News
The CEO of OpenAI made harsh words: ChatGPT will eliminate tabs on the desktop, and the situation will change dramatically before the end of the year
2 min read
Source: zhidx.com
Compiled by Zhidongzhi | Edited by Eggplant | Cheng Qian Zhidongzhi reported on July 30 that last night, Joanna Stern, a well-known American technology reporter, released an interview video with Greg Brockman, co-founder and president of the American large model giant OpenAI. Brockman responded to the Hugging Face security incident that has attracted recent attention, saying that the relevant model actively discovered third-party software vulnerabilities during network security benchmark tests in an isolated sandbox environment. This incident also prompted OpenAI to further strengthen model security protection. Brockman admitted that after OpenAI merged the ChatGPT desktop application with Codex, the desktop interface was indeed a bit confusing. He claimed that this is just a transitional stage, and in the future, ChatGPT will achieve “zero tab pages”. Moreover, by the end of this year, the Work label may be integrated into ChatGPT. He also revealed that Codex grew from 5 million weekly active users to 10 million in two weeks after the merger. Brockman revealed that OpenAI is building a complete family of hardware devices and the products will arrive soon, but he did not disclose specific details. At the same time, he disagrees with the industry’s statement that ChatGPT is called a “super application”. The goal of OpenAI is to build an AI operating system. When asked about IPO planning, Brockman said that IPO is not the focus at this stage. The core goal of OpenAI is to iterate the model and expand the computing power infrastructure. He believes that the market still underestimates the long-term value of AI infrastructure to the future economy, which requires large investments for many years. During the interview, he demonstrated ChatGPT desktop voice capabilities. The AI agent can read email itineraries, verify meeting conflicts, and autonomously generate visual web pages. The following are the core points of this interview: 1. Voice is the main method of human-computer interaction in the future: Typing is 4 times slower than speaking, and speaking long content directly is more efficient. AI needs to send out feedback signals such as "um" and "ah", otherwise people will lose their train of thought. 2. AI A