Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ODCA startup established by self-driving pioneer Oliver Cameron And Jeff HawkCreates an AI model that allows users to “interact” with streaming videos.
A “Initial Demo” is available on the web, creating and streaming video frames per 40 milliseconds of model. Through basic controls, viewers can explore territories in a video like 3D-rendered video games.
“The current status of the world, an coming step and the history of state and activity, the model tries to predict the next situation in the world,” Explains the Odyssey in a blog postThe “It is strengthened by a new world model, showing efficiency of pixels, which feels realistic, maintains spatial continuity, outputs the learning verbs from the video and outputing the consistent video streams for 5 minutes or more.”
Several Startups and Large Technology Companies are chasing World modelWith DipmindThe dominant your researcher fee-VII Li’s World Labs, MicrosoftAnd DeacartThe They believe that world models can one day be used to make interactive media such as games and films and to play realistic simulations like the training environment for robots.
However, creatives have mixed feelings about technology. Ay Recent Wired investigations have shown that game studios like Activision Blizzard, which exceeded a number of workers, are using the angles and using the attitude of the fighting attraction. And a 2024 Study Operated by a union animation guild representing Hollywood animators and cartoonists, it has been estimated that over 5 US-based films, television and animation work will be disrupted by AI in the coming months.
As part of this, the Odyssey promises to cooperate with creative professionals – do not replace them.
“Interactive Video … A whole new entertainment door opens, where stories can be generated and explored, according to demand, free from the constraints of the traditional production production and expense,” the company wrote in its blog post. “Over time, we believe in all the videos today – entertainment, advertising, education, training, travel and more – all interactive videos driven by Odysse will develop.”
The Odyssey’s demo is a bit rough around the edges, which the company acknowledges in its post. The environment that produces the model is obscure and perverted and becomes unstable in the sense that their layouts are not always the same. For a while, move on one side or turn around, and the surroundings can suddenly look different.
However, this company is committed to rapid improvement, which currently spends “users-hours” from $ 1 to $ 2 per frame from the cluster of NVDia H1 GPUs.
“Looking at the front we are researching the more prosperous world presentation that captures more loyalty with more loyalty while increasing temporary stability and endless state,” the Odyssey writes in his post. “In parallel, we are expanding the action space from speed to world interaction, learning exposed verbs from large -scale videos.”
Many AI labs in the ODC World Modeling Space are taking a different approach than the labs. It is designed to be a 360-degree, backpack-mounted camera system To capture the real-world landscapes, the Odyssey thinks that only the public can serve as the basis of high quality models than trained models in public data.
To date, the Odyssey has collected $ 27 million from investors, including EQUE Ventures, GV and Air Street Capital. One of Pixar’s co-founders and former President of Walt Disney Animation Studios, Ed Katmul Startup, are on the board of directors.
Last December, Odyssey D It was working on the software that allowed the Creator’s models to load the equipment such as unrealistic engines, blender and adobe so that after the effects they could be edited by the hand.