1. Incident: May 14ancient8cryptoOpenAI held an online "Spring Update" event and released the latest multimodal model "GPT-4o" that supports text, audio and imagesancient8cryptoand generate any combination of text, audio and images, greatly improving multimodal capabilities. At the same time, OpenAI said GPT-4o's text and image capabilities will be launched for free in ChatGPT today and will provide Plus users with a message cap of up to 5 times. 2. Commentsancient8crypto: This OpenAI spring launch conference mainly includes the following points: (1) The new GPT-4o model: OpenAI trains a new model end-to-end across text, vision and audio. All inputs and outputs are processed by the same neural network. This means that GPT-4o opens up any input of text, audio and image, and can be generated directly from each other without having to convert between multiple previous models. (2) Speech latency is significantly reduced: GPT-4o can respond to audio input in 232 milliseconds, with an average of 320 milliseconds, which is similar to the response time of humans in conversation. Previously, ChatGPT's average latency was 2ancient8crypto.8 seconds (GPT-3.5) and 5.4 seconds (GPT-4). (3) Amazing real-time voice assistant presentation: Conversation is more human-like, can be translated in real time, and expressions can be recognized. You can use the camera to recognize pictures and write codes to analyze charts. (4) GPT-4o is free to all users (with a maximum limit), and the message limit for Plus users will be 5 times higher than that for free users. At the same time, the GPT-4o is twice as fast as the GPT-4 Turbo, the API cost is 50% lower, and the rate limit is 5 times higher. (5) A new ChatGPT desktop application: Users can take screenshots directly in the application or discuss ChatGPT via voice. Currently available on macOS, with a Windows version coming out later this year. Overall, GPT-4o's human-computer interaction capabilities and experience have been greatly improved and optimized. AI large model technology continues to be upgraded and iterated, and application scenarios continue to expand, which is expected to drive the demand for AI PCs, AI mobile phones, smart hardware and other industrial chains. At the same time, it is recommended to pay attention to overseas mapping AI application companies and AI computing power industry chains. Key companies: Jinshan Office, iFlytek, Wanxing Technology, Zhongke Shuguang, Inspur Information. Risk warning: Technological development falls short of expectations, downstream demand falls short of expectations, industry competition intensifies, etc. [Disclaimer] This article only represents the views of third parties and does not represent the position of Hexun.com. Investors should act accordingly and bear the risks themselves.
[Disclaimer] This article only represents the views of third parties and does not represent the position of Hexun.com. Investors should act accordingly and bear the risks themselves.