Kerris: AI in Media & Entertainment Must Be Rebuilt Around

Nvidia's VP and GM of Media and Entertainment, Richard Kerris, has spent his career at the intersection of computing and storytelling. He believes that AI is a new compute model worth rebuilding around, and that its potential is being widely misread.

Kerris thinks that the industry's use of AI so far has been limited to using it like 'a better version of Google', as a feature to make tasks easier, but not yet leveraging its full capabilities. He reads the moment through a longer arc, one that began with Nvidia's RTX technology making rendering real-time eight years ago.

The point of RTX was never just about speed, but about enabling simulation and eventually achieving what we're doing today with AI. Generative AI then inherited the old bottleneck in a new form, prompting worked, but the heavy lifting still took time. Nvidia's Vera Rubin platform is built to eliminate this wait.

A partnership between Nvidia and Runway resulted in the ability to bring 4.5 software up in a day on Vera Rubin, with 'blowing away' results that were faster than generating scenes in minutes. This speed change has enabled generation to become a live creative loop, closer to conversation than production.

Kerris argues that AI dissolves the traditional pipeline relay of handoffs and can turn a studio's past into working material. He recommends rebuilding around one workflow at a time, dissecting it, building agents to address inefficiencies, and training the organization on the narrative behind the new technology.