When Nvidia Was Ready for Its Close-Up at GTC, Hollywood Was There
The biggest story in the business world this week is AI giant Nvidia and its multiday confab GTC in San Jose, Calif., where researchers, developers and other stakeholders from industries ranging from health care to automotive convened.
Hoping not to get lost amid all the hoopla surrounding the $2 trillion-plus company were representatives from the media and entertainment category, which Richard Kerris, vice president of developer relations and head of media and entertainment at Nvidia, pointed to its history in content creation and asserted that M&E is part of the “DNA” of the company.
“We emphatically believe [AI] is going to be a workflow enhancement if used correctly, if sourced correctly,” said Kerris in an interview with Variety Intelligence Platform. “With that comes a lot of responsibilities to ensure that content is sourced properly.”
Kerris is touching on both the promise and peril for media and entertainment in the AI world. In the week prior to GTC, Nvidia became the latest tech firm to be sued for alleged copyright infringement in the AI space. A group of authors sued the company, alleging their work was used to train the company’s NeMo language model without their permission.
But Nvidia founder and CEO Jensen Huang also firmly positioned his company at the forefront of the generative AI movement at GTC with a string of announcements, including the introduction of the company’s most powerful AI chip, Blackwell, which he hailed as “a processor for the generative AI era.”
Media and entertainment certainly had a presence at GTC with its own track of sessions devoted to the industry, featuring panelists from Pixar and WME and speaking to AI themes.
AI developers in the M&E space that flocked to San Jose are among what Kerris estimates to be roughly 2,000 startups such as Runway, which last June announced a $141 million extension to its Series C funding with participation from companies including Google and Nvidia.
Such startups investing in the Hollywood market believe AI — driven by enabling computing tech — holds the potential to increase productivity, enhance creativity and democratize content creation. Runway co-founder and CTO Anastasis Germanidis noted that each year there are millions of books published and thousands of films produced, but AI could close that gap. “[There’s] a whole new set of stories that have not been told in audiovisual form,” he said. “The barrier for that is really just the resources.”
Beyond the Silicon Valley confab, AI remains a polarizing topic in Hollywood — a central subject during the 2023 WGA and SAG-AFTRA strikes that is now critical in the IATSE and AMPTP Basic Agreement negotiations that began this month. These talks will impact talent such as cinematographers, editors and production designers.
At GTC, Ed Ulbrich, chief content officer and executive vice president of production at AI startup Metaphysic, acknowledged this divide, saying some Hollywood studios are “all in — and then just down the street at another major studio, we are heretics.”
Stakeholders believe AI can advance areas such as virtual production, having the technique be accessible to smaller studios by making it easier to create content for the LED walls; and digital actors, with a range of new possibilities for uses ranging from film to social media.
Metaphysic is developing tech for uses such as the aging/de-aging of actors, and Ulbrich cited the company’s previously announced work on Robert Zemeckis’ upcoming movie “Here,” which stars Tom Hanks and Robin Wright, as an example of what might be possible. Through the use of their AI tools, these actors would age throughout the movie, starting as teens and concluding in their senior years.
Cloud apps could be an attractive way to get new AI software tools into the hands of a wider base of users, but there are currently limitations.
Thobey Campion, founder of Lore Machine, a visualization system that came out of its private beta earlier this month, asserted that designing AI cloud apps is “an absolute unmitigated nightmare, but I think it’s worth it.” Ulbrich added that he’d like to see more commercially available cloud options that could support these tools.
During this period of rapid AI R&D, stakeholders in M&E see advantages to embracing technical standards and open-source tools.
A notable one highlighted at GTC is OpenUSD (Universal Scene Description), a 3D content creation and interchange file format that was developed at Pixar (and last month collected a SciTech Award from the Academy of Motion Picture Arts and Sciences).
The Alliance for Open USD — whose partners include Pixar, Adobe, Apple, Autodesk and Nvidia — used GTC to promote the framework, as did the Academy Software Foundation. At GTC, Nvidia extended it potential, announcing the ability to send OpenUSD scenes to the Apple Vision Pro.
