"GenAI can do in seconds what a Hollywood studio does at enormous costs"
"GenAI can do in seconds what a Hollywood studio does at enormous costs"
Oded Granot, head of the visual effects department at Hour One, spoke at Calcalist's AI conference: "Nearly a thousand animators work on an animated film, it's a complex and expensive process until two years ago when GenAI came along and made the process cheap and immediate."
"Generative artificial intelligence (GenAI) can do in seconds what a Hollywood studio does at enormous costs," said Oded Granot, head of the visual effects department at Hour One, at Calcalist's AI conference. Granot, a special effects artist who worked for 15 years in Hollywood on films and series such as Spider-Man and Game of Thrones, estimates that this new technology will fundamentally change the industry.
"This is a disruptive industry," he said. "We don't know how big the disruption will be, but we believe in it. Quality video is created through a complex and difficult process. Close to a thousand animators work on an animated film, and each animator produces just one to two seconds of footage a week. Each shot weighs more than 5,000 terabytes, and there are thousands of shots in a film. It's a complicated and expensive process. However, two years ago, GenAI made the process more affordable. I can describe everything I want—a location, an animal, a character, a camera movement, or a type of lens—in words, and it is created in a second at virtually no cost.
"However, artificial intelligence is not perfect, and there are many challenges. It's a kind of Russian roulette; every time you run it, you get a different result. This inconsistency is a significant problem for people like me who require consistency throughout the production process. Even in the shot itself, if I want to change the color of a balloon from red to blue, I may get a completely different result. The procedure is also not simple enough; you need prior knowledge to produce it. On the other hand, the industry is advancing at a rapid pace, with new models continually emerging to address these issues.
"At Hour One, we create animated characters and avatars, allowing people to produce content without standing in front of a camera. Reed Hoffman, the founder of LinkedIn, has a character in our system connected to a language model that can create content live. He is interviewed for television and delivers messages without investing a second of his life. We believe several developments will change the landscape of media and content consumption. The first is real-time video content production, and the second is content creation beyond mobile platforms. Soon, we will be able to produce vast amounts of live content. The volume of video we will see in the next few years is astounding, featuring avatars that know your personality and represent you.
"There are various questions that arise from this. With such an overwhelming amount of content, who owns my avatar? How do I prevent someone else from using it? What if I want to sell my avatar online? Alongside these concerns, there are questions about what is real. Why should we believe someone if you can make everyone say what you want them to say?"
"There are all kinds of questions that arise from this. Since there will be an insane amount of content, who owns my avatar? How do I prevent someone from using it? What if I want to sell my avatar online? Alongside this there are questions of what is real. Why can we believe If you can make everyone say what you want them to say?"
