The news that Weng Li, the Chinese vice president of OpenAI, is about to leave the company has attracted widespread attention. The editor of Downcodes will explain in detail the career trajectory of this female leader who has made outstanding contributions in the field of artificial intelligence and her contribution to OpenAI. During his six years at OpenAI, Weng Li participated in a number of important projects, ranging from deep reinforcement learning to model security, and made indelible contributions to the development of artificial intelligence technology. The formula "Agent = large model + memory + active planning + tool use" she proposed provides important theoretical support for AI research.
Recently, OpenAI’s Chinese vice president Weng Li announced that he will officially resign on November 15. Weng Li graduated from Peking University. He was the head of the OpenAI security system team and one of the important contributors to ChatGPT.
She joined OpenAI in 2018 and participated in many important projects including GPT-4, especially in aspects of model security, pre-training and reinforcement learning.
During her career, Weng Li proposed the famous formula of "Agent = large model + memory + active planning + tool use", which provided important theoretical support for the research and application of AI. Weng Lizeng delivered a keynote speech on "AI Safety and Training" at the 2024B Station Super Science Night event, which received widespread attention and discussion.
In her resignation letter, Weng Li expressed her gratitude to OpenAI and reviewed her growth process in the company. She mentioned that she was full of curiosity and enthusiasm when she first joined OpenAI. She participated in multiple projects from deep reinforcement learning algorithms to robot challenges, and successfully taught the robot to solve the Rubik's Cube with the joint efforts of the team. She also said that when exploring the application of AI models in the real world, she established the first application research team and laid the foundation for subsequent security work.
Under her leadership, OpenAI's security system team has continued to grow and now includes more than 80 outstanding scientists and engineers. Weng Li is proud of the team’s achievements in AI safety, especially the newly released o1 preview model, which demonstrates excellent safety and practicality. Her strict requirements for the team in security testing and red team testing ensure the robustness and transparency of the model.
Although leaving OpenAI, Weng Li said she will continue to update her blog and looks forward to having more time to devote to programming and writing. Her departure attracted the blessings of many colleagues. It not only marked a new starting point for her personal career, but also meant that the key position of OpenAI security director would usher in a new candidate.
Although Weng Li's resignation is regrettable, her contribution to the field of artificial intelligence will forever be remembered in history. We look forward to her greater success in the future! The editor of Downcodes will continue to pay attention to her latest developments and bring more exciting reports to readers.