DeepSeek引发的AI创新和开源生态发展的思考

Thoughts on AI innovation and open source development:Lessons from DeepSeek

        武延军(中国科学院软件研究所 北京 100190)

         人工智能;基础软件;开源;基础设施;软硬件协同

        artificial intelligence (AI);system software;open-source;infrastructure;software-hardware collaboration

        在人工智能（AI）领域竞争激烈的重要时刻，DeepSeek发布了V3/R1等基础大模型产品，性能比肩国际领先机构OpenAI，不仅展现了中国AI领域科技创新的实力，更为全球AI发展提供了来自中国的创新路径：一是低成本训练推理，打破高端算力的垄断封锁，降低研发应用门槛；二是全栈、全系列的开源开放，支持按需自主部署，普惠各行各业。这一来自中国的技术创新和开源实践，值得学习借鉴。文章将从“以软补硬”“开源传播”“生态优先”3个步骤归纳DeepSeek的开源模式创新之处。同时，也从大模型入口、开源软件供应链、开源基础设施3个方面，分析当前我国AI开源创新仍然面临的问题和风险。最后从大模型操作系统创新、软件供应链保障、开源基础设施建设、软硬件协同发展4个角度，提出加强我国AI创新与开源软件基础能力的建议。

        At a critical moment of intense competition in the artificial intelligence (AI) field, DeepSeek has released foundational large language models (LLM) such as V3/R1, with performance comparable to leading international organizations like OpenAI. This not only demonstrates China's technological innovation capabilities in AI but also provides a Chinese innovative pathway for global AI development. Firstly, through low-cost training and inference, break the monopolistic barriers of high-end computing power and lower research and development thresholds. Secondly, through full-stack and comprehensive open-source strategies, support customizable and local deployment that benefits various industries. This technological innovation and open-source practice from DeepSeek deserves in-depth discussion and learning. This study summarizes DeepSeek's innovative model from three perspectives, namely, compensating hardware with software, acquiring users through open source, and ecosystem priority. Meanwhile, it analyzes the current challenges and risks for China's AI open-source innovation, in terms of ecosystem portal, open-source software supply chain, and infrastructure. The study concludes by proposing suggestions to strengthen China's AI technological foundation in four aspects, i. e., LLM operating system innovation, software supply chain governance, open-source infrastructure construction, and software-hardware collaboration.