The original AIoT was only vertical AI (artificial intelligence), not universal AI. AIoT powered by large models is the 'true AI'. On May 31, 360 (hereinafter referred to as "360") Smart Life Group held a visual big model and AI hardware new product launch conference. Zhou Hongyi, the founder of 360 Group, delivered a speech under the theme of "Big Models Opening a New Era of AIoT". He said that the emergence of large models marked the arrival of universal artificial intelligence, and AI completed the evolution from the perception layer to the cognitive layer. It is not only a disruptive revolution for traditional artificial intelligence, but also can promote the development of fields such as autonomous driving and robot control. The big model will bring about a new industrial revolution. "Zhou Hongyi believes that various software, apps, websites, and industries are worth reshaping with the big model, while intelligent hardware is a hardware based app. From the perspective of the development trend of large models, multimodality is a necessary path for the development of large models, and the most important change of GPT-4 is its ability to handle multimodality. Zhou Hongyi stated that the combination of multimodal technology and intelligent hardware is the trend. In the future, large models will become the brain of the Internet of Things, and IoT devices will act as the sensing end of large models, allowing them to evolve "eyes and ears". Large models may also manipulate IoT devices, evolve mouths, hands, and feet, and thus possess mobility, ultimately achieving a transition from perception to cognition, and from understanding to execution. At the meeting, Zhou Hongyi released the "360 Intelligent Brain Visual Model". He stated that the big language model is the foundation for building a visual big model, and the core of enhancing multimodal ability is to utilize the cognitive, reasoning, and decision-making abilities of the big language model. At the same time, the visual big model is also an important component of the "360 Smart Brain", enabling it to understand images, and in the future, it can also understand videos and sounds. It is understood that on the basis of its visual perception ability, 360 integrates the "360 Smart Brain" big model with hundreds of billions of parameters, conducts cleaning training based on billions of levels of internet graphic and text data, and fine-tuning security scenarios with millions of levels of industry data, Finally, we created a professional visual and multimodal model - the 360 intelligent brain visual model. At present, the capabilities of large models are mainly reflected in the software layer. When large models are connected to intelligent hardware, it will shift their capabilities from the digital world to the physical world, "Zhou Hongyi said. (New News Agency)