CRISTÓBAL VALENZUELA is no stranger to weird, scary creatures. As cofounder and CEO of Runway, a five-year-old New York City startup that develops AI tools for video, his company’s product has brought to life a parade of surreal on-screen characters, from shape-shifting dolls that melt into walls to dancing giants with contorted faces.
When I talk to Valenzuela on a winter afternoon, the conversation is about why some scary creatures—even the most menacing giants—aren’t really scary if you know how to destroy them.
“Sometimes a sling and a stone is all you need,” Valenzuela tells me.
We’re talking about slings and stones because of the imposing shadow stretching over Valenzuela’s company. Days before our chat, OpenAI, the $86 billion, Microsoft-backed juggernaut of generative AI, unveiled its latest creation: a text-to-video tool called Sora that essentially does what Runway does; in some cases, maybe better.
As with Runway’s product, Sora lets users type a description of a scene into
