At its annual developer conference this week, Alphabet Inc’s Google unveiled Veo 3, the latest upgrade to an AI-powered video-generating tool developed by Google’s DeepMind unit.
The new model not only produces high-resolution, movie-like visuals, but also includes realistic audio, such as ambient sounds, effects, and even voice dialogue, which makes it stand out from competing tools.
Google’s new model is different from OpenAI’s Sora, Meta’s MovieGen, or Runway’s Gen-4, none of which currently offer audio support. Instead, Veo 3 blends visuals with synchronized sound, which has sparked a wave of viral videos on social media. Many online users comment that it is difficult to tell whether what they are seeing is AI or reality.
“It’s amazing how well Veo 3 does at modeling intuitive physics,” Demis Hassabis, co-founder and CEO of Google DeepMind, said on Friday, adding that the advance could provide important insights into the computational complexity of the universe.
Tesla CEO Elon Musk responded: “Awesome. Good work.” Musk also expressed admiration for Google’s Veo 2 last year.
While Musk has been praising Google’s video model, he has previously expressed concerns about Google’s Gemini, accusing the model of bias.
Meanwhile, reports emerged earlier this week that Musk’s artificial intelligence company xAI is preparing a major investment to deploy 1 million GPUs near Memphis, which could benefit Nvidia significantly.
+ There are no comments
Add yours