At the 2025 Inclusion · Bund Conference main forum, Dr. Wang Jian, former Chief Technology Officer of Alibaba Group and Director of Zhijiang Lab, delivered a keynote speech, delving into the topic of "openness" in the context of artificial intelligence. He pointed out that artificial intelligence is undergoing a revolutionary transformation from code openness to resource openness, and emphasized that in the AI era, "openness" has become a key variable determining the direction of the industry.
The Debate on "Open Source": From Code to Resources
Wang Jian reviewed several important events that occurred at the beginning of 2025 to reveal the evolution of the concept of "openness." He mentioned that on January 13th, the United States released export controls on artificial intelligence model weights, but specifically excluded "open source" weights, which reflected different understandings of "open source." Subsequently, with the open-sourcing of models such as Qwen and DeepSeek, discussions on "openness" in the field of artificial intelligence entered a new stage. Wang Jian quoted Sam Altman's words—"OpenAI stood on the wrong side of history"—to emphasize that this was not just a strategic mistake, but a historical choice.
He pointed out that in the internet era, the open-sourcing of the Netscape browser was a watershed moment. However, in 1998, the concept of "open source" was officially established, referring specifically to open source code. However, in today's AI era, merely opening the source code is no longer sufficient to solve the problems.
Wang Jian believes that with the explosive growth of model size, opening data and computing resources has become crucial for driving industry development. He prefers to call today's "open source" as "open resources" (Open Resource), because its essence is sharing data and computing power, allowing individuals to avoid re-investing large amounts of resources.
Space Computing: The Next Big Opportunity for AI
Wang Jian further proposed that in the context of AI's "resource openness," space will become a huge new resource. He said that just as mobile phones were redefined as computers, the existence of communication, navigation, and remote sensing satellites will give rise to a fourth type of satellite—“computing satellites.”
He shared a breakthrough achievement from Zhijiang Lab: On May 14th this year, a constellation of 12 computing satellites named “Three-Body Computing Constellation” successfully entered orbit, marking the first time an 8B AI model identical to that on the ground was sent into space. This constellation achieved interconnectivity between satellites, ensuring data processing can be done anywhere a satellite reaches, thus breaking free from the limitation that data must be transmitted back to the ground.
Wang Jian explained that the constellation was named “Three-Body” to break the dilemma of the unsolvable “Three-Body Problem,” and through the collaboration of countless entities, it aims to share space resources together. He is confident that only by sending AI and computing power into space can humanity truly leave Earth and explore deep space.
