The article discusses the evolution of AI-generated video, highlighting a viral clip of Will Smith eating spaghetti that initially showcased poor quality, later improved with Google DeepMind's Veo 3 video generation model. The upgraded model produces more recognizable and photorealistic content, along with AI-generated sound that, while impressive, can be jarring for some listeners. Google revealed this model during its I/O 2025 conference, emphasizing its ability to include user-suggested dialogue, though accessibility comes with a hefty subscription fee. The moral implications of such realism in video generation remain under debate.
The clip, which captured quite the audience, utilized AI technology that has since evolved remarkably from its initial deranged output—a testament to advancing video generation capabilities.
A significant highlight during Google I/O was the introduction of the Veo 3 model, allowing for the creation of photorealistic videos with accurate sound and dialogue inputs.
Collection
[
|
...
]