequin that may create Hollywood-grade visible results, identical to OpenAI’s Sora device that’s but to be launched by the corporate. Odyssey’s co-founder says that the AI mannequin will let customers edit and management the output at a granular degree, including that the agency is working with a number of giant language fashions (LLMs) to generate totally different layers of the output video, which will be managed individually.
How Odyssey’s AI Visible Mannequin Works
In a collection of posts on X (previously Twitter) Odyssey CEO and Co-Founder Oliver Cameron stated that the AI startup had raised $9 million (roughly Rs. 75.1 crores) in its seed spherical funding led by Google Ventures and was at the moment constructing a device that might ship high-quality video that might be customised and edited by customers.
Cameron additionally shared particulars about Odyssey’s AI know-how and claims that it was designed to generate “Hollywood-grade” video. The manager additionally stated that the startup was coaching 4 generative fashions to customers to take “full management of the core layers of visible storytelling”.
Individually, every mannequin will allow you to exactly configure the minutia of your scene.
Mixed, these fashions will generate video or scenes, however precisely as you wished.
Going additional, our mannequin outputs combine into present pipelines in use in Hollywood and past. pic.twitter.com/jHZoevLV9n
— Oliver Cameron (@olivercameron) July 8, 2024
Cameron highlighted the issue in present AI text-to-video fashions, which is the shortage of instruments or choices to manage or edit the output. “As a storyteller, you might have little capacity to direct your atmosphere or characters, or to iterate on the finer particulars of your shot till it is excellent. Extra highly effective fashions are required,” he added.
To unravel the issue, the corporate is utilizing a number of AI fashions that can generate a single layer of the composite video. As per Cameron, 4 fashions will independently generate geometry, supplies, lighting, and movement. These 4 layers can be generated concurrently based mostly on a single textual content immediate after which mixed to current the ultimate video.
The corporate claims that customers may have the choice to configure every layer individually for larger management over the output. Odyssey may also combine its video outputs into present Hollywood instruments and techniques used to generate visible results.
At the moment, the AI video mannequin is in its early growth stage. There isn’t any launch date for the AI mannequin. Nonetheless, the corporate has highlighted that it’ll maintain sharing common updates about its progress. Notably, Cameron has beforehand labored for Cruise and Voyage, two startups working with self-driving autos.
Jeff Hawke, the opposite Co-Founder and the CTO of the corporate was beforehand working because the Vice President of Expertise at Wayve, an AI agency which is growing autonomous driving techniques.
