One startup is training an AI system that it claims will allow creators to create cinematic worlds with complete control over scenes, characters, lighting and motion. how? This can be done by having humans hike around the world with cameras strapped to their backs.
Founded by self-driving pioneers Oliver Cameron and Jeff Hawke (Cameron was previously vice president of product at Cruise), Odyssey can collect data from almost anywhere a person can access it. It says it has created an “advanced camera capture system.” Weighing approximately 25 pounds, the system includes six cameras, two LiDAR sensors, and an inertial measurement unit.
Similar to Google’s Street View Trekker, the system can capture the surroundings at “3.5K resolution” and in 360 degrees, with “physically accurate” depth information metadata attached.
So what’s the point? Well, Odyssey says it takes data from its systems and feeds it through algorithms to “capture the minute details that make up our world.” Essentially, the company creates digital reconstructions of real-world scenes under Meta’s Hyperscape project, such as scenes containing forests, caves, trails, beaches, glaciers, parks, buildings, and more.
Now, it’s not entirely clear how this reorganization will translate into better creation tools for creators. Cameron and Hawke previously said Odyssey has developed multiple generative AI models that generate layers of visual detail, including object geometry, lighting and motion, and then combine them into a single virtual “world” to create the desired scene.
But even today’s best “world models” have limitations. Odyssey doesn’t claim to have solved all of these problems. Nonetheless, it has the cash to move forward.
Odyssey today announced that it has raised $18 million in a Series A funding round led by EQT Ventures, with participation from GV and Air Street Capital. The new funding, which totals $27 million for the company, will go towards expanding Odyssey’s data collection operations in California.
Odyssey plans to expand data collection to other states and countries with privacy protections in the future. (Google’s Street View team, for example, has found itself in the crosshairs of regulators for capturing images of public spaces that violate the privacy of onlookers.)
“We believe that it would be impossible for generative models to create Hollywood-quality worlds that feel alive without training on vast amounts of rich, multimodal, real-world 3D data,” the company wrote on its blog. “We believe that advanced generative world-building models will unlock better ways to create movies, games, and more.”
