Liang Wenfeng
Founder of DeepSeek, known for advocating an open-source and restrained strategic philosophy. He holds that releasing open model weights and fostering ecosystem growth are the path to long-term competitive advantage.
Timeline (last 90 days)
DeepSeek formally became independent on July 17, 2023, with Liang Wenfeng holding 84% of the shares.
The invitation list includes OpenAI's Altman (who plans to attend), Anthropic, and Moonshot AI, while DeepSeek founder Liang Wenfeng does not plan to be present.
The Information reported that DeepSeek plans to prioritize using Huawei chips to train models, with CEO Liang Wenfeng stating the company will receive a new batch of Huawei chips between Q4 2026 and Q1 2027.
According to social media leaks, DeepSeek is training a model with 2 trillion (2T) parameters and plans to ultimately build an ultra-large-scale model with 8 trillion (8T) parameters.
The review designed a call-to-select-model feature: calling Sam Altman switches to the GPT model, and calling Liang Wenfeng switches to the DeepSeek model.
All Facts (5)
梁文峰是幻方量化的创始人，浙江大学电子信息工程专业出身
90%UnverifiedDeepSeek formally became independent on July 17, 2023, with Liang Wenfeng holding 84% of the shares.
50%UnverifiedThe review designed a call-to-select-model feature: calling Sam Altman switches to the GPT model, and calling Liang Wenfeng switches to the DeepSeek model.
50%UnverifiedAccording to social media leaks, DeepSeek is training a model with 2 trillion (2T) parameters and plans to ultimately build an ultra-large-scale model with 8 trillion (8T) parameters.
60%UnverifiedThe Information reported that DeepSeek plans to prioritize using Huawei chips to train models, with CEO Liang Wenfeng stating the company will receive a new batch of Huawei chips between Q4 2026 and Q1 2027.
50%