Fei-Fei Li, the renowned computer scientist and "godmother of AI," has launched a startup called World Labs. According to the Financial Times, the startup has already reached a valuation of over $1 billion within just four months. It is backed by major investors, including Andreessen Horowitz and Radical Ventures.
World Labs aims to develop AI that processes visual data with human-like reasoning abilities. This ambitious goal is akin to the capabilities of generative AI models like ChatGPT. Li's expertise in computer vision, a field focused on enabling machines to interpret visual information, is central to World Labs' mission. She is known for her pioneering work on ImageNet, a comprehensive visual database for object recognition research. She has previously led AI efforts at Google Cloud. She currently advises the White House task force on AI.
An anonymous venture capitalist familiar with Li's work explained that World Labs is developing a model to understand the three-dimensional physical world, including objects' dimensions, locations, and functions. This innovative approach is part of Li's broader vision of "spatial intelligence," where algorithms can extrapolate images and text into 3D environments and make predictions based on that data.
World Labs has completed two funding rounds, the latest securing around $100 million. Li co-founded the startup while on partial leave from Stanford University, where she co-directs the Human-Centered AI Institute.
In an April TED Talk, Li elaborated on the research areas her startup will advance, emphasizing the potential applications of spatial intelligence in robotics, augmented reality, virtual reality, and computer vision. The successful implementation of these technologies could revolutionize industries such as healthcare and manufacturing.
The rapid investment in World Labs reflects the growing interest of venture capitalists in ambitious AI ventures, spurred by the remarkable success of OpenAI’s ChatGPT, which quickly achieved a valuation exceeding $80 billion.
Commentaires