中国开源的三大任务
大家早上好！
Good morning!
中国开源软件推进联盟成立至今已有20周年了，我们举办举世瞩目的《开源中国开源世界大会》今年也是19届了，在这段时间中，开源在中国、在全球均取得了很大的发展。
It has been 20 years since the establishment of the China OSS Promotion Union (COPU), and the world-famous Open Source China Open Source World Conference marks its 19 years this year. During this period, great progress in the field of open source at home and beyond has been made.
20年来开源联盟致力于完成三大任务。我们在国内，不遗余力地推动开源在中国的发展和传播，力图做好开源的布道者，团结开源的开发者、维护者、应用者和爱好者，支持开源的教育者、法律工作者；这些年来更围绕建设中国式的现代化及推动自主创新而努力奋斗。在国际，我们积极推动与开源的国际组织、社区的合作，创建“东北亚开源软件推进论坛”，协同抵抗美政府对我开源企业的无理打压。
For 20 years, COPU has worked to accomplish three major tasks. At home, we have spared no effort to promote the development and spread of open source software, try to be a good pace setter, unite developers, maintainers, users and enthusiasts, and support educators and legal workers in this field. In recent years, we have been striving to build a Chinese-style modernization and promote independent innovation. Internationally, we have actively promoted cooperation with international OSS organizations and communities, and established the Northeast Asia Open Source Software Promotion Forum to jointly resist the unreasonable suppression set upon our OSS enterprises by the US government.
我们致力于完成的三大任务是：
The three major tasks include:
一
解决“缺芯少魂”的短板，即自主研发、设计、制造为计算机、通信设备、 OS生态等配套需要的半导体芯片，海思已经攻克并完全掌握了14纳米以上的芯片设计的EDA,我们自主协同开发RISC-V开源架构。
First, to solve the "lack of chips" problem, that is, independent research and development, design and manufacture of semiconductor chips for computers, communication equipment, and other needs for sound OS ecology, Hisilicon has fully mastered the EDA chip design above 14 nm, and we independently co-developed RISC-V open source architecture.
经过近十多年来的努力，自主开发了鸿蒙、欧拉、龙蜥、统信、麒麟、中兴、xiaomi Vela等30多种优秀的操作系统，并在同一通用的社区版OS上(如龙蜥、欧拉)协同开发商业发行版(如与统信、麒麟、浪潮、SUSE合作等），防止操作系统无序分散和碎片化发展。
After nearly 10 years of efforts, we have independently developed more than 30 kinds of excellent operating systems such as Harmony, openEuler, Anolis, UOS, Kylin, My, xiaomi Vela, etc., and co-developed commercial releases (such as UOS, Kylin, Inspur, SUSE, etc.) on the same general community version of OS (such as openEuler and Anolis) to prevent disorder and fragmentation in operating systems.
在目前全球智能手机移动操作系统中，已形成安卓、ios和鸿蒙，三足鼎立的局面，现在有9亿台设备接入鸿蒙。
In the current global smartphone mobile operating systems, Android, IOS and Harmony are the three major pillars. And there are now 900 million devices that access HarmonyOS.
今年6月COPU组织了一次大模型LLM的操作系统与鸿蒙发展的讨论会，大家认为在当前数智时代，包括鸿蒙在内要与时俱进，在OS中增加智能模块，开发设计AI操作系统。银河麒麟敢为人先，8月6日发布首个OS的AI版本，使银河麒麟更具智能化的OS。
In this June, COPU organized a discussion on LLM operating systems and Harmony’s development. We believe that in the current era of digital intelligence, everyone (including Harmony) should keep pace with the times, increase intelligent modules in operating systems, and develop and design AI operating systems. Kylin was the innovator, releasing the first AI OS on August 6, so that the operating system can be more intelligent.
二
开发基于开源 的深度信息技术（如移动互联、物联网、云原生、区块链、人工智能等）
我们加入了人工智能适用于开源还是闭源的大讨论。
Second, develop deep information technologies based on open source (such as mobile Internet, Internet of Things, cloud native, blockchain, artificial intelligence, etc.)
We joined the debate about whether AI is applicable in open source or closed source.
三
建设、改造供应链，保障供应链和网络的安全运行。
Third, build and transform the supply chain to ensure the safe operation of supply chain and network.
我们这届开源大会讨论的主题是：
（一）发展基于开源的深度信息技术
（二）发表2024北京AI 安全国际共识
（三）发布2024中国开源发展蓝皮书
（四）由中国优秀大模型对标、平替或超越OpenAI声明所谓关停中国GPT4o API
（五）由阿里魔塔社区介绍医疗影像AI胰腺癌全球筛选
（六）由阿里云汇报负责巴黎奥运会全球转播和开发云上AI（及巴赫主席表彰）
（七）汇报RISC-V开源架构开发情况
（八）讨论供应链建设改造问题
（九）汇报部分操作系统开发、运维情况及由开源社区发行版向开源商业发行版转换合作情况
The topics of this forum include:
1. Develop deep information technologies based on open source
2. Publish the 2024 Consensus Statement on Red Lines in Artificial Intelligence
3. Publish the 2024 China Open Source Blue Book
4. Develop homegrown LLM to substitute GPT4o API (shutdown by OpenAI)
5. ModelScope introduces medical imaging AI for pancreatic cancer screening globally
6. Alibaba Cloud reports the global broadcast of the Paris Olympic Games and the development of cloud AI (and President Bach's commendation)
7. Report the development on RISC-V open source architecture
8. Discuss the construction and transformation of supply chains
9. Report on the development, operation and maintenance of some operating systems and the progress of distribution transitioning from open source community to open source commerce
在这些讨论主题中，作为亮点的是COPU与Linux基金会捍卫基于开源的人工智能，全球21位人工智能大师和专家发表“北京AI 安全国际共识”，我们发布了第三个年度的“2024中国开源发展蓝皮书”，中国的一批优秀的大模型完全可以对标、平替被OpenAI关停的API，以中国为主力的RISC-V开源架构成功开发，我们还可听到国际奥委会主席巴赫的重磅之言：若没有中国的全力相助，巴黎奥委会决然达不到现今的程度，中国的阿里巴巴作为全球领域的高科技企业之一，他们的云技术和人工智能解决方案提供了有力支撑，超过半数的全球转播面都用阿里巴巴的云服务8k传播，全球权威医学杂志《Nature Medicine》介绍的阿里云医疗影像AI胰腺癌全球筛选这个造福人类的AI项目，也作为亮点之一列入本届会议，还有供应链建设改造问题、操作系统开发、运维和开源社区发行版向开源商业发行版转换合作问题，也作为亮点列出。
Among these topics, the highlight were COPU and the Linux Foundation defending open source-based AI, 21 AI experts from around the world publishing the Consensus Statement on Red Lines in Artificial Intelligence, the third China Open Source Blue Book, a group of excellent China LLM which can completely substitute and replace the API shut down by OpenAI, and the successful development of RISC-V open source architecture (with China as the main force). We also heard the words of President Bach: China's contribution to the success of the Paris 2024 Olympics is manifold. China's Alibaba, as one of the high-tech enterprises in the global field, used its cloud technology and artificial intelligence solutions to provide strong support, and more than half of the global broadcast is transmitted by Alibaba's cloud service 8k. Also, the global authoritative medical journal Nature Medicine introduced Ali Cloud’s medical imaging AI for global pancreatic cancer screening, which is an AI project benefiting mankind. And the development, operation and maintenance of some operating systems and the progress of distribution transitioning from open source community to open source commerce are also listed as highlights.
本届大会重点之一是讨论基于开源的人工智能，评论生成式人工智能和通用人工智能，现将全球对此的异议罗列如下，供大家讨论参考：
有人非要用闭源来捆绑AI，势将束缚AI的发展；而开源将使AI具有更大潜力来提升其创造力和协同能力，而对AI发展尤为关键的安全更离不开开源，Meta的扎克伯格认为，“开源AI是最强大的模型）”，似乎不无道理。
One of the focuses of this forum is to discuss AI based on open source, generative AI and general AI. The global objections to these are listed as follows for discussion and reference:
Some people insist on using closed sources to bind AI, which will hinder the development of it. Open source will give AI greater potential to enhance its creativity and synergy, and security, which is particularly critical to the development of AI, is inseparable from open source. Meta's Zuckerberg holds a reasonable view that open source AI is the most powerful model.
目前在全球火红的是语言大模型生成式AI，在肯定其巨大进步同时也要议一下它存在的局限性，其中突出的是产生“幻觉”，一些AI大师认为作为语言大模型缺乏对世界底层结构（物理空间）的理解，使这个模型能否实现真正的智能质疑。有人提出应该关注发展依托于实体经济的“工业互联AI”大模型。人们对通用人工智能看作为AI的发展前沿似乎取得了共识，但也存在讨论的空间，如有人指出更现实的做法是首先发展服务于各行各业的专用人工智能。
At present, the most popular LLM globally is generative AI. While affirming its great progress, we also need to discuss its limitations, of which a prominent one is “illusion”. Some AI experts believe that as a LLM, it does not truly understand the underlying structure of the world (as in physical space), so that it might not represent “real intelligence”. Some people have proposed that attention should be paid to the development of a large model of "industrial connected AI" based on the real economy. There seems to be a consensus that general AI is the frontier of AI, but there is room for discussion, such as pointing out that a more realistic approach is to first develop specialized AI that serves various industries.
人类是否应在可能具有自主系统的通用人工智能建成前制定限制其摆脱人类控制的红线，似乎也有共识，进一步我们要研究和讨论的是：开源在制定这条红线中的作用如何？研究通用人工智能是否应做到安全第一、安全为先？全球同步？技治并举?
There seems to be a consensus on whether humans should set a red line limiting general AI that may have autonomous systems from human control before it is built. Furthermore, we need to study and discuss: What is the role of open source in setting this red line? Should the research of general AI put safety first, make even progress globally and be developed and managed at the same time?
我希望在本届大会上及以后，为大家的讨论开一个好头。最后，我预祝大会成功举办！谢谢。
I hope to make a good start to our discussions at this session and beyond. In conclusion, I wish the forum a great success. thank you.
