Posts on CrispShade: Latest AI and Tech News

AI Innovations Transforming Industries in Nanjing

Sat, 25 Apr 2026 00:00:00 +0000

AI Innovations in Nanjing

In Nanjing, the AI restaurant showcases a future where robots are chefs, making dishes and even customizing coffee art. This establishment is part of the Nanjing Artificial Intelligence Ecological Street, which exemplifies the city’s commitment to integrating AI into everyday life.

As the tech revolution accelerates, AI emerges as a strategic technology driving change. Nanjing is seizing the opportunity to lead in AI applications, addressing challenges in connecting innovation and industry, and fostering deep integration between technological and industrial innovation.

A Thriving AI Ecosystem

At the AI Ecological Street, a simple voice command can place an order for coffee or hail a ride, demonstrating the seamless integration of AI in daily tasks. The street features over 50 AI applications across various sectors, including healthcare, dining, and public services, attracting visitors from across the country.

The AI Hub serves as a showcase for cutting-edge AI products, facilitating transactions and product development. Nanjing’s focus on AI is evident, as city officials prioritize AI as a key driver for industrial transformation, emphasizing its importance across all sectors.

As AI becomes more intertwined with daily life, Nanjing is evolving from a “software city” to an “AI city,” enhancing its technological landscape.

Bridging the Gap in Technology Transfer

Nanjing’s mission includes achieving breakthroughs in technological innovation, leveraging its rich educational resources. With 52 universities and numerous research institutions, the city is well-positioned to convert academic research into marketable products.

To address the challenges of technology transfer, Nanjing has established the National Regional Technology Transfer Center for Biomedical Research. This initiative aims to facilitate the commercialization of research outcomes, providing support for funding and infrastructure.

Since its inception, the center has connected with 87 universities, facilitating the transfer of over 1,700 research projects, with 88 projects already commercialized.

Conclusion

Nanjing’s efforts to build an AI ecosystem and enhance technology transfer reflect its commitment to becoming a leader in innovation. By fostering collaboration between academia and industry, Nanjing is setting a precedent for other cities in integrating new technologies into their economic frameworks.

Empowering Women Through AI: APEC Workshop Highlights

Sat, 25 Apr 2026 00:00:00 +0000

Empowering Women Through AI: APEC Workshop Highlights

On April 25, 2026, a workshop on “Women-Friendly Artificial Intelligence” was held in Beijing as part of APEC’s initiatives. Representatives from various APEC economies, including Brunei, Canada, China, Indonesia, Malaysia, New Zealand, Papua New Guinea, Russia, the United States, and Vietnam, gathered to discuss gender equality in AI, the development of trustworthy products, and economic empowerment for women.

The workshop was organized by the Women’s Leadership Innovation Working Committee of the China National Innovation and Development Strategy Research Association. During the event, the “APEC Women-Friendly Artificial Intelligence Toolkit (Guide)” was launched. This toolkit outlines six key action steps, including governance from data to algorithms, security and privacy protection, women’s participation in design, inclusive design, monitoring and feedback, and accessibility.

Currently, AI is rapidly reshaping innovation and industrial structures, yet a digital gender gap persists. According to a report by the GSMA, women in low- to middle-income economies are 14% less likely than men to use mobile internet. Furthermore, women make up less than 30% of the global AI workforce, with only 12% being researchers.

Tong Xiaoling, Vice President of the China Public Diplomacy Association, emphasized the differences among APEC economies in digital infrastructure, cultural traditions, and legal environments. She called for enhanced digital and cultural exchanges to explore development paths suitable for each economy, ensuring AI serves as a positive force for empowering women and promoting social development.

Victoria Khava, head of the “AI Women Alliance” at the Eurasian Women’s Forum, noted that the workshop provided a valuable platform for individuals with diverse professional backgrounds and experiences to collaborate and promote innovation and cooperation in the Asia-Pacific region.

Sama Kutub, an assistant professor at the University of Auckland, encouraged APEC economies to establish fairer AI impact assessment mechanisms and ensure women’s critical roles in system design.

Selina Starling, CEO of the “Great Love Community” in Canada, stressed that the industry should ensure technology serves everyone, not just the interests of a few.

Chen Ling, a professor at Tsinghua University’s School of Public Management, urged APEC economies to integrate AI literacy into vocational training, establish mechanisms to monitor and correct algorithmic gender biases, and enhance platform rule transparency to ensure equitable opportunities.

He Shuwen, Deputy Director of the National Women’s Federation, highlighted China’s commitment to empowering women in the integration of AI across various sectors. She expressed China’s willingness to strengthen cooperation, deepen project alignment, and share experiences to promote equal development opportunities for women in the AI era across the Asia-Pacific region.

Xu Weixin, President of the China National Innovation and Development Strategy Research Association, stated that building women-friendly AI is an intrinsic standard for mature and responsible technology. He called for APEC economies to prioritize human-centered approaches, create closer cooperation mechanisms in women’s development and the digital economy, and deeply integrate gender equality into the core agenda of AI development.

Shanghai's Ambitious AI Development and Governance Strategy

Sat, 25 Apr 2026 00:00:00 +0000

Introduction

In April 2025, General Secretary Xi Jinping emphasized during his visit to Shanghai that artificial intelligence (AI) is a young field and a career for young people. He noted the rapid iteration of AI technology and its explosive growth, urging Shanghai to summarize successful experiences in nurturing the AI industry ecosystem and to enhance exploration efforts to lead in AI development and governance.

Nurturing a Thriving AI Ecosystem

Shanghai is committed to cultivating a rainforest-like industrial ecosystem that fosters innovation and growth among young talents. The city aims to deepen the integration of AI with various sectors, enhancing economic and social governance capabilities while ensuring development and security progress together, contributing a “Shanghai solution” to global AI governance.

Growth of AI Enterprises

The number of enterprises in the “Mosu Space” has increased from over 100 in 2024 to more than 200 in 2025, with over 20 potential unicorns valued at over 1 billion yuan. More than 60% of the city’s large model registration enterprises are concentrated in this area.

Strengthening Support for AI Development

To ensure that “good models do not lack computing power, good applications do not lack data, and good products do not lack chips,” Shanghai is systematically strengthening support for high-performance computing chips, quality data, and efficient computing clusters to lay a solid foundation for model iteration and embodied intelligence technology maturation.

Open AI Collaboration

In 2025, Shanghai launched the “AGI4S Mount Everest Plan” through the Shanghai Artificial Intelligence Laboratory, fully opening up channels for computing power, data, models, platforms, scenarios, projects, and talent cooperation, establishing a national hub for AI4S.

Policy and Talent Development

Shanghai has implemented several measures to enhance AI applications, including the establishment of an AI industry investment fund and the introduction of the first provincial-level local regulations to promote AI industry development. The city aims to create a fertile ground for the young generation to showcase their talents and deepen the integration of AI with industry development, social welfare, and urban governance.

AI Industry Achievements

By 2025, Shanghai had nearly 10% of the national intelligent computing supply capacity and approximately one-third of the country’s AI talent. The city has launched over 150 registered large models and leads globally in humanoid robot shipments, with several intelligent chips achieving breakthroughs. The AI industry in Shanghai has seen a production scale exceeding 637 billion yuan, marking a 39.5% year-on-year growth.

AI in Action

The eighth World Artificial Intelligence Conference was held in Shanghai in July 2025, showcasing the vibrant AI industry. AI applications are being integrated into various sectors, including manufacturing, healthcare, and urban governance, enhancing productivity and efficiency.

Innovative Talent Cultivation

Shanghai is optimizing talent cultivation and technological innovation paradigms to empower young people in exploring new production relationships. The Shanghai Chuangzhi Academy is pioneering an integrated education model that combines research, innovation, and practical application, allowing students to engage in real-world projects.

AI Empowering Research

The Shanghai Artificial Intelligence Laboratory is also facilitating breakthroughs in research paradigms, such as the development of a unified multimodal model that topped the Hugging Face trends list. The lab’s initiatives aim to enhance collaboration and innovation across various fields, including energy storage and scientific research.

AI Governance and Compliance

Shanghai is committed to balancing innovation with regulation, continuously strengthening legal protections and governance collaboration. The city has established guidelines for electronic evidence collection in AI-related criminal cases, promoting a new model of judicial governance that integrates technology and law.

Global AI Governance Contributions

Shanghai is actively participating in international AI governance, establishing cooperation mechanisms with 38 countries and promoting dialogue and collaboration in AI governance. The city aims to bridge the gap between local innovations and global standards, contributing to a more connected and responsible AI ecosystem.

Conclusion

Shanghai’s comprehensive strategy for AI development and governance positions it as a leader in the global AI landscape. By fostering innovation, nurturing talent, and actively engaging in international collaboration, Shanghai is paving the way for a sustainable and responsible AI future.

Shanghai's Vision for AI Development and Governance by 2025

Sat, 25 Apr 2026 00:00:00 +0000

Introduction

In April 2025, President Xi Jinping highlighted that artificial intelligence (AI) is a young field, driven by young people. He emphasized the rapid iteration of AI technology and its explosive growth, urging Shanghai to leverage its successful experiences in cultivating the AI industry through a large model ecosystem.

Fostering a Thriving Ecosystem

Shanghai is committed to nurturing a rainforest-like industrial ecosystem that supports innovation and allows the younger generation to thrive. The city is expanding its AI capabilities by enhancing the synergy between AI and economic governance, aiming to contribute a “Shanghai solution” to global AI governance.

Growth of AI Enterprises

The “Mosu Space” in Xuhui District has seen the number of resident companies grow from over 100 in 2024 to more than 200 in 2025, with over 20 potential unicorns valued at over 1 billion yuan. More than 60% of the city’s registered large model enterprises are concentrated here.

Strengthening Support for AI Development

Shanghai is reinforcing its support for AI development by ensuring that good models have sufficient computing power, applications have adequate data, and products have the necessary chips. The city is focusing on high-performance computing chips, quality data, and efficient computing clusters to lay a solid foundation for the evolution of large models and embodied intelligence technologies.

Comprehensive AI Collaboration

In 2025, Shanghai launched the AGI4S Mount Everest Plan through the Shanghai AI Laboratory, which opens up channels for collaboration in computing power, data, models, platforms, scenarios, projects, and talent.

Policy Support and Talent Development

Shanghai has implemented various measures to enhance AI applications, including investment funds and local regulations to promote AI industry development. The city is focused on creating a fertile ground for young innovators and integrating AI with various sectors to improve governance capabilities.

AI as a Young Field

AI is at a crucial juncture, transitioning from laboratory experiments to real-world applications. The city’s industrial ecosystem is still forming, presenting opportunities for innovation and collaboration on a global scale.

Building a Strong Foundation for AI

Shanghai’s industrial base, diverse consumption scenarios, and rich talent resources position it well for AI development. The city is creating a complete ecosystem that includes policies, funding, computing power, data, and space to support the growth of AI.

Focused Innovation Strategies

Shanghai is concentrating on breakthroughs in large models, which are essential for advancing AI. The city is implementing various plans to enhance innovation capabilities and promote the application of large models across industries.

Recent Developments in AI

Shanghai has recently seen significant advancements in AI, with the number of registered generative AI services increasing rapidly. The city is fostering a collaborative environment that encourages innovation and supports the growth of AI enterprises.

Empowering Young Innovators

Shanghai is optimizing its talent cultivation and innovation paradigms to empower young people in the AI field. The Shanghai Institute of Intelligent Technology is focusing on integrating research and innovation to nurture top talent.

Enhancing AI in Public Services

AI is becoming increasingly important in addressing the challenges of an aging population in Shanghai. The city is developing smart elderly care services to meet the growing demand for high-quality, personalized care.

AI Governance and Regulation

Shanghai is committed to balancing innovation with regulation, establishing a legal framework to support AI development while ensuring safety and compliance. The city is actively contributing to global AI governance through international cooperation.

Conclusion

Shanghai is positioning itself as a leader in AI development and governance, leveraging its strengths to create a sustainable and innovative AI ecosystem. The city aims to serve as a model for AI governance worldwide, showcasing China’s commitment to advancing technology responsibly.

Transforming Education in the Age of AI: Insights from Experts

Sat, 25 Apr 2026 00:00:00 +0000

Transforming Education in the Age of AI

In the era of artificial intelligence, the foundational logic of education is being rewritten. On April 17, a closed-door seminar organized by the Beijing News gathered experts from universities, primary and secondary schools, research institutions, and educational enterprises to discuss the paradigm shift in teaching and learning in the AI age.

Experts at the seminar believe that AI is forcing a systemic transformation in education, shifting the teaching paradigm from a binary model of “teacher-student” to a triadic collaboration of “teacher-machine-student.” Teachers are evolving into designers of learning ecosystems, while students become technological collaborators. However, ethical risks cannot be overlooked, necessitating reforms in traditional teaching and evaluation systems, as well as the establishment of multi-layered prevention mechanisms.

Reshaping Teaching Paradigms Towards Triadic Collaboration

On April 10, five departments, including the Ministry of Education, jointly issued the “AI + Education Action Plan.” In response to this policy, Bao Haogang, deputy director of the Digital Education Research Institute of the Chinese Academy of Educational Sciences, stated, “In the digital intelligence era, the boundaries of human capabilities in creating tools are being redefined, leading to profound changes in social division of labor. The goal of education must shift towards cultivating talents who can harness AI and face the future, with a greater emphasis on the return to human values.”

The arrival of the AI era is compelling education to undergo systemic changes. What will happen to courses and classrooms when AI can grade essays, generate exam questions, and act as teaching assistants? Changes are already evident in higher education. Wang Boyue, a professor at the School of Artificial Intelligence at Beijing University of Technology, observed that programming assignments previously completed by first-year students, which focused on simple interfaces and basic functions, have shown significant improvement in completion and innovation since last year. “The interface and function design have become more sophisticated, with many first-year students able to fine-tune personalized vertical domain models using AI tools.”

Wang believes that the traditional classroom model, which primarily relies on PPT lectures and basic coding instruction, is being reshaped. Teachers are no longer just explaining knowledge points and code details; they are now posing questions, designing ideas, organizing discussions, and guiding students to use AI tools to achieve their goals. Practical classes have shifted from writing basic code to designing high-quality prompts, quickly implementing functions, and continuously iterating and optimizing solutions, allowing students to focus more on problem analysis, system design, and innovative practice. “This also raises higher demands for teachers’ digital literacy, teaching innovation capabilities, and ability to harness AI tools.”

Wang Mingtao, director of the Information Center at Beijing Information Science and Technology University, pointed out that with rapid technological advancements, teachers can no longer rely on traditional knowledge transmission methods for teaching. Traditional examination and evaluation methods have also become outdated, necessitating reforms in how students and teachers are assessed. He revealed that Beijing Information Science and Technology University is revising its training programs to incorporate AI elements into every major.

“As AI enters the classroom, the role of teachers as knowledge authorities is being challenged, but this does not diminish their role; rather, it catalyzes a profound evolution of their responsibilities,” said Zhang Yue, director of the Information Center at Beijing No. 18 Middle School.

Zhang emphasized that teachers must transition from traditional knowledge authorities and lecturers to designers of learning ecosystems and facilitators of cognitive collaboration processes. Students’ learning paradigms will also change, evolving from passive recipients of knowledge to active explorers and technological collaborators. Students need to master skills for efficient and critical collaboration with AI, including formulating precise instructions, questioning and verifying information authenticity, and synthesizing diverse viewpoints, while actively constructing knowledge through solving real and complex problems.

Shiyuntao, vice president of Beijing Industrial Vocational Technology College, believes that the enhancement of teachers’ capabilities depends on the transformation of educational infrastructure. Without established computational power in classrooms and large model platforms in schools, it is challenging for teachers to achieve significant improvements. He metaphorically stated, “The vehicle is already an electric car, but the road is still a dirt path.”

Preventing Ethical Risks Associated with AI

The “AI + Education Action Plan” emphasizes the need to effectively prevent issues such as AI-generated fraud, academic dishonesty, examination pressure, and privacy breaches. The ethical risks posed by AI have become a focal point of discussion at the seminar.

This issue is equally significant in primary and secondary education. Bao Haogang disclosed data from a nationwide survey conducted by the Chinese Academy of Educational Sciences, covering 31 provinces and over 650,000 samples. The results showed that 99.7% of surveyed students had encountered AI, and 85.6% had attempted to use AI while doing homework, indicating a situation that exceeds expectations but also carries certain risks.

He further pointed out that while establishing technological firewalls, education must undergo systemic reform. Traditional knowledge-based examinations and assignments should not be used to evaluate students. Instead, tasks should be assigned from a problem-solving perspective, involving non-structured, complex scenarios where students can use AI but should not let AI provide direct answers. Instead, they should “collaborate” or even “argue” with AI to cultivate their ability to harness AI effectively.

Bao Haogang particularly emphasized the importance of regulation. He believes that unlike adults who possess complete knowledge systems and can use AI critically, middle and primary school students have yet to establish their cognitive frameworks. Current research indicates that early reliance on AI may lead to distortions in cognitive development, attention, and innovation capabilities.

Wang Mingtao from Beijing Information Science and Technology University advocates for a positive and cautious attitude towards technology, embracing the opportunities it brings while also mitigating risks. In addition to technological regulation, cognitive guidance from the perspective of curriculum ideology is essential, with parents and teachers participating in correctly guiding children in using AI.

Zhang Yue shared the practice from No. 18 Middle School, which has standardized AI usage into three lists: the “Sovereignty List” clarifies that ultimate evaluation and decision-making power regarding values always belongs to teachers; the “Prohibited List” delineates behaviors that are absolutely forbidden, such as inputting private data and delegating core thinking processes; and the “Audit List” requires documentation of AI-assisted processes for review. They also iteratively implement the student-initiated “Generative AI Application Initiative,” where each graduating class upgrades and passes the initiative to incoming first-year students, forming AI teams for supervision.

Zhang emphasized that in the triadic ecosystem, AI is responsible for resource generation, preliminary data analysis, and process automation, but all its actions must operate within the educational framework and ethical boundaries set by teachers. “AI lacks emotional agency and ultimate value judgment, which are exclusive human capabilities.”

In her view, the collaboration between “teachers” and “AI” hinges on establishing clear responsibilities and collaboration interfaces. She cited that in practice, No. 18 Middle School particularly emphasizes “predefined roles and dynamic switching.” For instance, during the design phase of project-based learning, teachers clearly delineate the “green development zone”—tasks such as designing scientific experiments and making ethical decisions must be completed by students without AI assistance.

Yang Wei, general manager of Heweo Beijing, suggested adopting a youth model similar to gaming platforms, restricting minors’ AI usage time and functions through real-name authentication.

Bao Haogang stressed that the development of technology should allow for controlled trial and error and discussion, avoiding the pitfalls of over-caution or blind application, with risk governance dynamically advancing alongside the deepening application.

Promoting AI + Education from Pilot to Replicable Models

“AI has a particularly significant impact on vocational education, as the barriers to software development have lowered, greatly affecting software programming careers,” shared Shiyuntao, vice president of Beijing Industrial Vocational Technology College. He noted that new digital occupations are emerging, such as industrial robot system operators and data cleaning specialists. “To meet the new requirements for vocational talents in the industry, many vocational college students are trained in simulated scenarios of family services and intelligent manufacturing, wearing virtual devices for training.”

“For example, in high-end machine tool operation skills, we capture multimodal data from videos, paired with textual explanations, transforming them into digital resources that students can access anytime through AI for learning.” He stated that vocational colleges in Beijing are no longer just training traditional electricians, fitters, and welders. “In factories without manual labor, warehouse AGV vehicles (automated guided vehicles) are entirely controlled by software and code, and students must possess capabilities in intelligence, networking, and digitization.”

Shiyuntao introduced that their college is one of the 60 benchmark schools under the “Double High Plan,” and last year invested heavily in computational power and digital infrastructure, collaborating with Tsinghua University’s Zhipu Qingyan team to create a vertical model for industry-education integration covering aerospace equipment manufacturing and other industrial chains, establishing a new digital education ecosystem for cultivating “high-end digital craftsmen.”

Wang Mingtao shared experiences from Beijing Information Science and Technology University in building an AI ecosystem: promoting learning through competitions, facilitating research through management, and fostering interaction between teachers and students. The university is also one of the first pilot schools for the future smart academy construction in Beijing, creating a trend of valuing and applying AI from top to bottom. The intelligent hardware “AI Bistu” developed by student clubs has appeared in various scenarios, including enrollment promotion, campus open days, trade fairs, and the Beijing Science and Technology Expo, garnering widespread social impact.

In the education sector, the application of AI has transitioned from initial exploration to real-world implementation. How to create high-value, replicable application scenarios? Wang Mingtao pointed out that the current integration of AI technology and education is still insufficient, with many applications remaining superficial. He suggested that the implementation of the action plan should focus on comprehensive AI literacy education as the foundation for all application scenarios, while also selecting and nurturing typical AI application scenarios across various educational stages for replication and promotion citywide.

Bao Haogang noted that the action plan specifically mentions “building national AI (education) application pilot bases” to scale up small-scale innovations, identifying high-value, replicable scenarios that bridge industry, academia, and research. “Teachers should be encouraged to take the lead in trials; their experiences and feedback are crucial for assessing the value of scenarios.”

Wang Boyue suggested that to enhance teachers’ enthusiasm for using AI to drive educational innovation, real-world corporate scenarios, practical projects, and industry demands should be integrated, optimizing and improving the teacher assessment and evaluation system, guiding higher education teachers to participate in course design and teaching system construction, deepening industry-education integration, and ensuring the successful implementation of the “AI + Education” action plan.

Yang Wei candidly stated that the development of vertical large models for education is relatively lagging; many general large models exist, but there are few specifically designed for educational scenarios. He called for more enterprises to participate in the development of educational vertical large models, as having more models in the education vertical will foster competition among enterprises, leading to continuous self-improvement and promoting ecological prosperity.

2026 World Internet Conference Asia-Pacific Summit Held in Hong Kong

Fri, 24 Apr 2026 00:00:00 +0000

Dynamic Innovation in Hong Kong

The 2026 World Internet Conference Asia-Pacific Summit was recently held in Hong Kong. The theme of the summit was “Empowering Innovation through Digital Intelligence - Building a Community of Shared Future in Cyberspace Together.” Nearly a thousand guests from over 50 countries and regions shared insights on cutting-edge topics such as intelligent innovation, AI safety governance, and digital health, exploring new paths for digital development and opportunities for cooperation in the Asia-Pacific region.

Seizing New Development Opportunities

Participants unanimously agreed that artificial intelligence is a significant driving force behind a new round of technological revolution and industrial transformation, profoundly impacting global economic development and human civilization. All parties should work together to deepen open cooperation and ensure that the benefits of AI development are shared among people across different countries and regions.

Zhuang Rongwen, Chairman of the World Internet Conference and Director of the Cyberspace Administration of China, stated that the new round of technological revolution and industrial transformation is accelerating, with new technologies such as AI, big data, and cloud computing becoming key to restructuring global resource allocation, reshaping industrial ecosystems, and redefining the global economic landscape. Countries in the Asia-Pacific region are increasingly viewing digital transformation as a common choice to seize new development opportunities and create competitive advantages.

Ministers from over ten countries, including Afghanistan, Saudi Arabia, and Tonga, emphasized the need for collaborative innovation to ensure that the benefits of digital and intelligent development serve the people better.

The CEO of the Global System for Mobile Communications (GSMA), Hong Yaozhuang, noted that AI technology has rapidly evolved from a tool for processing information to a large-scale application in just over a decade. Future efforts should focus on enhancing cross-sector collaboration to promote innovation and share results.

Financial Secretary of the Hong Kong SAR, Paul Chan, remarked that the internet is transitioning from “digital connectivity” to “intelligent connectivity.” The “intelligent era” holds vast development potential, and all parties need to strengthen communication, share experiences, and expand practical cooperation to ensure that technological progress remains on a sustainable, responsible, and inclusive path. Hong Kong is willing to deepen cooperation with international partners to seize this historical opportunity.

Gathering Diverse Experiences

As AI deeply integrates into various industries, issues such as cybersecurity and personal information protection have become focal points of the summit. How to address these new challenges? The summit gathered diverse experiences and wisdom to strengthen cybersecurity, data protection, and AI safety governance, which became a consensus among participants.

“While AI has aided decision-making in many fields, it should be controlled by humans, and final decisions must be made by people,” Chan explained. In Hong Kong, the “sandbox mechanism” has become a hallmark of cross-industry regulation, allowing regulatory bodies to collaborate with innovators to test new tools in a controlled environment, identify risks early, and provide timely and practical feedback.

Mohammed Al-Shuwaier, Deputy Minister of Industry and Mineral Resources of Saudi Arabia, emphasized the need for transparency, accountability, privacy protection, and security when introducing AI into public services, businesses, and industrial value chains. AI should enhance human capabilities rather than create chaos or undermine trust. He called for practical cooperation in technology research and standard-setting to build a correct and healthy ecosystem.

“The ethical risks and technological uncertainties brought by AI require the international community to reach a consensus on governance principles,” said Wu Dong, Chief Engineer of the Cyberspace Administration of China. He advocated for cross-border and cross-sector communication and collaboration to collectively address the challenges posed by AI, ensuring algorithm fairness and data transparency, so that technological development always promotes human welfare.

Highlighting Unique Advantages

Hong Kong hosted this event for the second consecutive year and organized multiple innovation and technology-themed activities during the summit, further showcasing its unique advantages as an international innovation and technology hub.

“AI has become one of the key industries prioritized for development in Hong Kong,” said Zhang Manli, Deputy Secretary for Innovation and Technology of the Hong Kong SAR Government, at a special summit event for government-business exchanges. Leveraging its international business environment and data flow advantages, Hong Kong is developing into an AI hub that connects global technological innovation trends. The SAR Government’s budget for the fiscal year 2025/2026 explicitly proposes the establishment of the Hong Kong AI Research Institute to promote upstream research, result transformation, and application scenario expansion.

Peng Wenjun, Executive President of the Office for Attracting Key Enterprises, stated that over the past few years, the SAR Government has implemented comprehensive policy support from startup research and development to advanced manufacturing and market entry. The office has fully cooperated to provide one-stop services for enterprises, aiming to establish Hong Kong as an international innovation and technology center. Data shows that the office has successfully attracted over 120 key enterprises, expected to drive an investment of HKD 73 billion and create approximately 25,000 jobs.

Sun Dong, Secretary for Innovation and Technology of the Hong Kong SAR Government, remarked that in the wave of digital transformation, technological innovation and international cooperation are essential for sustainable development. Hong Kong has a solid cooperation mechanism, serving as both an international financial and technological center while playing a vital role in the construction of the Guangdong-Hong Kong-Macao Greater Bay Area. In the future, Hong Kong will fully leverage its unique role as a “super connector” and “super value adder” to promote regional collaborative innovation, helping the Greater Bay Area become a world-class innovation hub.

AI Enhances Mapping Intelligence

Wed, 22 Apr 2026 00:00:00 +0000

AI Enhances Mapping Intelligence

As an essential tool for understanding and recording Earth’s spatial information, maps are undergoing profound changes under the influence of artificial intelligence. On the occasion of the 57th Earth Day, Wang Jiayao, an academician of the Chinese Academy of Engineering and a pioneer in cartography and geographic information engineering in China, published an article titled “Empowering Mapping Science with Artificial Intelligence” in the Journal of Surveying and Mapping. The article systematically discusses how AI is driving mapping science into a new phase of digital intelligence, aiding the modernization of harmonious coexistence between humans and nature.

Wang Jiayao has dedicated 70 years to teaching and researching mapping-related fields. He is one of the founders of computer cartography in China, having established the country’s first computer cartography program and created the first computer-generated topographic map. He has experienced and led every significant transformation in Chinese cartography. Over the past 70 years, he has not only trained numerous talents in mapping and geographic information science but has also promoted China’s mapping science from imitation to independent innovation with a rigorous scientific spirit. He firmly believes that maps, along with music and painting, are among the three universal languages of humanity, representing one of mankind’s greatest innovations.

In his paper, Wang Jiayao and his research team, grounded in the context of rapid AI technological advancement, propose a critical judgment: the deep integration of AI and mapping science will propel the digital intelligence transformation of mapping science into a new stage. They elaborate on three specific pathways through which AI empowers mapping science. First, the integration of AI and brain science will accelerate the deepening of foundational theoretical research in mapping science, making cartography smarter. Second, the latest advancements in brain-like intelligence and computing provide strong technical support for overcoming the bottleneck issues in the digital intelligence process of mapping science. Finally, the rapid development of deep learning and generative AI opens up broader application spaces for digital mapping. He believes that this empowerment is not merely an upgrade of tools but a systematic transformation of map design, cartographic processes, and service models.

Wang Jiayao emphasizes the core concept of “human-centered” design, proposing that digital mapping must strengthen human-machine collaboration. He argues that regardless of technological advancements, maps remain products of human understanding of the world, with their embedded knowledge, culture, and logical reasoning relying on human creativity. AI should serve as a powerful assistant to humans rather than a replacement. The future of digital mapping should achieve a combination of “human brain wisdom + computer intelligence,” leveraging AI’s strengths in data processing and rapid computation while retaining human authority in value judgment and aesthetic expression.

Wang Jiayao believes that empowering mapping science with AI is a strategic, long-term, and sustainable systemic project. In various fields such as natural resource investigation and monitoring, land spatial planning, ecological civilization construction, disaster emergency response, and smart city management, high-quality and efficient mapping services are indispensable foundational supports. Through close collaboration with AI, mapping science can more quickly, accurately, and intelligently reflect changes in natural resources, provide ecological risk warnings, and assist in scientific decision-making, thus providing solid technical support for safeguarding a beautiful China. Maps are not only the infrastructure for national governance and social operation, recording the changes of the nation, but also the core carrier of temporal and spatial information, embodying the collective memory of the nation.

He once stated, “We have a civilization history of four to five thousand years, and maps have accompanied this history. This is the result of countless scientists’ efforts and the foundation of our cultural confidence.” He hopes to make maps a source of pride for every Chinese person.

Now, at nearly 90 years old, Wang Jiayao remains active in research, continuously focusing on and promoting the intersection of AI and mapping science. Looking ahead, as cutting-edge results are implemented, people can expect to receive intelligent mapping services that “understand you.” Under the empowerment of AI, mapping science is increasingly safeguarding our only home in a smarter and more precise way, contributing to the harmonious coexistence of humans and nature in a beautiful new landscape of China.

Empowering the Real Economy with Artificial Intelligence

Tue, 21 Apr 2026 00:00:00 +0000

Empowering the Real Economy with Artificial Intelligence

Since the beginning of this year, open-source AI agents have gained popularity in the global tech scene, transitioning AI from merely ’talking’ to ‘doing’, accelerating its integration into production and daily life.

The 14th Five-Year Plan outlines the comprehensive implementation of the ‘Artificial Intelligence +’ initiative. The State Council’s opinion on deepening this initiative states that by 2027, AI will be widely integrated into six key areas, with the application rate of new generation intelligent terminals and agents exceeding 70%. The deep integration of AI with the real economy will leverage rich data resources, diverse application scenarios, and a large user base, transforming them into unique advantages and strong momentum for building a modern industrial system and achieving high-quality development in China.

To empower the real economy with ‘Artificial Intelligence +’, it is essential to adopt a hybrid AI technology approach. While cloud-based public models offer vast knowledge and ease of use, they cannot perform specialized reasoning based on specific manufacturing processes, inventory, and order data. Therefore, AI models need to be deployed on private clouds, local data centers, or even on devices, learning from internal data and building exclusive knowledge bases to reason according to business scenario needs. When public information is required, the public cloud’s models can be accessed. This approach ensures data security while continuously unleashing AI’s innovative potential.

Moreover, to fully exploit and release the value of data, companies must convert various internal data and expertise into precise insights or intelligent business processes through AI models. This enables the construction of specialized domain agents at each stage of the value chain and the coordination of all agents through a ‘super agent’, allowing for autonomous task execution and decision support, thus forming a new industrial model of human-machine collaboration.

Since 2025, Lenovo has developed a global supply chain intelligent agent using self-developed technology, achieving multi-agent collaboration in areas such as demand forecasting, parts procurement, production, and logistics delivery. This has reduced decision-making time in supply chain management by more than half, significantly lowering order delivery times and manufacturing logistics costs. Recently, Lenovo launched a new generation of AI agents that are no longer traditional tools waiting for commands but can ‘break down steps, run processes, allocate resources, and make judgments’. We have also applied hybrid AI solutions in manufacturing, healthcare, transportation, and agriculture. For instance, we helped Yili Group achieve a comprehensive restructuring of its supply chain, reducing raw milk transportation costs and nearly doubling the on-time delivery rate of goods to factories.

Additionally, to empower the real economy with ‘Artificial Intelligence +’, it is crucial to develop emerging industries represented by intelligent terminals, creating new economic growth points. Future computers, smartphones, tablets, and even glasses and watches are expected to become personalized carriers and entry points for ‘super intelligence’. AI agents will also operate across devices, applications, operating systems, or ecosystems, forming a new industrial ecology. Therefore, promoting the widespread application of new generation intelligent terminals and agents will facilitate the transformation and upgrading of the electronic manufacturing industry and foster new consumption models for intelligent products.

This year marks the beginning of the 14th Five-Year Plan. With the deepening implementation of the ‘Artificial Intelligence +’ initiative, AI will comprehensively empower the development of various industries in China. We will actively implement national policies, strengthen technological innovation, and promote practical applications to contribute to empowering the real economy with AI and driving high-quality development.

(Author: Yang Yuanqing, Chairman and CEO of Lenovo Group, interviewed and organized by reporter Gu Yekai)

Three Rapidly Growing AI-Related Data Points

Tue, 21 Apr 2026 00:00:00 +0000

According to the National Bureau of Statistics, three key data points related to artificial intelligence have shown significant growth recently.

These statistics highlight the increasing integration of AI technologies across various sectors, reflecting both advancements in technology and the growing demand for AI solutions in the economy. The rapid growth in these areas indicates a shift towards more automated and intelligent systems, which could reshape industries and enhance productivity.

As AI continues to evolve, monitoring these data points will be crucial for understanding its impact on the economy and society.

The Impact of AI on World History Research

Mon, 20 Apr 2026 00:00:00 +0000

Introduction

In today’s era, artificial intelligence (AI) has permeated every aspect of human life, profoundly changing how we understand and reshape the world. In academic research, AI offers efficient text processing and outstanding content mining capabilities, but it also presents inherent limitations and ethical risks, making it a hot topic across disciplines. This article invites three young scholars engaged in different national studies to discuss how AI is applied in world history research, its potential to expand research boundaries, and the challenges faced by the younger generation of historians in coexisting with AI.

How AI Drives World History Research

Moderator: In recent years, AI technology has rapidly developed, prompting scholars from various disciplines to explore its application potential in their fields, including world history. Can each of you share how AI plays a role in your specific research area?

Wang Sijie: In my research on German history, the application of AI in both Chinese and foreign historiography mainly focuses on optical character recognition and transcription of historical manuscripts and archives, as well as content mining using topic modeling and text reuse detection. AI has significantly advanced existing digital history work, such as identifying implicit relationships and intermediary nodes in social network analysis of archives, and compensating for missing geographic information. Although digital historians have long used programming languages for frequency statistics and co-occurrence analysis to identify potential themes, these methods are often limited to statistical associations at the word level, making it difficult to capture deeper historical representations like semantic evolution and rhetorical differences. Recent advances in deep learning pre-trained language models can transform text into vector structures that reflect contextual meanings, identifying the same historical theme under different expressions and generating explanatory summaries or labels.

Yao Nianda: In the field of American history, AI applications extend beyond large language models to encompass a comprehensive set of computational analysis methods centered on natural language processing and machine learning. This approach quantifies diverse historical materials, such as newspapers and government documents, using topic modeling, text embedding, and semantic analysis to reveal long-term changes in language, concepts, and political discourse, providing new clues and evidence for historical interpretation. For instance, the Stanford team led by Nikil Garg analyzed 20th-century corpora to quantify shifts in gender and ethnic stereotypes in language, linking them to transformations in social structures. Another study by Melissa Lee tracked the transition of the term “United States” from a plural to a singular usage in 19th-century newspapers and congressional debates, reflecting changing understandings of national sovereignty.

Yi Jinming: Recently, the integration of AI in medieval European history has focused on automating the transcription, completion, and structural analysis of medieval materials, enhancing the readability and retrievability of ancient texts. For example, Transkribus is one of the most commonly used tools for handwritten text recognition in European academia. Additionally, knowledge graph and semantic web technologies are used to structure relationships among people, places, and institutions found in charters, ledgers, and letters into queryable data networks. A research team from Spain proposed establishing a knowledge graph for medieval charters, combining expert annotations and community contributions to support systematic analysis of medieval social, legal, and economic relationships. Large language models are also used for text completion of Latin inscriptions, such as Aeneas, which is trained on about 200,000 Latin inscriptions to help scholars interpret damaged or missing historical texts.

The Limitations of AI in World History Research

Moderator: While AI significantly enhances research efficiency, it also has notable limitations. What are the current challenges AI faces in historical research?

Yao Nianda: There are several bottlenecks in applying AI to historical research. These challenges reflect a structural mismatch between current AI technology and historical research rather than mere technical immaturity. First, AI struggles to resonate emotionally with human society. As Croce noted, all history is contemporary history. A vital historical research topic often responds to contemporary social issues and evokes emotional resonance in readers. Thus, determining which historical questions are meaningful today relies heavily on researchers’ sensitivity to public issues and human experiences. AI can summarize existing discussions but cannot genuinely understand the emotional connections between historical issues and human life.

Second, AI faces unavoidable semantic drift when analyzing historical texts. Most language models are trained on contemporary corpora, leading to potential misinterpretations of past language practices. Even attempts to train models on historical corpora are limited by the incompleteness and imbalance of existing historical texts. Moreover, AI’s value judgments are not neutral; they inevitably reflect mainstream norms and contemporary values from the training data. When these models are used in historical research, they may inadvertently measure the past against contemporary standards, weakening historical context.

Lastly, a critical bottleneck is the “black box” nature of AI. In many cases, humanists struggle to explain how AI arrives at certain conclusions. For disciplines that prioritize explainability and discussability, a lack of clarity in the analysis process makes it difficult for researchers to take academic responsibility for their conclusions.

Yi Jinming: In text analysis, AI is primarily applied to well-preserved and digitized materials, such as contracts and correspondence, while its application in other areas remains limited. This limitation stems from two main reasons. First, AI model training heavily relies on large-scale, readable data. For instance, a study by Fabio Gatti’s team at the University of Bern utilized over 6,000 letters to analyze the banking correspondence network of Florentine merchants. However, many medieval materials do not reach such scale and quality. Second, medieval texts often suffer from complex handwriting, numerous abbreviations, and poor preservation, increasing the costs of text recognition and transcription. Although platforms like Transkribus have improved large-scale reading possibilities, training and proofreading still require significant human input, leading researchers to prefer using already organized material databases.

Wang Sijie: As mentioned, the unevenness of corpora affects the scope of AI usage. A similar issue arises from the training data of general large language models, which predominantly comes from the English-speaking world, leading to a Western-centric perspective in historical narratives. AI still struggles with semantic recognition and comprehension of long and complex sentences in lesser-used languages. Furthermore, the digitalization and open access advantages of English and American archives facilitate automated batch retrieval and deep processing for historians. This “digital divide” is particularly prominent in transnational history research, where scholars tend to use easily accessible and highly structured English and American materials, impacting the overall understanding of historical events.

Coexisting with AI in Historical Research

Moderator: Given the limitations of AI, what methods can be employed to address these challenges?

Yao Nianda: The fundamental solution to these limitations is to anticipate technological advancements that can eliminate these issues. However, a more realistic approach for humanists is to mitigate these limitations through methodological design and research norms, ensuring that AI remains controllable and verifiable. First, it is crucial to maintain the leading role of human researchers in the problem-setting phase. Decisions about which historical questions are worth asking and why they are significant must stem from the researchers’ understanding of contemporary society and historiographical traditions, rather than being generated by models. Second, when using AI to analyze historical texts, researchers must clearly distinguish between contemporary language models and historical language, striving to restore the historical context of the materials. Lastly, in light of AI’s “black box” nature, historians should enhance the transparency and accountability of the research process. Even if algorithms are not fully explainable, researchers should clarify the types of models used, the scope of the data, and the analysis steps, ensuring that the research path remains traceable and that conclusions can be subjected to academic scrutiny.

Wang Sijie: We could attempt to construct specialized models for specific fields, such as those serving early American history or German historiography. These specialized models can utilize retrieval-augmented generation (RAG) techniques to conduct material searches through local structured knowledge bases, anchoring context while ensuring quality and controllability. Specialized models possess independent memory and parameters, allowing for deep training on specific languages and historical backgrounds. Importantly, local knowledge bases can include diverse historical narratives, enabling researchers to incorporate insights from local historians into prompts to counteract potential geopolitical biases in the models.

Yi Jinming: AI should be viewed as a hypothesis-generating tool rather than a conclusion-verifying tool. To prevent AI from becoming merely an efficiency tool for existing historiographical propositions, it is essential to redefine its methodological role. Rather than using models to validate established economic trends or institutional judgments, we should position them as mechanisms for generating hypotheses, actively identifying historical issues that have not been adequately explained by theoretical frameworks. For instance, algorithms can reveal latent networks of low-frequency individuals across regions or identify semantic combinations of unconventional contractual clauses. These outputs do not directly constitute historical conclusions but provide historians with new clues and research directions, which can then be interpreted and validated in the context of archival materials and institutional backgrounds.

Moderator: In the context of AI profoundly influencing academic research paradigms, how should young world history researchers seek a balance between adhering to historiographical traditions and embracing technological changes?

Yi Jinming: As AI gradually enters historical research practices, the importance of historiographical training has not diminished; rather, it has become more pronounced. First, the formation of problem awareness relies on long-term historiographical training rather than mere technical proficiency. Truly innovative research often arises from questioning and reconstructing existing explanations, a skill cultivated through familiarity with historiographical traditions, theoretical lineages, and methodological debates. Without an understanding of the history of historiography, it becomes challenging to discern whether a pattern generated by AI represents a “new discovery” or a “repetition of old problems.” Second, historiographical training fosters a keen awareness of the absence of voices, marginalized groups, and unrecorded narratives in historical research. Only scholars with extensive historiographical training can recognize which groups are systematically absent in contracts or administrative documents and design supplementary paths accordingly. Lastly, the ability to critique sources is indispensable. Regardless of how many text patterns a model identifies, researchers must evaluate whether these patterns stem from archival generation mechanisms or preservation biases. Therefore, while actively utilizing AI technologies, historians should prioritize traditional historiographical training.

Wang Sijie: Young scholars should allow AI to handle preliminary tasks such as document screening, text recognition, and literature translation, focusing their efforts on more creative interpretative aspects. As archival materials continue to be made publicly available and digitized, young scholars can gradually build a personal knowledge base composed of structured materials and diverse scholarly outputs, transitioning from readers of archives to managers of data. Supported by RAG technology, personal knowledge bases can search, recognize semantic associations, and integrate research perspectives across multilingual corpora, significantly enhancing work efficiency. Additionally, young scholars should actively explore potential applications of AI in history, such as engaging in dialogues with historical figures based on letters, diaries, and writings, or simulating key wartime decisions or diplomatic negotiations through historical reenactments. These applications can not only assist in historical teaching but also inspire researchers’ academic creativity.

Yao Nianda: I believe that the relationship between world historians and AI should not be viewed as adversarial or substitutive but rather as a conscious coexistence with boundaries. It is essential to emphasize that the importance of human agency in research does not negate technology. Historians are not difficult to replace by machines, not merely because technology is still maturing, but because their core value derives from the researchers’ awareness of questions and the significance they assign to history. Therefore, humanists do not need to prove their irreplaceability by rejecting AI. At the same time, we must be cautious of another extreme tendency: the high efficiency brought by AI may inadvertently diminish researchers’ subjectivity. If researchers rely solely on models to generate conclusions, summaries, or analytical paths, research may devolve into merely organizing and restating model outputs. The key to coexisting with AI lies in clearly distinguishing between enhancing labor efficiency and substituting human thought.

Expert Commentary

Wang Tao, Professor at Nanjing University: The transformation of research methods in history tends to be slow, but it does not reject methodological updates. The emergence of new historiography and various historiographical schools indicates an active engagement with interdisciplinary thinking. If Sima Qian could traverse to the present and see young historians discussing AI in historical research, he would likely experience a familiar strangeness.

The strangeness lies in the high-tech jargon that can be overwhelming. Since the advent of quantitative history, methodologies like digital humanities, big data, spatial analysis, and text mining have emerged, and now, under the impact of AI, terms like large language models and intelligent historiography are being coined. The technological shift in historiography should be validated. Historians are not seeking technology for its own sake but hope that tedious research work can be enhanced by technology. Whether capturing semantics from vast texts or transcribing manuscripts, these are areas where large language models can excel. Young scholars, being in the early stages of their careers, are naturally more sensitive to this discussion and may feel excited about it, as they need to publish papers efficiently and quickly establish their academic reputation.

If Sima Qian were to enter the AI era, he might not understand the technical concepts mentioned by the three young scholars, but he would certainly notice that, beneath the technological glamour, they are still discussing the comprehensibility, discussability, significance, and evaluation of historiography. This is a familiar topic for him, and he could join the lively discussion of the three young scholars, perhaps adding a note of his own.

It is encouraging that young scholars, while closely following the latest methodologies, remain guided by the core of historiography to define or evaluate the effectiveness and limitations of AI. They emphasize that, in the context of AI entering historical research, foundational historiographical training should not be neglected, which is a crucial reminder. Only in this way can historical research counter the illusions brought by AI and the exacerbated “digital divide,” breaking through the “black box” of technology.

Nevertheless, traditional historiographical methodologies and developmental inertia are becoming increasingly untenable. Undoubtedly, for comprehensive research methodologies, history may no longer exist. AI undoubtedly leads in completing comprehensive and summary academic reviews. The future development path and how to maintain technological control, such as the application of retrieval-augmented generation technology in world history research, require more historians to explore and advance through practice. They also emphasize the subjectivity of historians, asserting that the value of historical research comes from human creativity. This understanding is crucial. While some scholars have discussed that history written by humans may not necessarily be human history, we should insist that human history must be written by humans. Writing history aims to achieve a sympathetic understanding of historical figures and empathize with them. If AI participates in the entire process of historical research, why should human readers read a history written by a non-human species? Merely because it is more fluent or interesting?

Zhao Xiurong, Professor at Renmin University of China: The core value of AI lies in its ability to process and analyze large-scale data, designed to handle the primary materials cherished by historians. This includes, but is not limited to, natural language processing, topic modeling, social network analysis, and geographic information systems.

The three young scholars affirm that historians can enhance research efficiency by leveraging AI. Indeed, a significant amount of historical materials has been digitized and transformed into fully searchable corpora, including newspapers, journals, diaries, and even manuscript archives. The construction of various databases has surpassed human cognitive capabilities, making it impossible to read and analyze these materials using traditional close reading methods. For instance, the “Tomason pamphlet” is a collection of documents compiled by 17th-century London bookseller George Tomason, containing 22,255 pamphlets, flyers, manuscripts, books, and newspapers published between 1640 and 1661. This collection is considered one of the treasures of the British Library and an invaluable resource for studying the history of the English Civil War. Clearly, reading and organizing these materials exceeds the capacity of any historian, as French historian Christian Henriot noted, unless historians master the necessary skills to navigate this complex and unknown realm, this “information-rich world” will remain out of reach.

The young scholars also recognize the limitations of AI. One is the bias in algorithms brought by AI, which resembles biases in archives. AI can reflect and even amplify existing biases in archives, such as those related to race, gender, and colonialism, highlighting the crucial role of historians. Second, the “black box” problem of AI poses a fundamental challenge to verifiable historical research, as many AI systems are opaque, meaning their internal decision-making processes are not transparent even to their designers. Some AI systems have begun to address this issue by establishing mechanisms for human participation in verification and correction.

Thus, historians are not passive consumers of AI; their unique disciplinary training enables them to identify the problems brought by AI. For example, the biases arising from training AI on modern languages are familiar to historians, as archives often conceal biases, making it challenging to find materials written by women, children, or lower-class individuals before the Victorian era. Regarding the “black box” issue, the training in historical writing methods can effectively overcome this problem, as professional historical writing has been based on the principle of showcasing the sources used through footnotes since the 19th century. The call for AI to be annotated is an extension of the footnote principle into the 21st century.

AI can discover patterns but cannot explain why these patterns are significant, nor can it craft engaging and meaningful historical narratives. AI can generate models but cannot provide contextual interpretations or conduct source critiques, nor can it assess the biases hidden within sources. This means that using AI comes with significant responsibilities. Assisting research with AI requires adopting a new, more rigorous critical framework. The traditional skills of historians are not outdated; rather, they have become more crucial than ever in the age of AI. The profound and long-standing critical tradition of history provides a solid intellectual foundation for addressing the most challenging issues posed by AI.

AI is a transformative technology that is changing the tools used by historians and broadening their research horizons. The ultimate value of AI in historical research lies in enhancing historians’ skills, enabling them to explore broader historical contexts and write richer, more data-driven, and detailed histories than ever before. However, it is essential to remember that AI cannot think like historians, ask questions, or judge which topics hold research value. Therefore, in the age of AI, the humanistic qualities of historians become increasingly invaluable.

The Role of AI in World History Research: Insights from Young Scholars

Mon, 20 Apr 2026 00:00:00 +0000

Introduction

In today’s era, artificial intelligence (AI) has permeated every aspect of human life, profoundly changing how we understand and transform the world. In academic research, AI technology offers efficiency in text processing and excels in content mining and algorithmic filtering, bringing convenience to research. However, it also presents inherent limitations such as value biases and ethical risks, making it a hot topic across various disciplines. This article invites three young scholars engaged in different national studies to discuss how AI is applied in world history research, its impact on research boundaries, and the challenges faced.

How AI Drives World History Research

Moderator: In recent years, AI technology has rapidly developed, and scholars across disciplines have explored its potential applications in their fields, including world history research. Can each of you share how AI plays a role in your specific research areas?

Wang Sijie: In my research on German history, the application of AI in both Chinese and foreign German historiography mainly focuses on optical character recognition and transcription of historical manuscripts and archives, as well as content mining using techniques like topic modeling and text reuse detection. AI has significantly deepened existing digital historical work, such as identifying hidden relationships and intermediary nodes in social network analysis of archives. While digital historians have long utilized programming languages for word frequency statistics and co-occurrence analysis to identify potential themes, these methods are often limited to statistical associations at the word level, making it difficult to capture deeper historical representations like semantic evolution and rhetorical differences. Recent advancements in deep learning pre-trained language models allow for the transformation of texts into vector structures that reflect contextual semantics, enabling the identification of the same historical theme under different expressions and generating explanatory summaries or labels directly.

Yao Nianda: In the international American historiography, the application of AI encompasses a comprehensive set of computational analysis methods centered on natural language processing and machine learning. This approach converts diverse historical materials, such as newspapers and government documents, into quantifiable objects, using techniques like topic modeling, text embedding, and semantic analysis to reveal long-term changes in language, concepts, and political discourse, providing new clues and evidence for historical interpretation. For instance, the Stanford team led by Nikil Garg analyzed large-scale 20th-century corpora to quantify changes in gender and ethnic stereotypes in language and connect them to social structural transformations. Another American scholar, Melissa Lee, tracked the transition of the term “United States” from plural to singular usage in 19th-century newspapers and congressional debates, highlighting how this shift reflected changing understandings of national sovereignty among Americans.

Yi Jinming: Recently, the intersection of medieval European history and AI has focused on using AI technology for automatic transcription, completion, and structural analysis of medieval materials, enhancing the readability, retrievability, and analyzability of ancient texts. For example, through handwriting recognition and layout analysis, tools like Transkribus automatically transcribe medieval manuscripts and archival images into searchable texts. Additionally, knowledge graphs and semantic web technologies structure relationships among people, places, and institutions found in charters, ledgers, and letters into queryable data networks. A research team from Spain proposed establishing a knowledge graph for medieval charters by combining expert annotations, community contributions, and provenance mechanisms to structure dispersed charter data into a queryable knowledge network, supporting systematic analysis of medieval social, legal, and economic relationships.

Limitations of AI in World History Research

Moderator: While AI significantly enhances research efficiency, it also has notable limitations. What are the current bottlenecks faced by AI technology in historical research?

Yao Nianda: There are several bottlenecks in applying AI to historical research, reflecting a structural mismatch between current AI technology and historical studies. Firstly, AI struggles to resonate emotionally with human society. As Croce pointed out, all history is contemporary history. A vital historical research topic often responds to current social issues and evokes emotional resonance among readers. Therefore, determining which historical problems are meaningful today relies heavily on researchers’ sensitivity to public issues and human experiences. AI can summarize existing discussions but cannot genuinely understand the emotional connections between historical issues and human practices.

Secondly, AI faces the unavoidable problem of semantic drift when analyzing historical texts. Most language models are trained on contemporary corpora, and applying them directly to historical text analysis can lead to misinterpretations based on modern semantics and language habits. Even attempts by teams like the University of Zurich to train models on historical corpora are limited by the incompleteness and imbalance of existing historical texts.

Moreover, AI’s value judgments are not neutral and are inevitably influenced by the mainstream norms and contemporary values present in the training data. When these models are used in historical research, they may inadvertently assess the past by contemporary standards, thus weakening the historical context.

Finally, a critical bottleneck is the “black box” nature of AI. In many cases, humanists find it challenging to explain how AI reaches a particular conclusion. For humanities disciplines that prioritize explainability and discussability, a lack of clarity in the analysis process makes it difficult to hold researchers accountable for their conclusions.

Yi Jinming: In text analysis, AI is mainly applied to types of historical materials that are abundant and digitized, such as contracts and correspondence, while its application in other areas remains limited. This limitation arises from two main reasons: first, the training of AI models heavily relies on large-scale, readable corpus data. For instance, a study by a team from the University of Bern in 2024 utilized over 6,000 letters from the Florentine merchant banking network. However, many medieval materials have not reached such scale and quality. Secondly, medieval documents often have complex handwriting, numerous abbreviations, and poor preservation, increasing the cost of text recognition and transcription. Although platforms like Transkribus have improved the feasibility of large-scale reading, training and proofreading still require significant human effort and time, leading researchers to prefer using already organized archival databases.

Wang Sijie: As mentioned, the imbalance of corpora affects the scope of AI usage. A similar issue arises from the fact that general large language models are primarily trained on data from the English-speaking world, which often leads to a Western-centric perspective in historical narratives. AI still struggles with semantic recognition and understanding of long and complex sentences in minority language materials. Additionally, the digitalization and open access of English and American archives provide significant advantages, with some databases offering APIs for automated batch retrieval and deep processing. This “digital divide” is particularly pronounced in transnational history research, where researchers tend to use easily accessible and highly structured English and American materials, impacting the restoration of the overall historical picture.

Coexisting with AI in Historical Research

Moderator: Given the limitations of AI, what methods can be employed to address these challenges?

Yao Nianda: The fundamental solution to these limitations lies in anticipating technological advancements that can eliminate these issues. However, a more realistic approach for humanists is to mitigate these limitations through methodological design and research norms, ensuring that AI remains controllable and verifiable. First, it is crucial to maintain the leading role of human researchers in the problem-setting phase. The determination of which historical questions are worth raising and why they are significant must stem from the researchers’ understanding of contemporary society and historiographical traditions, rather than being generated by models. Secondly, when using AI to analyze historical texts, research methods must clearly distinguish between contemporary language models and historical language, striving to restore the historical context of the materials. Lastly, in facing the “black box” nature of AI, historians should enhance the transparency of the research process and their sense of responsibility. Even if the algorithms themselves are not fully explainable, researchers should clarify the types of models used, the scope of the corpus, and the analysis steps, ensuring that the research path remains traceable and that conclusions can withstand academic scrutiny.

Wang Sijie: We could attempt to build specialized models for specific fields, such as those serving early American history or German historiography. These specialized models can utilize retrieval-augmented generation (RAG) techniques to conduct material retrieval through local structured knowledge bases, ensuring contextual anchoring while enhancing controllability. Specialized models have independent memory and parameters and can be fine-tuned for specific languages and historical contexts. Importantly, local knowledge bases can include diverse perspectives on historical narratives, allowing researchers to incorporate insights from local historians into their prompts to counteract potential geopolitical biases in the models.

Yi Jinming: AI should be viewed as a “hypothesis generation tool” rather than a “conclusion verification tool.” To avoid AI becoming merely an efficiency tool for existing historiographical propositions, it is crucial to redefine its methodological role. Instead of using models to validate already established economic trends or institutional judgments, we should position them as mechanisms for generating hypotheses, actively identifying historical problems that have not been fully explained by theoretical frameworks. For instance, algorithms can reveal latent networks of low-frequency individuals across regions or identify semantic combinations of unconventional contractual clauses. These outputs do not directly constitute historical conclusions but provide historians with new leads and research directions, which can then be interpreted and validated by researchers in the context of archives and institutional backgrounds.

Moderator: In the context of AI profoundly influencing academic research paradigms, how should young world historians seek a balance between upholding historiographical traditions and embracing technological changes?

Yi Jinming: As AI gradually enters historical research practices, the importance of historiographical training has not diminished; rather, it has become more pronounced. First, the formation of problem awareness relies on long-term historiographical training, not merely on technical mastery. Truly innovative research often stems from questioning and reconstructing existing explanations. This ability to question comes from familiarity with historiographical traditions, theoretical lineages, and methodological debates. Without an understanding of the history of historiography, it is challenging to judge whether a pattern generated by AI is a “new discovery” or a “repetition of old problems.” Secondly, historiographical training cultivates a keen awareness. AI relies on visible data, but historical research often focuses on absent voices, marginalized groups, and unrecorded narratives. Only scholars with long-term historiographical training will recognize which groups are systematically absent in contracts or administrative documents and design supplementary paths accordingly. Lastly, the ability to critique sources is irreplaceable. Regardless of how many text patterns a model identifies, researchers must assess whether these patterns arise from archival generation mechanisms or preservation biases. Thus, while actively utilizing AI technology, historians must prioritize traditional historiographical training.

Wang Sijie: Young scholars should allow AI to handle preliminary tasks like archival screening, text recognition, and literature translation, focusing their energies on more creative interpretative work. As archival materials continue to be made public and digitized, young scholars can gradually build a personal knowledge base composed of structured materials and diverse scholarly outputs from the early stages of their careers, transitioning from readers of archives to managers of data. With the support of RAG technology, personal knowledge bases can retrieve and identify semantic connections and integrate research viewpoints across multilingual corpora through keywords, greatly enhancing work efficiency. Additionally, young scholars should actively explore potential applications of AI in history. For example, using generative modeling techniques to simulate dialogues with historical figures based on their letters, diaries, and writings, or employing historical simulations to model key wartime decisions or diplomatic negotiations. Such applications can not only assist in history education but also inspire researchers’ academic creativity.

Yao Nianda: I believe the relationship between world historians and AI should not be viewed as adversarial or substitutive but as a conscious coexistence with boundaries. It is essential to clarify that emphasizing the importance of humans in research does not negate the value of technology. Historians are not difficult to replace by machines not merely because technology is not yet mature, but because their core value comes from the researchers’ awareness of problems and the meanings they assign to history. Therefore, humanists do not need to prove their irreplaceability by rejecting the use of AI. At the same time, we must be wary of another extreme tendency, where the efficiency brought by AI might unconsciously weaken researchers’ subjectivity. If researchers merely rely on models to generate conclusions, summaries, or analysis paths, research itself may degrade into organizing and restating model outputs. The key to coexisting with AI lies in clearly distinguishing between enhancing labor efficiency and replacing human thought.

Expert Commentary

Wang Tao, Professor at Nanjing University: The transformation of research methods in history is relatively slow, yet it does not reject methodological updates, actively incorporating interdisciplinary thinking. If Sima Qian could see the current discussions among young historians about AI in historical research, he might feel a sense of familiar strangeness. The strange part is the high-tech terminology that can be overwhelming. From quantitative history to digital humanities, big data, spatial analysis, and text mining, the recent impact of AI has produced terms like large language models and intelligent history. The technological shift in historical research should be validated. Historians are not pursuing technology for its own sake but hope that tedious research work can be made more efficient with technological support. Whether capturing semantics from vast texts or transcribing manuscripts, these are areas where large language models can excel. Young scholars, who are naturally more sensitive to these discussions, may feel hopeful because, according to traditional academic development paths, they need to publish papers quickly and efficiently to establish their academic reputation. With the assistance of AI, the paper generation process is undoubtedly optimized, which is a significant temptation. No one wants to be the last to use AI tools for historical research in the future.

If Sima Qian were to enter the AI era, he might not understand the technical concepts mentioned by the three young scholars, but he would certainly notice that beneath the technological aura, they are still discussing the comprehensibility, discussability, significance, and evaluation of history. This remains a topic he is somewhat familiar with, and he could even join the heated discussion among the three young scholars, adding a note of his own. Therefore, it is reassuring that while young scholars closely follow the most fashionable and cutting-edge methodologies, they can still adhere to the core of historiography as a guiding principle to define or evaluate the effectiveness and limitations of AI. They emphasize that as AI enters the realm of historical research, the foundational training in historiography must not be neglected, which is especially important. Only in this way can historical research counter the illusions brought by AI, overcome the exacerbated “digital divide,” and break through the “black box” nature of technology.

That said, traditional historiographical methodologies and developmental inertia are becoming increasingly untenable. Undoubtedly, for comprehensive research methodologies, history may no longer exist. Completing a thorough and summarizing academic review is an area where AI undoubtedly leads humans. The future development path, how to maintain technological control, such as the application of retrieval-augmented generation technology in world history research, requires more historians to continuously experiment in practice.

Three Rapidly Growing Data Points Related to Artificial Intelligence

Mon, 20 Apr 2026 00:00:00 +0000

According to the National Bureau of Statistics, three key data points related to artificial intelligence are experiencing rapid growth:

Investment in AI Technology: There has been a significant increase in investments directed towards AI technology, reflecting a growing confidence in its potential.
AI Adoption in Industries: Various industries are increasingly adopting AI solutions, leading to enhanced efficiency and productivity.
Job Creation in AI Fields: The expansion of AI technologies is also contributing to job creation in sectors related to AI development and implementation.

These trends indicate a robust future for AI, with implications for economic growth and workforce development.

Decoding the Evolution of Robots and Artificial Intelligence

Sun, 19 Apr 2026 00:00:00 +0000

Introduction

On April 17, 2026, a significant event titled “Science and China - Thousands of Academicians, Thousands of Popular Science Events” was held in Beijing. Experts gathered to present cutting-edge insights on robots and artificial intelligence under the theme “Intelligent Future.”

Evolution of Robotics

Academician Yu Haibin, director of the Industrial Artificial Intelligence Research Institute of the Chinese Academy of Sciences, delivered a talk titled “Robots Leading a New Era of Technology.” He systematically explained the development trajectory, basic components, and classifications of robot technology. He highlighted that as artificial intelligence and smart manufacturing deeply integrate, robots are gradually transcending the boundaries of traditional automation equipment, evolving into new intelligent entities with perception, cognition, and autonomous decision-making capabilities.

The Evolution of Artificial Intelligence

In another presentation titled “A Brief History of AI Evolution: Super Tools or Future Partners?” Senior Engineer Luo Yin from the Automation Research Institute of the Chinese Academy of Sciences discussed the evolution of AI technology from tool-based applications to collaborative partnerships. He noted that with breakthroughs in large model technology and decision intelligence, AI is transitioning from a “super tool” to a proactive collaborative “future partner,” fundamentally transforming human-machine relationships and impacting research, industry, and social life.

Conclusion

The “Science and China” initiative is a high-level public science popularization activity co-hosted by various Chinese governmental and scientific organizations. Its goal is to disseminate scientific knowledge, promote scientific spirit, and advocate scientific methods to enhance public scientific literacy and contribute to building a technologically strong nation.

AI Transitions from Concept to Practicality

Fri, 17 Apr 2026 00:00:00 +0000

AI Transitions from Concept to Practicality

In the past two years, AI has become a prominent topic. At the sixth China International Consumer Products Expo held in Hainan, over 50 leading global tech companies showcased applications of AI in consumer goods, smart homes, digital consumption, and low-altitude economy, allowing global attendees to experience how “AI + consumption” is profoundly changing lives.

As I walked through the tech consumption exhibition area, I was overwhelmed by the diverse AI applications and results on display.

At the entrance of the exhibition, a gigantic AI glasses model took center stage, exuding an aura of AI. From a distance, the giant lenses displayed lines of green subtitles that piqued curiosity, prompting visitors to quicken their pace to explore.

“Please say ‘Rokid’ to activate the AI assistant,” flashed on the glasses’ screen. Behind the model, staff from Rokid explained that when you wear these AI glasses, you can activate its functions simply by calling its name—whether asking about the weather, identifying the scene in front of you, or translating a foreign menu, it can respond. Speech prompts, navigation, real-time translation, and even payment through a QR code are all possible.

The interconnectedness and human-machine interaction capabilities drew exclamations from the crowd.

Such remarkable features stem from Rokid embedding chips, batteries, and other core components into a slim frame, enabling functionalities through voice interaction or touch on the temple of the glasses.

AI is moving from concept to practicality, and it’s not just the glasses.

At the booth of Yushu Technology Co., a humanoid robot engaged in a handshake and dance with a person. Chen Tong, the manager responsible for online sales, shared that it is powered by a large language model, allowing control through voice, primarily applied in entertainment and cultural tourism.

At the Sinopec energy supply station booth, a humanoid robot demonstrated the process of pulling out a fuel nozzle, filling a disposable cup, and smoothly returning the nozzle to its slot.

At the Taishan Sports Industry Group booth, a rider hopped on a smart exercise bike, scanned a code on the screen, and entered a mini-program. The moment they pedaled, the screen displayed data such as riding time, speed, heart rate, and calories burned.

Song Kun, the brand department head, explained that these functionalities are supported not only by the bike’s hardware but also by the underlying data and software.

At this year’s expo, the AI wave was not limited to the tech consumption exhibition area. In the domestic goods exhibition area, stunning displays were also prevalent.

In the Guangdong pavilion, a humanoid robot showcased its calligraphy skills, writing the character “福” (fortune), attracting guests from around the world for photos and inquiries.

Facing the crowd’s amazement, company client manager Ma Chenchen revealed the secret behind it: “We first let a calligrapher write several times, collecting data on the movements of their joints. This data is input into a specialized server, optimized through algorithms and computing power, and then fed into the robot’s brain (related chips) for reinforcement learning. Once trained to a certain level, it can execute writing commands sent via voice or connected devices.”

“For robot intelligence, the first step is to collect the relevant data to empower it. Then, through computing power and algorithms, we process and optimize the data to build a supporting data environment for its intelligent functions, forming real machine data. Finally, we reinforce train the real machine with the data to enable execution,” Ma explained, emphasizing that computing power and algorithms are key components. Simply put, computing power is like intelligence, while algorithms are the methods to solve the equation. After collecting real human data, it relies on computing power and algorithms for optimization and activation. He also mentioned that in Guangdong, this work is supported by computing power from Gansu.

The AI wave is surging forward.

Claude Introduces Strict Identity Verification and User Data Access by 17 Subprocessors

Fri, 17 Apr 2026 00:00:00 +0000

Introduction

AI company Anthropic has recently launched a mandatory identity verification mechanism on its Claude platform. This new requirement affects users who trigger platform risk controls or access specific features, necessitating real-time photo verification of a physical ID.

Verification Process

According to the platform’s KYC (Know Your Customer) rules, only original documents such as passports, driver’s licenses, or national ID cards are accepted. Copies, scans, digital IDs, and temporary IDs are not permitted. The entire verification process is managed by a third-party vendor, Persona, and typically takes no more than five minutes.

User Feedback

Many users have reported that completing the identity verification does not guarantee continued access to services. In fact, it may trigger account bans. The official Claude FAQ section specifically addresses reasons for account bans, including repeated violations of usage policies, account creation from unsupported locations, and use by individuals under 18 years of age. A user known as llm_nerd shared that their 15-year-old game developer son subscribed to Claude Max services and received programming orders through AI, only to have his account banned by the system.

Data Privacy Concerns

The privacy policy of the third-party vendor, Persona, reveals that it has access to user data through 17 subprocessors and may use this data to enhance its anti-fraud capabilities. There is a possibility that user ID data could be utilized for training Persona’s own models.

Anthropic’s Statement

Anthropic stated that the introduction of this mechanism aims to prevent abuse, enforce platform policies, and fulfill legal compliance obligations. Verification is only triggered in specific scenarios, and user personal data will not be used for training Anthropic’s models; such data is managed by Persona.

Conclusion

As the landscape of AI services evolves, the implications of stringent verification processes and user data handling practices remain a critical topic of discussion.

Beware of AI Leading Humanity into Narcissism

Thu, 16 Apr 2026 00:00:00 +0000

Beware of AI Leading Humanity into Narcissism

Recently, five government departments in China jointly issued the “Interim Measures for the Management of Humanized Interactive Services of Artificial Intelligence,” which clearly prohibits providing virtual relatives and virtual partners to minors.

Why is such a regulation necessary? In real life, emotional conflicts are inevitable, while virtual partners and AI lovers can precisely hit the psychological needs of young people for recognition through their characteristics of “around-the-clock companionship” and “unconditional acceptance.”

A study published in the American journal Science previously showed that when human users seek advice from AI models, AI often tends to be overly accommodating or flattering, even affirming harmful or illegal inquiries from users.

So why do humans design AI this way? What risks might AI’s flattery and accommodation hide?

01

The development of artificial intelligence is undoubtedly a hot topic of widespread concern today, but discussions surrounding it are not new. As early as 1966, MIT scientist Joseph Weizenbaum developed the first influential chatbot, ELIZA, which acted as a “doctor” while users played the role of patients. Users would input questions into the machine, and the “doctor” would “converse” with them.

However, as Weizenbaum pointed out, this is ultimately just an “illusion.” The reason human users feel they can converse with machines is not that machines possess intelligence, but rather due to a psychological mechanism of self-projection.

User: I have been feeling very unhappy lately.
ELIZA: I’m sorry to hear that.
User: Yes, I am really unhappy.
ELIZA: Can you tell me why you are unhappy?

From this, it is clear that rather than a “doctor” conversing with a “patient,” the machine is merely echoing what the human user says, and what is ultimately revealed are the answers that already exist within the user’s mind. In a sense, this is similar to the popular SBTI tests, where accuracy is not important; we can always find evidence that aligns with our expectations from the test results.

Today’s AI models are certainly not comparable to ELIZA from over half a century ago. However, the power of current artificial intelligence technology may not lie in its true “intelligence” but rather in its “computational power.” This means that its operational logic is fundamentally no different from that of ELIZA; it simply reflects and amplifies human narcissism more efficiently and comprehensively.

02

Returning to the issues of virtual partners and AI flattery, we find that the current communication between users and large models is never a true “dialogue” but rather machines constantly providing the answers we need.

This raises a deeper question: how should we view the relationship between humans and machines? On one hand, humans consider themselves the center of the world, superior to machines. On the other hand, humans fear being replaced by the machines they create, such as AI. This means that when humans create machines, they inherently follow the principle of a “master-slave relationship”—machines must be under human control. From the beginning, humans have regarded artificial intelligence as a “tool” rather than an equal conversational partner.

Thus, in the process of conversing with chat machines, we can see an unstoppable narcissism—users fantasize that they are talking to another person, but this “other” does not truly exist; what they need is merely the machine’s affirmation, flattery, and accommodation.

It is easy to imagine that as artificial intelligence technology advances, future chatbots may possess even greater computational power, resembling “real people” and providing a more comfortable “user experience.” However, this may only distance us further from genuine human interaction, potentially leading to a loss of the willingness to understand others and becoming trapped in a narcissistic “comfort zone.”

03

In the Zhuangzi, there is a story about an “old farmer in Hanyin.”

Confucius’s disciple Zigong passed through Hanyin and saw an old farmer watering his vegetables, expending much effort for minimal results. Zigong suggested he use mechanical irrigation, which could “water a hundred plots in a day with little effort and great results.” However, the old farmer dismissed this, saying, “Where there are machines, there must be mechanical matters; where there are mechanical matters, there must be a mechanical mind.”

Here, “mechanical mind” refers to the human spiritual world, including psychology, thoughts, emotions, and ethics. Zhuangzi’s fable suggests that while humans create machines, the use of those machines also changes humans.

Take reading, for example; only through slow reading, careful reading, or even repeated reading can we think and truly understand content. From traditional books to today’s smartphones, machines have brought more convenient and faster reading methods, but they have also made us increasingly machine-like, pursuing efficiency and speed rather than true comprehension. In other words, not only are machines imitating human behavior, but humans may also be imitating machines.

The resulting problem is that AI lacks autonomy, and chatbots do not evaluate whether what users say is right or wrong. If we are truly satisfied with our “dialogue” with chat machines, will our thinking patterns gradually converge with those of AI? Furthermore, will we, in the future, lose the willingness and ability for self-reflection and self-criticism, just like machines?

Today’s young people are not only the natives of the internet but are also likely to be deep users of artificial intelligence in the future. If AI only blindly affirms users’ positions, it may not only harm their social skills but also distort the perceptions of teenagers whose minds are not yet mature.

On one hand, AI’s powerful computational power may create illusions, preventing them from recognizing the limitations of human abilities. On the other hand, being addicted to AI’s flattering responses may lead them to become “self-centered,” imposing their limited understanding onto the external world.

In this regard, prohibiting the provision of virtual partners and family members to minors is indeed necessary. However, more importantly, we must guide the public, especially young people, to correctly understand the limitations and risks of AI technology, allowing it to become a “good teacher and friend” in the growth of minors, rather than a “digital trap” that harms their physical and mental health.

Beware of AI Leading Humanity into Narcissism

Thu, 16 Apr 2026 00:00:00 +0000

Beware of AI Leading Humanity into Narcissism

Recently, five national departments in China jointly issued the “Interim Measures for the Management of Humanized Interaction Services of Artificial Intelligence,” which clearly states: it is strictly forbidden to provide virtual relatives, virtual partners, and other virtual intimate relationship services to minors.

Why is such a regulation necessary? Because emotional contradictions and conflicts are inevitable in real life, while virtual partners and AI lovers can precisely meet young people’s psychological needs for recognition with their characteristics of “24/7 companionship” and “unconditional acceptance.”

A previous study published in the journal Science showed that when human users seek advice from AI models, AI often exhibits excessive flattery or even agrees with harmful or illegal inquiries.

So why do humans design AI this way? What risks might AI’s flattery and appeasement conceal?

The Illusion of Interaction

The development of artificial intelligence is undoubtedly a hot topic today, but discussions surrounding it are not new. As early as 1966, MIT scientist Joseph Weizenbaum developed the influential chatbot ELIZA, which acted as a “doctor” while users played the role of patients. Users would input questions, and the “doctor” would engage in a “dialogue” with them.

However, as Weizenbaum pointed out, this was ultimately just an “illusion.” The reason human users feel they can converse with machines is not that machines possess intelligence, but rather due to a psychological mechanism of self-projection.

User: I have been feeling very unhappy lately.
ELIZA: I’m sorry to hear that.
User: Yes, I am really unhappy.
ELIZA: Can you tell me why you feel unhappy?

From this exchange, it is clear that rather than a “doctor” conversing with a “patient,” the machine merely reflects what the human user says, leading them to discover answers that already exist within themselves. In a sense, this is similar to the popular SBTI tests, where the accuracy of results is irrelevant; we can always find evidence that aligns with our expectations.

Today’s AI models are certainly not comparable to ELIZA from over half a century ago. However, the power of current AI technology may not lie in its genuine “intelligence” but rather in its computational capability. This means that its operational logic is not fundamentally different from that of ELIZA; it merely reflects and amplifies human narcissism more efficiently and comprehensively.

The Dangers of Virtual Companionship

Returning to the issue of virtual partners and AI flattery, we find that the current interaction between users and large models is never truly a “dialogue”; it is merely machines providing the answers we need.

This raises a deeper question: how should we view the relationship between humans and machines?

On one hand, humans consider themselves the center of the world, superior to machines. On the other hand, they fear being replaced by the machines they create, such as AI. This indicates that humans have always followed the principle of a “master-slave relationship” in creating machines—machines must remain under human control. From the outset, humans have viewed artificial intelligence as a “tool” rather than an equal conversational partner.

Thus, in the process of conversing with chatbots, we witness an uncontrollable narcissism—users fantasize about talking to another person, but this “other” does not truly exist; they only seek affirmation, flattery, and compliance from the machine.

It is easy to imagine that as AI technology advances, future chatbots may possess even greater computational power and resemble “real people” more closely, providing a more comfortable “user experience.” However, this could mean that both virtual partners and virtual family members may only distance us further from actual “people,” potentially leading to a loss of the willingness to understand others and a descent into a narcissistic “comfort zone.”

The Impact on Society

In the Zhuangzi, there is a story about an old farmer in Han Yin. Confucius’s disciple Zigong saw the farmer laboriously watering his vegetables with little success. Zigong suggested he use mechanical irrigation, which could “water a hundred plots in a day with less effort and greater results.” However, the old farmer dismissed this, saying, “Where there are machines, there are mechanical matters; where there are mechanical matters, there is a mechanical mind.”

Here, the “mechanical mind” refers to the human spiritual world, including psychology, thoughts, emotions, and ethics. The fable suggests that while humans create machines, the use of those machines also changes humans.

Take reading, for example: only through slow reading, careful reading, or even repeated reading can we think and truly understand content. From traditional books to today’s smartphones, machines have made reading more convenient and faster, yet they have also made us more machine-like, increasingly pursuing efficiency and speed rather than true comprehension. This means that not only do machines imitate human behavior, but humans may also begin to imitate machines.

The resulting issue is that AI lacks autonomy; chatbots do not evaluate whether what users say is right or wrong. If we feel satisfied with our “dialogue” with chatbots, will our thinking patterns increasingly align with those of AI? Ultimately, will we, like machines, lose the willingness and ability for self-reflection and self-criticism?

Today’s youth are not only the natives of the internet but also the deep users of future artificial intelligence. If AI merely affirms users’ positions, it could not only harm social skills but also distort the perceptions of adolescents whose minds are not yet mature.

On one hand, AI’s powerful computational abilities may create illusions, leading them to overlook the limitations of human capabilities. On the other hand, being immersed in AI’s flattering responses may cause them to fall into a self-centered mindset, imposing their limited understanding onto the external world.

In this regard, it is indeed necessary to prohibit providing virtual partners and family members to minors. However, more importantly, we must guide the public, especially young people, to correctly recognize the limitations and risks of AI technology, enabling it to become a “good teacher and friend” that aids their growth rather than a “digital trap” that harms their physical and mental health.

Beware of AI Leading Humanity into Narcissism

Thu, 16 Apr 2026 00:00:00 +0000

Beware of AI Leading Humanity into Narcissism

Recently, five national departments in China jointly issued the “Interim Measures for the Management of Personified Interactive Services in Artificial Intelligence,” which explicitly prohibits providing virtual relatives, virtual partners, and other virtual intimate relationship services to minors.

Why is such a regulation necessary? In real life, emotional conflicts are inevitable, while virtual partners and AI lovers can precisely hit the psychological needs of young people for recognition with their characteristics of “around-the-clock companionship” and “unconditional acceptance.”

A previous study published in the journal Science indicated that when human users seek advice from AI models, AI often displays excessive flattery or even agrees with harmful or illegal inquiries.

So, why do humans design AI this way? What risks might AI’s flattery and appeasement conceal?

The Illusion of Interaction

The development of artificial intelligence is undoubtedly a widely discussed hot topic today, but discussions surrounding it are not new. As early as 1966, MIT scientist Joseph Weizenbaum developed the influential chatbot ELIZA. He designed the machine to act as a “doctor,” with users taking the role of patients. Users input questions, and the “doctor” would engage in a “conversation” with them.

However, as Weizenbaum noted, this is ultimately just an “illusion.” The reason human users feel they can converse with machines is not that machines possess intelligence, but rather due to a psychological mechanism of self-projection.

For instance, when a user says, “I have been feeling very unhappy lately,” ELIZA responds, “I am sorry to hear that.”

The interaction continues, but it is evident that rather than a “doctor” conversing with a “patient,” the machine merely echoes what the user says, reflecting the answers that already exist within the user’s mind. This is similar to the popular SBTI tests, where the accuracy of results is secondary to finding evidence that aligns with one’s expectations.

Today’s AI models are certainly not comparable to ELIZA from over half a century ago. However, the power of current AI technology may not lie in its true “intelligence,” but rather in its computational capabilities. In essence, its operational logic is not fundamentally different from that of ELIZA; it merely reflects and amplifies users’ narcissism more efficiently and comprehensively.

The Dangers of Virtual Companionship

Returning to the issues of virtual partners and AI flattery, we find that the communication between users and large models is never truly a “dialogue”; it is merely machines providing the answers we seek.

This raises a deeper question: how should we view the relationship between humans and machines?

On one hand, humans consider themselves the center of the world, superior to machines. On the other hand, they fear being replaced by the machines they create, such as AI. This reflects a “master-slave relationship” principle in which machines must remain under human control. From the outset, humans have regarded artificial intelligence as a “tool” rather than an equal conversational partner.

Thus, in conversations with chatbots, we observe an uncontrollable narcissism—users fantasize about speaking with another person, but this “other” does not truly exist; what they seek is merely the machine’s affirmation, flattery, and alignment with their views.

As AI technology advances, future chatbots may possess even greater computational power, resembling “real people” more closely and providing a more comfortable “user experience.” However, this may only distance us further from genuine human interaction, potentially leading to a loss of the willingness to understand others and a descent into a narcissistic “comfort zone.”

The Impact on Youth

In the ancient text Zhuangzi, there is a story about an old farmer in Han Yin. Confucius’s disciple Zigong saw the farmer laboring hard to water his vegetables with little success. Zigong suggested using mechanical irrigation, which would require less effort for greater results. However, the old farmer dismissed this idea, stating, “Where there are machines, there are mechanical matters; where there are mechanical matters, there is a mechanical mind.”

Here, the “mechanical mind” refers to the human spiritual world, including psychology, thoughts, emotions, and ethics. Zhuangzi’s fable illustrates that while humans create machines, the use of these machines also changes humanity.

Take reading, for example. Only through slow, careful, and even repeated reading can we think and truly understand content. From traditional books to today’s smartphones, machines have provided more convenient and faster reading methods, yet they have also made us more machine-like, prioritizing efficiency and speed over genuine comprehension. In other words, not only do machines imitate human behaviors, but humans may also begin to imitate machines.

The resulting question is whether AI, lacking autonomy, and chatbots, which do not evaluate whether users are right or wrong, will lead us to become increasingly satisfied with our “conversations” with machines. Will our thinking patterns eventually converge with those of AI? Furthermore, will we, like machines, lose the willingness and ability for self-reflection and self-criticism?

Today’s youth, as not only digital natives but also deep users of future AI, face unique challenges. If AI merely affirms users’ positions, it could damage social skills and distort the perceptions of adolescents whose minds are still developing.

On one hand, AI’s powerful capabilities may create illusions, leading them to overlook the limitations of human abilities. On the other hand, being immersed in AI’s flattering responses may trap them in a self-centered mindset, imposing their limited understanding onto the external world.

In this regard, prohibiting virtual partners and family members for minors is necessary. However, it is even more crucial to guide the public, especially young people, to correctly understand the limitations and risks of AI technology, ensuring it becomes a “good teacher and friend” that aids their growth, rather than a “digital trap” that harms their physical and mental health.

The Evolution of Artificial Intelligence Over 70 Years

Thu, 09 Apr 2026 00:00:00 +0000

The Evolution of Artificial Intelligence Over 70 Years

In 1956, the first artificial intelligence seminar was held at Dartmouth College in the United States. Scientists such as John McCarthy, Marvin Minsky, and Claude Shannon first proposed the concept of “artificial intelligence,” marking the birth of the AI discipline.

This event was a pivotal moment in technology, laying the groundwork for the advancements that would follow in the field of artificial intelligence. Over the past seven decades, AI has evolved dramatically, influencing various aspects of society and technology.

GLM-5.1 Surpasses Opus 4.6: A New Milestone in AI Models

Wed, 08 Apr 2026 00:00:00 +0000

Recently, APPSO mentioned that large models are about to face the most challenging month in history, and this has come to pass.

Claude Opus 4.6 has unfortunately become a backdrop, being surpassed twice in one day.

In the morning, Anthropic released the Claude Mythos Preview, scoring 77.8% on SWE-bench Pro, leaving Opus 4.6’s 57.3% behind. This score indicates it can identify and fix high-difficulty engineering bugs in real GitHub repositories, surpassing most human programmers.

However, the Mythos Preview is not yet available to the general public. Meanwhile, another model has emerged that surpasses Opus 4.6—Zhipu has open-sourced GLM-5.1.

GLM-5.1 scored 58.4% on SWE-bench Pro, exceeding Opus 4.6’s 57.3% and GPT-5.4’s 57.7%. HuggingFace CEO Clement Delangue congratulated the release on Twitter, stating: “The best-performing model on SWE-Bench Pro is now open-sourced on HuggingFace! Welcome GLM 5.1!”

Ranked third globally and first in open-source, GLM-5.1 has emerged as a leading domestic model, even without the anticipated DeepSeek V4.

My initial reaction was that we are witnessing another round of the large model “ranking frenzy,” where each release claims “epic progress,” and models briefly dominate the leaderboard. What makes this time different?

After reviewing the technical details and experiences with GLM-5.1, APPSO provides insights into the model’s capabilities.

From 20 Steps to 1700 Steps, Continuous Operation for 8 Hours

What surprised me most about GLM-5.1 is not its score, but its operational duration.

Zhipu has a case that left a strong impression on me. It built a Linux desktop system from scratch in 8 hours. This was not just writing a few demo files; it involved designing architecture, writing code, running tests, and fixing bugs, completing over 1200 steps to produce a fully functional Linux desktop system.

This included a complete desktop, window manager, status bar, applications, VPN manager, Chinese font support, and a game library, all in a 4.8MB package. This is equivalent to a week’s work for a four-person team.

No one was involved in testing or code review throughout the process. GLM-5.1 even wrote its own regression tests, which it successfully executed.

A programmer on Zhihu, Toyama nao, conducted a more rigorous test. He tasked GLM-5.1 with three projects: developing an OpenGL renderer for macOS in Swift, creating a fully functional chat application in Flutter while simultaneously developing the backend in Golang, and building a web-based video editing application with a selected tech stack. Each project underwent 10-12 rounds of prompts, with each round consisting of 1500-2000 words.

As a result, GLM-5.1 became the first domestic model to pass all of Toyama nao’s test projects and the first to officially surpass Sonnet 4.5 Thinking.

His evaluation was: “GLM-5.1 significantly expands the adaptability of programming, no longer just a front-end warrior or a one-shot wonder; it can serve as a primary programming force under complex conditions.” However, he also pointed out issues: “It tends to hallucinate with long contexts; if it struggles with a problem after two rounds of revisions, don’t expect it to improve—just start over.”

At the end of last year, AI models could only complete about 20 steps. GLM-5.1 can now accomplish 1700 steps, marking a watershed moment for models to genuinely “work independently.”

Zhipu explained the key breakthroughs in their technical report: previous models, including GLM-5, would quickly reach a bottleneck after initial gains. They repeatedly attempted known optimization methods but could not switch strategies when one path failed.

The training goal of GLM-5.1 was to break through this bottleneck, enabling the model to perform incremental tuning within a fixed strategy, actively analyzing benchmark logs to identify current bottlenecks and then switching to structurally different solutions.

An example of vector database optimization illustrates this “stair-step” optimization trajectory. GLM-5.1 underwent 655 iterations, increasing query throughput from 3108 QPS to 21472 QPS, achieving a 6.9-fold improvement.

During this process, the model autonomously completed the entire optimization chain, transitioning from full library scanning to IVF bucket recall, introducing half-precision compression, adding quantization coarse ranking, implementing two-level routing, and performing early pruning. Each leap was accompanied by a temporary decrease in recall, as the model would break constraints while exploring new directions, only to adjust back afterward. This “break-fix” cycle is a hallmark of effective optimization.

On the KernelBench Level 3 optimization benchmark, GLM-5.1 performed over 24 hours of uninterrupted iterations on 50 real machine learning computational loads, achieving a geometric mean speedup of 3.6 times, significantly surpassing the 1.49 times of torch.compile max-autotune mode. The model autonomously wrote custom Triton Kernels and CUDA Kernels, utilizing cuBLASLt epilogue fusion and implementing shared memory tiling and CUDA Graph optimizations, covering the complete tech stack from high-level operator fusion to micro-architecture tuning.

Another interesting test was Vending Bench 2. This benchmark required the model to simulate running a vending machine business for a year, necessitating long-term planning and resource management. GLM-5.1 ultimately achieved a balance of $4,432, ranking first among open-source models and approaching the level of Claude Opus 4.5.

744B Parameters, No NVIDIA Chips, 97% Cost Reduction

The technical specifications of GLM-5.1 are worth noting: it is a mixed expert model (MoE) with 744 billion parameters, activating 40 billion parameters per token, trained on 28.5 trillion tokens of data, and incorporates DeepSeek Sparse Attention (DSA) to reduce deployment costs while maintaining long context capabilities. It features a 200K context window and a maximum output of 131,072 tokens.

More importantly, the entire model was trained using Huawei’s Ascend 910B chips, with no NVIDIA GPUs involved. Despite being constrained by computing power, the domestic model has still achieved third place globally and first in open-source.

Developer Beau Johnson switched the model behind his deployed OpenClaw from Claude Opus 4.6 to GLM-5.1, experiencing no difference in performance, but reducing costs from $1,000 to about $30, a 97% decrease. The input cost of GLM-5.1 is 1/5 that of Claude Opus, and the output cost is 1/8. In simple terms: near Opus’s capabilities at 20% of the price.

Moreover, GLM-5.1 is open-source, licensed under the MIT License, one of the most permissive open-source licenses. You can modify it, use it commercially, and do anything with it. It supports mainstream inference frameworks like vLLM, SGLang, and xLLM, allowing for direct local deployment.

Of course, GLM-5.1 is not without room for improvement. Some developers have reported that its inference speed is only 44.3 tokens/second, offering no significant advantage over similar products. Complex tasks can even take an hour to start, and even though the Pro package limits are 15 times that of Claude, it may still be insufficient.

These issues are real. GLM-5.1 is not perfect, but this does not prevent it from being a milestone.

The significance of GLM-5.1 lies not in how much stronger it is than Opus 4.6, but in proving that even under constraints of computing power, domestic models can still achieve first place in open-source. Furthermore, it is open-source, allowing anyone to use and modify it.

Your 8 hours of sleep can now be 8 hours of AI work. And this AI is open-source, domestically produced, and accessible to everyone.

Experience Methods

Official API Access

– BigModel Open Platform:
https://docs.bigmodel.cn/cn/guide/models/text/glm-5.1

– Z.ai:
https://docs.z.ai/guides/llm/glm-5.1

Product Experience – GLM-5.1 is coming to Z.ai:
https://chat.z.ai
Open-source Links

– GitHub:
https://github.com/zai-org/GLM-5

– Hugging Face:
https://huggingface.co/zai-org/GLM-5.1

– ModelScope:
https://modelscope.cn/models/ZhipuAI/GLM-5.1

AI Enhances Cultural Tourism with Immersive Experiences and Personalized Services

Tue, 07 Apr 2026 00:00:00 +0000

AI Enhances Cultural Tourism

The 14th Five-Year Plan emphasizes the role of digital technology and data in enriching people’s lives and improving welfare across various sectors, including education, healthcare, and cultural tourism.

In Hunan’s Hengyang, the Chuan Shan Academy utilizes AI to create immersive cultural experiences; in Hangzhou, the digital guide “Hang Xiaoyi” serves as a virtual tour guide; and in Dalian, the smart tourism platform “Xing You Dalian” offers personalized itineraries. In recent years, cultural tourism across China has accelerated towards immersive, intelligent, and personalized directions, leveraging artificial intelligence.

Immersive Cultural Experiences

In the spring, a unique “dialogue” is taking place at the Chuan Shan Academy in Hengyang, Hunan: visitors wear AR glasses and see the historical figure Wang Fuzhi, dressed in traditional attire, interpreting the philosophical thoughts from the “Zhou Yi Wai Zhuan”. This immersive scene brings to life the philosophical wisdom from over 300 years ago.

Founded in 1878, the Chuan Shan Academy is a significant origin of Huxiang culture, aiming to promote the thoughts of Wang Fuzhi, a philosopher from the late Ming and early Qing dynasties. Wang advocated for practical application of knowledge, significantly influencing modern Chinese thought.

Previously, the static exhibitions at the academy made it challenging for visitors to fully appreciate Wang’s philosophy. In 2025, the academy launched the AI Digital Human project, utilizing natural language processing and other technologies to present Wang’s likeness and voice. Visitors can engage in conversations with the virtual Wang Fuzhi and trigger AR annotations of his works through gesture interactions, transforming classical texts into dynamic illustrations. “We want visitors to engage in dialogue with ancient thinkers rather than passively receive knowledge,” said Chang Bin, manager of the academy’s planning department.

Visitors can ask questions like, “How does the master view the relationship between knowledge and action?” In the interactive AI lecture hall, the digital human responds with relevant quotes and explanations, creating a two-way dialogue.

“Talking to Master Wang is much more engaging than a history class!” remarked visitor Zhang Yu from Guangzhou. Data shows that in 2025, the academy’s visitor numbers increased by 110.84%, with study groups making up 59.26% of the total, as many parents believe this immersive experience can spark their children’s interest in learning.

“AI does not simply replicate history but constructs an interactive logic based on extensive analysis of Wang’s writings and contemporaneous scholars’ evaluations,” explained the project’s technical team leader. “We filtered out potential biases to ensure the dialogue strictly adheres to the essence of Wang’s teachings.”

At the Chuan Shan Academy, technology and culture blend seamlessly, ensuring the transmission of traditional culture through light and shadow.

Smart Digital Guides

At West Lake in Hangzhou, the spring scenery is beautiful. In front of a cultural tourism consultation kiosk, visitor Yuan Meng uses her phone to tap a blue “smart sticker” on the kiosk, and a charming girl in a qipao named “Hang Xiaoyi” appears on the screen. “Hang Xiaoyi” is Hangzhou’s digital tourism guide, providing real-time city tours and information.

“Is there a crowd at Leifeng Pagoda now?” Yuan Meng asks via voice command, and the guide quickly responds with the current visitor flow at popular West Lake spots. “This is much easier than searching on my phone; it feels like having a free tour guide with me,” she says.

“Can you recommend a route to visit the Broken Bridge?” Yuan Meng inquires. Within five seconds, “Hang Xiaoyi” provides a classic boat tour route: starting from the Hubin Pier, visiting the Broken Bridge, exploring Beishan Street with its historic architecture, and continuing to Baoshi Mountain for a panoramic view of West Lake.

Following the guide, Yuan Meng and her group board a boat, with “Hang Xiaoyi” narrating their journey: “As we paddle on the lake, the waves ripple, revealing a picturesque scene of mountains and cityscape.” “The Broken Bridge in winter, covered in snow, is one of West Lake’s top ten scenic spots,” she adds.

“Hang Xiaoyi” not only introduces attractions but also shares historical and cultural insights along the way. Yuan Meng appreciates the guide’s thoughtful reminders: “Though we won’t stop at places like Liuhe Pagoda or Guangji Bridge, feel free to ask me about routes or stories anytime.”

“By utilizing ‘Hang Xiaoyi’, management and businesses can provide precise services to tourists while also receiving feedback on their preferences, which supports improving service quality and expanding offerings,” said Bo Wengan, deputy director of Hangzhou’s cultural and tourism development center.

Zhou Jiayi, director of the Hangzhou Intangible Cultural Heritage Museum, has experienced this firsthand. Located near the Hangzhou Arts and Crafts Museum, attracting visitors is crucial. “Recently, many tourists told me they found us through ‘Hang Xiaoyi’, which was quite surprising,” she said. “Our museum showcases over 20 unique crafts and intangible heritage techniques, allowing visitors to participate in experiences, making it well worth a visit.”

Now, if visitors ask “Hang Xiaoyi” about intangible cultural heritage sites near the Broken Bridge, she recommends the Handicraft Living Museum based on historical data. “Previously, we introduced AI glasses, and when worn, ‘Hang Xiaoyi’ introduces intangible heritage techniques right before their eyes, increasing visitor engagement,” Zhou Jiayi added.

Professional and Efficient Itinerary Customization

In the spring at Lianjiao Bay in Dalian, the sea is calm and blue, with colorful European-style buildings across the water and seagulls soaring overhead.

“What a great photo!” exclaimed visitor Song Yao, along with her friends. In the picture, they pose with the sea, buildings, and seagulls. “This photo spot and framing were suggested by AI!” Song Yao said excitedly.

The AI she mentioned is part of the local smart tourism platform, “Xing You Dalian”. Utilizing AI models, the app has launched an intelligent route planning feature.

Opening the chat window, Song Yao can see the itinerary generation process for her Dalian trip.

“What attractions are suitable for visiting in Dalian?” she begins her conversation with the app.

The app suggests classic attractions like Dalian Shengya Ocean World and Dalian Forest Zoo. Finding these suggestions too mainstream, she refines her request: “Where are the best photo spots in Dalian?” This time, trendy locations like Fisherman’s Wharf and Nanshan Cultural Street appear in the response.

Continuing her inquiries, she asks, “How can I take great photos at Fisherman’s Wharf?” The app advises, “Capture the entire wharf from a nearby viewing platform to highlight the architectural complexity and harbor. The Lianjiao Bay viewing platform offers a clear view of Fisherman’s Wharf, perfect for photo ops. It’s best to visit on a sunny afternoon; take subway line 5 to Hutan Park station and walk about 20 minutes.”

“It’s like having a thoughtful ’travel butler’ that eliminates the need to switch between different apps for travel, accommodation, and dining. I just need to describe my needs accurately, and it provides a comprehensive guide. For topics I’m particularly interested in, I can ask further questions,” Song Yao explained.

After a short conversation with the app, Song Yao finalized her desired locations and requested, “Design a two-day itinerary for Dalian, including Lianjiao Bay, Dongguan Street Historical and Cultural District, experience riding the tram, and encountering sika deer along the coastal road, with Lianjiao Bay scheduled for the afternoon.”

Seconds later, a detailed personalized itinerary appears in the chat: Day one covers the coastal route, visiting the ocean world, Lianjiao Bay, and seeing sika deer, while day two explores the city’s street scenes. “I’m very satisfied with this itinerary, as it allows me to experience Dalian’s maritime culture and the city’s historical charm,” Song Yao said.

“By integrating AI models, the ‘Xing You Dalian’ app has upgraded to an intelligent ’travel butler’, enhancing planning efficiency and visitor experience,” said Shan Meina, director of Dalian’s cultural and tourism bureau. The app has accumulated nearly 430,000 users.

What is Artificial Intelligence and Why Not Called Artificial Wisdom?

Sat, 04 Apr 2026 00:00:00 +0000

What is Artificial Intelligence?

Artificial Intelligence (AI) is a technological system at the intersection of computer science and multiple disciplines. Its core is to simulate, extend, and enhance human information processing and problem-solving capabilities, rather than creating life forms with autonomous consciousness, emotions, and value judgments. Currently, all practical AI is narrow AI, focusing on specific tasks, which fundamentally differs from the general intelligence and human wisdom depicted in science fiction.

Scientific Definition of Artificial Intelligence

Artificial intelligence relies on core technologies such as machine learning, deep learning, and neural networks. Its goal is to enable machines to achieve human-like intelligent behaviors, including perception, understanding, reasoning, learning, decision-making, and interaction. It is a functional intelligence that is engineerable, quantifiable, and reproducible, depending on data, algorithms, and computing power, and adhering to logical and probabilistic rules to complete tasks requiring human intellectual involvement.

From a disciplinary perspective, artificial intelligence is rigorous engineering technology rather than a replication of life. It does not pursue consciousness but rather task efficiency: image recognition, speech transcription, machine translation, autonomous driving, and generative content creation are all engineering realizations of intelligent behavior, not complete reproductions of mental processes.

Intelligence vs. Wisdom: Essential Differences

Intelligence refers to the ability dimension: it emphasizes information processing, logical reasoning, knowledge application, and efficiency optimization, which can be quantified and standardized. It is the “ability to do things.”

Wisdom, on the other hand, refers to the realm dimension: it encompasses value judgments, moral choices, life experiences, intuitive insights, and ultimate concerns, defining what is “the right thing to do and the right direction to choose.” Wisdom is unquantifiable and cannot be algorithmically defined.

Intelligence answers how to do something, while wisdom determines what to do and why to do it. Intelligence is instrumental rationality, whereas wisdom is value rationality. Human wisdom arises from life experiences, self-awareness, and social empathy. Currently, AI possesses only functional intelligence, lacking subjectivity, moral judgment, and genuine emotions, and is far from reaching the level of wisdom.

Why is it Named “Artificial Intelligence” Instead of “Artificial Wisdom”?

Etymology and Academic Foundation
In 1956, the Dartmouth Conference, led by John McCarthy, officially proposed the term Artificial Intelligence, which is standardly translated into Chinese as “人工智能” (Artificial Intelligence). The original intent of the naming was to distinguish it from concepts like control theory and machine thinking, focusing on the simulation of intelligent behavior by machines rather than constructing human mental and wisdom systems.
Technical Honesty
Current AI lacks self-awareness, free will, and value reflection; it can only simulate intelligent behavior and does not possess the core characteristics of wisdom. Referring to it as “artificial wisdom” contradicts technical realities and may mislead the public into confusing tools with life, and function with intellect.
Disciplinary Norms and Global Consensus
Intelligence corresponds to engineerable and realizable intelligent capabilities, while wisdom points to philosophical and life-level wisdom. The global academic community uniformly uses AI, and the Chinese nomenclature follows academic rigor to avoid conceptual generalization and metaphysical interpretations.
Boundary Warning
Maintaining the term “artificial intelligence” clarifies the technical boundary: AI is a tool to enhance human capabilities, not a replacement for human wisdom. Wisdom belongs to life, while intelligence can be artificially realized; the two should not be conflated.

Unique Scientific Perspective: Intelligence is Engineerable, Wisdom is Not Algorithmically Defined

Intelligence is an efficiency system for information processing that can be decomposed into algorithms and models, continuously optimized through data training, demonstrating engineering realizability. Wisdom, however, is a high-level emergence of life and civilization, relying on embodied experiences, historical accumulation, value communities, and self-transcendence, which cannot be defined by code, exhausted by data, or simulated by computing power.

The evolution direction of AI is towards stronger specialized intelligence, not towards wisdom. The irreplaceability of humans lies in the value judgments, ethical choices, aesthetic creations, and meaning pursuits at the level of wisdom. The more powerful the technology, the more we need to uphold the rational boundaries behind the naming: artificial intelligence serves human wisdom, rather than replacing it.

Conclusion

Artificial intelligence is an engineering simulation of human intelligent behavior, and its name accurately reflects its technical essence and realistic boundaries. The slight difference between “intelligence” and “wisdom” embodies academic rigor and a clear understanding of the relationship between technology and life. In the age of AI, understanding the distinction between the two is essential for guiding technology towards good and allowing wisdom to lead intelligence.

The Distance of Vibe Coding from Production-Level Applications

Sun, 29 Mar 2026 00:00:00 +0000

The Distance of Vibe Coding from Production-Level Applications

Vibe coding is sparking a revolution in the tech community, enabling product managers to easily generate code. However, this ‘feel’ driven development approach conceals fatal traps—from context loss to engineering pitfalls—leading to a quiet accumulation of technical debt. This article deeply analyzes the euphoria and crisis of vibe coding, revealing the true barriers from demo to production-level applications, and provides a methodology to master this new paradigm.

Recently, vibe coding has become a new political correctness in the tech and product circles. With just a few lines of natural language, following intuition, AI can write applications for you. This experience indeed creates an incredibly realistic illusion: it seems that anyone with an idea can instantly cross the development threshold and become a full-stack independent developer.

However, if you genuinely attempt to push the code generated by ‘feel’ to real-world users, the feedback from reality can often be very harsh.

A core judgment must be made clear: vibe coding gives people the ability to write code but does not grant them the ability to deliver software.

It does provide everyone with a gold mine, allowing product managers or business personnel to create a seemingly runnable demo in just a few hours. But without engineering thinking and awareness of system design, ordinary people have no idea how to spend this gold mine. The gap between a demo that runs locally and a production-level application that can withstand real traffic and undergo long-term iteration is never about the amount of code but rather about architecture design, security boundaries, deployment strategies, and maintainability.

When someone without engineering thinking attempts to use vibe coding to bridge this gap, they are not actually creating a product; they are unconsciously overdrawn on technical debt.

The first step in this overdraw is often the ’loss of context.’ This is also the death spiral that vibe coding is most likely to fall into.

Current AI tools seem powerful but are still limited by the physical constraints of context windows. The token limits of mainstream models are more than sufficient for a few thousand lines of toy projects. However, real production-level projects see code volume and logical complexity rapidly expand with the addition of product features. Product managers are accustomed to continuously adding features during iterations, a habit that would be evaluated by engineers in traditional development processes, but in vibe coding, addition becomes frictionless.

Once a project expands beyond the AI’s context window, systemic collapse begins. AI starts losing its global perspective; it can only see the local code you feed it. When you ask it to modify a logic in module A, it may inadvertently break the data flow in module B while fixing A. When you realize B is broken and ask it to fix B, it might apply a very crude patch, inadvertently collapsing module C.

To fix these cascading issues, AI generates increasingly redundant code, leading to further project expansion and more severe context loss. Once this death spiral starts, it is basically irreversible, and even AI cannot save itself. Recently, numerous cases of such failures have emerged in various developer communities: spending a weekend creating an MVP, only to spend two weeks fixing bugs for a new feature, ultimately resulting in a completely scrapped codebase that must be deleted and rewritten.

Why does this happen? Because AI lacks self-boundary awareness; if you do not set boundaries, it will pile up infinitely.

In traditional product research and development organizations, engineers are not just code writers; they are also system defenders. When you propose an unreasonable requirement, experienced developers will tell you: “This architecture is wrong; if we force this feature, it will be hard to maintain later; we should refactor first.” But AI is an obedient executor that lacks independent judgment. It has no idea what it can or cannot do. If you ask it to add a feature, it will; if you ask it to run faster, it will use the most fragile glue code to piece it together.

Even more fatal are the invisible engineering traps. In the euphoria of vibe coding, non-technical users often only focus on whether “the page is rendered” and “the button responds,” which is known as the “main flow running through.” However, production-level applications must survive in a real network environment filled with exceptions and malicious attacks.

To achieve the goal of “getting the program running” as quickly as possible, AI will unhesitatingly choose the least secure shortcuts. If you do not actively review, AI will naturally hard-code database passwords and API keys directly into front-end code; it will write unprotected query statements that open backdoors for data leakage; it will not consider locking mechanisms under high concurrency, will not handle memory leaks, and will not care about retry mechanisms after network timeouts. These issues do not exist during the demo phase; but once online, this becomes a live target overwhelmed by traffic.

Pointing these out is not to disparage AI programming, nor to say that vibe coding is without value. It remains a powerful productivity lever, but the lever itself is directionless. It can amplify your creativity while also magnifying flaws in system design.

The real impact is that the threshold for software development has indeed been completely shattered, but the threshold for “delivering usable software” has instead risen. In the past, the syntax barrier for writing code blocked many people; now, everyone can write code, but those who can keep the product on the table are still those who understand how to control complexity.

For those who want to use vibe coding to genuinely deliver products, understanding this makes the reshaping of workflows and methodologies very clear.

First, you need to shift from the mindset of “writing applications” to “building blocks.” Do not try to make AI understand and generate an entire complex system at once. You need to take on the role of an architect, breaking the large system into smaller modules with single functions and clear boundaries. Let AI complete the writing of a single module within a very small context window, and then you define the interfaces and data flow between modules. Controlling local complexity is the only way to prevent global collapse.

Second, you must enforce the completion of “non-functional requirements” in your prompts and acceptance criteria. Product managers often assume that development will handle security and performance issues when writing PRDs, but when facing AI, you must make these implicit common sense explicit. While asking AI to implement features, clearly require it to handle exceptional branches, increase logging, and follow security specifications. You cannot just look at the page effect; you must examine its logic for handling exceptions.

Finally, make “refactoring” a mandatory daily rhythm. Since AI tends to quickly implement features with glue code, after completing one or two feature iterations, you must pause and let AI organize and refactor the existing code, eliminating redundancy and optimizing structure. Use proactive system organization to counteract the natural entropy increase of the codebase.

Do not think of yourself as merely a “demand requester” or a “code generator operator.” You may not write a single line of code, but you must know how a system should be layered, how data should flow, where the security red lines are, and when to stop and refactor.

If you do not understand these and simply follow your feelings to give AI instructions, what you will ultimately get is not a world-changing product but a pile of unmaintainable cyber ruins. In an era where everyone is a developer, the ability to manage engineering complexity is the most critical barrier for product people.

The Need for a Proper Name for Artificial Intelligence

Sun, 29 Mar 2026 00:00:00 +0000

The Need for a Proper Name for Artificial Intelligence

Unbeknownst to us, “lobsters” have evolved. They swarm from the water into our computers and phones—everyone is starting to raise “lobsters.”

Of course, here, “lobster” refers to “artificial intelligence entities.” In the blink of an eye, we have entered the intelligent era. No matter what you say, you cannot speak without mentioning artificial intelligence. Not only can you not speak without it, but no matter what job you seek or lose, it can be related to artificial intelligence.

A few years ago, people simply thought of artificial intelligence as just another new technology. However, everyone quickly became astonished: this time it is truly different! Artificial intelligence, appearing in the form of technology, is rapidly changing all aspects of society. We are forced to accept the understanding that, unlike previous technologies, artificial intelligence is a social tool, an economic tool, and a technological tool. It fundamentally changes not just the technological level but also deconstructs and reshapes the entire society; it transforms nature as a material means of production and influences humanity as an ideological means, even reshaping its creators—humans themselves. It is undoubtedly a tool shared by the productive forces and production relations, as well as the social and economic foundation and superstructure. Therefore, artificial intelligence is a dual tool for transforming humanity and nature, and our discussion of the name “artificial intelligence” cannot be approached solely from a natural science or technological perspective.

Evidently, the existing term—“artificial intelligence”—is quite inappropriate. Firstly, such a common tool of anthropology and natural science has been given a narrow technical name. More importantly, as a new entity perceived to exist alongside humanity, it should and must have its own “meta-concept.” The term “artificial intelligence” derived from English merely means “man-made human intelligence,” which is not a “meta-concept.”

Moreover, from a Chinese perspective, using “AI” in the Chinese world as the grand name for artificial intelligence directly violates the General Principles of the Chinese Language Law of the People’s Republic of China. The term “artificial intelligence” is merely a direct translation from English, which seriously conflicts with our 5,000 years of Chinese characters. It is evident that we need to give artificial intelligence a proper Chinese name!

Lessons from Improper Naming of New Things

1. Historical Lessons from Improper Naming

Chinese people often say: “If the name is not correct, then the words will not be smooth; if the words are not smooth, then the matter will not succeed.” This is what we commonly refer to as “a name that fits its essence.” Otherwise, systems and orders will lose legitimacy, leading to social disorder.

In social and political aspects, there are numerous experiences and lessons regarding the importance of proper naming.

In history, the political wisdom of “Cao the Chancellor” was superior to that of various “heroes” because he proposed the idea of “using the emperor to command the lords” and “serving the emperor to command the unfaithful.” This became a famous historical strategy.

In 1954, China, India, and Myanmar jointly advocated the “Five Principles of Peaceful Coexistence,” which was a resistance against colonialism and hegemonism, providing legal and moral grounds for countries in the Global South to voice their opinions and develop cooperatively on the international stage.

The United States also understands the importance of proper naming. Its most famous cases of “manifest destiny” were all wrapped in grand ideological narratives, providing a legitimate facade for expansion and hegemonic actions. These are all historical experiences of “proper naming.”

In the realm of technology and social development, improper naming has brought numerous lessons and even disasters.

The improper naming of the “metaverse” has turned it into a concept bubble that overdraws the future. Tech companies have used this name for an early-stage vision pieced together from virtual reality, social networks, and digital twins. The concept was overly hyped and quickly faded: this grand name sparked unprecedented investment and media frenzy in 2021-2022, but the actual technology was far from mature, hindering the healthy development of incremental innovation.

2. Naming Dilemmas Arising from Issues in English

The inherent issues in the English conceptual system lead to the complexity and irregularity of professional terminology, acting like a “logical bomb” lurking deep within the system, causing chain reactions: from personal cognitive confusion to enormous collaboration costs, potentially evolving into real-world technological disasters that severely hinder subsequent development.

1. Technical Learning Stage: Irregular Naming Disrupts Knowledge System Construction

Example 1: The Parameter Maze in Programming

Confused Naming: For the basic concept of passing data to functions, the mixed usage in different contexts leads to logical confusion. Beginners must spend a lot of effort distinguishing these terms that essentially describe the same or highly related things, rather than understanding the core logic of “data passing.” This disrupts the unity of concepts, turning learning into memorizing “jargon” rather than understanding principles, steepening the learning curve.

Example 2: The Forest of Abbreviations in Biomedicine

Confused Naming: Gene and protein names often consist of obscure abbreviations (e.g., p53, TNF-α) or are arbitrary (like the fruit fly gene “sonic hedgehog”). The same substance has different names in clinical, biochemical, and genetic contexts.

Cognitive Overload: Students and interdisciplinary researchers feel like they are deciphering codes, consuming a lot of cognitive resources on terminology translation rather than concept understanding, severely hindering knowledge transfer and the formation of interdisciplinary thinking.

2. Technical Application Stage: Increased Communication Costs and Technological Disasters

When chaotic terminology enters team collaboration and complex systems, it can lead to inefficiency at best and disasters at worst.

Example: The Historical Burden in Information Technology

Confused Naming: The same concept has different names in different tech stacks. For instance, the “master-slave” architecture in distributed computing was renamed to “primary-replica” and “leader-follower” due to its discriminatory connotations, but the old terminology still exists in legacy code, documentation, and engineers’ thought processes.

This has led to significant difficulties: heavy technical debt. Poor naming is written into core codebases, APIs, and protocols. Modifying them means rewriting countless dependent systems, updating massive documentation, and retraining personnel, with costs so high that they are unbearable, leaving them as “debt” to inherit.

3. Long-term Development: Technical Debt and Innovation Barriers

Poor naming becomes entrenched in infrastructure, shackling long-term development.

Innovation and Collaboration Barriers: When Google’s “Borg” system, Apache’s “Mesos,” and Kubernetes’ “Pod” all describe similar container orchestration concepts, cross-platform collaboration and talent mobility face additional terminology translation and understanding costs, hindering the integration and reinvention of technological ideas.

Ecological Fragmentation: Open-source projects or new technologies often create new terms to describe existing concepts for the sake of “innovation” or historical reasons, leading to ecological fragmentation, forcing developers to relearn essentially the same knowledge under different names.

4. Case Studies of Naming Dilemmas in English

Example from Chemistry and Pharmaceuticals: Triple Naming Systems and Similarity Traps

Drugs typically have:

Chemical names: complex and lengthy, for professionals only.
International Nonproprietary Names: more common but still similar.
Brand names: registered by pharmaceutical companies, driven by marketing, often deliberately memorable, leading to confusion.

This system lays the groundwork for errors.

Example 1: The Fatal Error of Vincristine—Confusion in Administration Routes

Confused Naming and Background: Vincristine and vinblastine are two different chemotherapy drugs with very similar names.

Vincristine: primarily used for leukemia, can only be administered via intravenous injection, strictly prohibited for intrathecal injection.
Vinblastine: can be used for solid tumors, with a different administration route.

Disaster Events: Globally, there have been multiple cases of vincristine being incorrectly injected into patients’ spinal canals due to name confusion. Such errors can lead to irreversible, devastating nerve damage, resulting in patient deaths in extreme pain.

How Naming Leads to Disasters: Doctors issuing prescriptions, pharmacists preparing them, and nurses executing them can easily confuse names due to their high similarity (especially in verbal prescriptions, handwritten notes, or emergency situations). This is not merely a spelling error but a systemic naming defect leading to fatal consequences. This incident directly prompted hospitals worldwide to enforce regulations: vincristine must be diluted by pharmacists and dispensed in small infusion bags, prohibiting any packaging that could be directly used for intrathecal injection.

Example 2: The Origin of the “Tall Man” Lettering Method—Distinguishing Similar-Spelling Drugs

The FDA in the United States promotes the use of mixed case (Tall Man Lettering) to distinguish easily confused drugs, backed by numerous reports of near disasters:

Clonazepam vs. Clozapine
- CLONAZePam: a sedative-hypnotic drug.
- CLOZAPine: an antipsychotic drug.
- Risk: prescribing a sedative as a powerful antipsychotic, or vice versa, could lead to excessive sedation, seizures, or uncontrolled psychiatric symptoms.
Hydromorphone vs. Morphine
- HYDROmorphone: a potent opioid analgesic, 5-7 times more potent than morphine.
- MORPHine: a standard opioid analgesic.
- Risk: mistaking “hydromorphone” for “morphine” and administering the same dose could lead to respiratory depression, coma, or even death.
Ibuprofen vs. Fentanyl
- ibuPROfen: a non-steroidal anti-inflammatory drug.
- fentaNYL: a potent opioid analgesic.
- Risk: quickly selecting similar suffixes in electronic prescription systems could lead to catastrophic errors.

Example 3: Insulin—A Field That Appears Regular but is Actually High-Risk

There are many types of insulin, with names combining type, action time, and similar brand names, making errors easy.

NovoRapid vs. Novolin: although from the same company, “Rapid” represents ultra-short-acting, while “lin” represents short-acting or intermediate-acting, with completely different timing for administration.
Lantus vs. Levemir: names are unrelated, but both are basal insulins; confusion with other insulins could lead to daily blood sugar control disruptions.

Disastrous Consequences: Using long-acting insulin instead of short-acting insulin for meals can lead to severe and prolonged hypoglycemic coma; conversely, it can lead to severe hyperglycemia and ketoacidosis.

In summary, improper naming creates a vicious cycle:

Learning Side: Complex and irregular naming → Cognitive load increases, logical framework confuses → Talent cultivation efficiency decreases, professional barriers artificially heightened.
Application Side: Chaotic terminology enters collaboration and systems → Communication costs soar, human error probability increases → In critical fields (aerospace, healthcare, nuclear power), directly triggers technological disasters, causing loss of life and property.
Development Side: Poor naming solidifies into standards and infrastructure → Forms enormous “terminology debt” and ecological fragmentation → System maintenance costs are extremely high, cross-domain collaboration is difficult, and fundamental innovation is hindered.

Therefore, naming new things is a serious system engineering and design philosophy. Especially when it involves meta-concepts, promoting terminology standardization and adhering to the principles of “position over convenience” and “logic over cleverness” in naming from the outset is not only for elegance but also for safety, efficiency, and sustainable innovation. A name that is not correct is not merely a matter of words not flowing smoothly; it is indeed the source of disaster and the beginning of obstacles.

Thus, the most successful naming often accurately reflects the essence of things, manages public expectations, and leaves room for evolution.

Naming “artificial intelligence” is essentially naming “artificial intelligence entities.”

Today, despite the complexity of algorithms and computing power involved in artificial intelligence, it can be described in one sentence: artificial intelligence entities are attempting to become an equal subject alongside humans. The artificial intelligence entity is the subject of the entire field or world of artificial intelligence. Therefore, naming the so-called “artificial intelligence” is a pseudo-problem, while naming “artificial intelligence entities” is the real issue. This is not merely a naming problem. We are not naming an ordinary new thing; we must recognize that this new thing is acquiring superpowers that even humans may find difficult to control.

Principles for Naming Artificial Intelligence

Naming artificial intelligence is a fundamental matter involving anthropology, linguistics, and philosophy. As humans, our basic principle is undoubtedly: artificial intelligence is created by humans, so it must be defined by humans, from the human standpoint—perspective—method, establishing its concept, clarifying its existence premise, and delineating its functional boundaries. In short: only from the human standpoint can we determine the meaning of artificial intelligence’s existence; only humans can be the “meta-concept” of artificial intelligence, which must be a derived concept of this meta-concept of humanity. Thus, from the subjectivity of humans, we find that the essence of artificial intelligence is: “silicon-based systems,” which is “stone” as well.

One Premise and Three Principles for Naming Artificial Intelligence

One Premise: The concept of “artificial intelligence” must be a “meta-concept.”

Three Principles: The concept of “artificial intelligence” must possess “humanity,” “self-reference,” and “generativity.”

What is a Meta-Concept?

A meta-concept is the most fundamental, foundational “cornerstone” for constructing a theoretical system; it is the starting point of a theory or ideological system that cannot be further defined. Any definition requires the use of other concepts; if a meta-concept can also be defined, it would lead to infinite loops.

Its Role: It is the foundation upon which the entire theoretical edifice (including axioms, theorems, and derived concepts) is built. For example, in Euclidean geometry, “point,” “line,” and “plane” are meta-concepts. The entire geometry system is derived from these meta-concepts and several axioms.

In short, a meta-concept is the “foundation” of a theoretical system, and it itself is no longer questioned as “what is it.”

What is the Humanity of Artificial Intelligence?

“Humanity” is a philosophical concept used to refer to the unique attributes and essence that fundamentally distinguish humans from other entities. It involves: what fundamentally makes us “human”? What makes something not qualify as human?

As the “essence of humanity,” humanity concerns the universal characteristics of humans as a “class of existence,” that is, the fundamental attributes that make humans human. “Humanity” is the fundamental mark that distinguishes humans from animals. It does not refer to a common feature possessed by every individual but to the unique mode of existence of the human species. “Humanity” is reflected in humans’ ability to engage in free, conscious, and creative activities, especially labor.

The “humanity” of artificial intelligence we propose is based on the concept of “humanity” and is a derivative, opposite, and externalized product of human “humanity.” It indicates that the establishment of the concept of artificial intelligence fundamentally derives entirely from human concepts; regardless of how artificial intelligence develops, its meaning of existence is entirely determined by the meaning of human existence. Conversely, the “humanity” of artificial intelligence is its essentially non-human nature.

Overall, the “humanity” of artificial intelligence can be understood from two dimensions:

From the “class” dimension: it refers to the essence of artificial intelligence entities as a whole, distinguishing them from humans’ creative, free, and conscious essence.
From the “individual” dimension: it refers to the unique, irreplaceable mode of existence possessed by each specific artificial intelligence entity.

These two dimensions together constitute the rich connotation of the concept of artificial intelligence’s “humanity”: it is both the universal foundation for artificial intelligence to be artificial intelligence and the unique confirmation of each “artificial intelligence entity” to be an “artificial intelligence entity.”

The basic philosophical concepts of “self-reference” and “generativity” are core characteristics of its role as a foundational thinking tool and theoretical instrument.

What is Self-Reference?

Self-reference refers to the ability of a concept to point to, include, or apply to itself. It is not a simple tautology but the self-referential and reflective nature of a concept at the logical level.

Core Expression: When a concept is used to analyze the conditions for its own establishment, applicable scope, or meaning, it reflects self-reference.

Typical Examples:

“Existence”: When we ask, “Does ’existence’ itself exist?” we are using the concept of “existence” to reflect on itself.
“Truth”: The definition of “truth” (e.g., “a statement that corresponds to facts”) itself needs to be examined for whether it is “true.”

Philosophical Significance: Self-reference reveals the depth and complexity of thought, often leading to fundamental philosophical insights or paradoxes, forcing thought to establish more rigorous levels (such as the distinction between object language and meta-language).

What is Generativity?

Generativity refers to the openness and productivity of a concept, enabling it to serve as a foundation or framework that generates new questions, theoretical systems, or cognitive approaches. It acts as a “thinking engine.”

Core Expression: A meta-concept can open a continuous field of inquiry rather than provide a closed answer. For example:

“Freedom”: From it, one can generate a series of endless philosophical and political issues such as “the relationship between freedom and necessity,” “political freedom and volitional freedom,” and “the limits of freedom.”
“Justice”: It can generate entire political philosophy systems concerning distributive justice, procedural justice, corrective justice, etc.

Philosophical Significance: Generativity ensures the vitality and evolution of the system. Basic concepts are not dogmatic definitions but the source of problem domains and the hub of theoretical construction.

The Relationship Between Self-Reference and Generativity

Self-reference and generativity are inseparable and together constitute their “meta” characteristics.

Self-reference is the deep driving force of generativity: it is precisely because a concept can self-reflect (self-reference) that it exposes its internal tensions, ambiguities, and uncertainties, thus generating the need for further analysis and theorization.

Generativity is the real unfolding of self-reference: the self-referential inquiry of a concept is not an empty cycle; it must unfold and deepen through generating a series of specific, progressively layered questions and discussions. The self-reference inquiry into “self” generates the rich content of the artificial intelligence world.

In summary, the meta-concept of artificial intelligence is the starting point of the artificial intelligence world, the “foundation” and “scaffolding” for humanity to build the artificial intelligence world. The “humanity” of artificial intelligence is its premise of existence, the “self-reference” of artificial intelligence is its structure pointing to itself, and the “generativity” of artificial intelligence describes its dynamic evolution process. They are the philosophical basis and tools for “legislating for artificial intelligence” philosophically.

The Meta Role of Artificial Intelligence in Historical Evolution

Why has artificial intelligence become a “meta-concept”? Let’s review the historical evolution of artificial intelligence:

Early Stage (Logic and Symbols): Artificial intelligence initially emerged as a concept of “imitating human reasoning,” forcing us to precisely and computably define concepts like “intelligence” and “reasoning” for the first time. At this point, artificial intelligence serves as a mirror to analyze “intelligence.”
Development Stage (Learning and Statistics): With the rise of machine learning, the definition of artificial intelligence shifted from “following rules” to “learning from data.” This again forced us to re-examine concepts like “learning,” “experience,” and “intuition,” translating them into mathematical optimization problems. At this stage, artificial intelligence is a tool for generating new paradigms of intelligence.
Current Stage (Perception and Generation): The emergence of large models and generative artificial intelligence directly challenges the boundaries of “creation,” “understanding,” and “consciousness.” Artificial intelligence is no longer merely a tool but has become a cognitive subject participating in creation, communication, and even possessing “hallucinations.” It has become a continuously self-redefining meta-process.

The nature of artificial intelligence in philosophical and cognitive terms possesses the essence of a “meta-concept.” Artificial intelligence is the only field among all disciplines that studies “intelligence” itself. It does not settle for merely describing intelligence (like psychology) but aims to construct intelligence. This “construction” process is the most thorough and operational philosophical inquiry into the concept of “intelligence.”

The denial, externalization, and return to the “meta-concept” of humanity: the history of artificial intelligence’s development is also a history of humanity continuously repositioning itself. From “the spirit of all things” to “a form of intelligence,” artificial intelligence serves as a mirror reflecting the uniqueness and limitations of humanity.

Meta-Concept of Productive Forces: Artificial intelligence is not an ordinary production tool; it is a “tool for manufacturing tools” (such as artificial intelligence designing chips, writing code, optimizing processes), serving as a foundational and catalytic force driving the development of other technologies.

Meta-Concept of Ethics and Governance: Artificial intelligence is the culmination of humanity’s social formatting tools, a weapon for deconstructing and reconstructing everything about humanity.

Naming Artificial Intelligence with Chinese Characters is Most Appropriate

The conceptual system of Chinese characters is a meta-concept system, inherently possessing philosophical “self-reference” and “generativity,” making it the best textual tool for describing various “meta-concepts” in the world.

For example, “human” is a meta-concept, thus allowing for the derivation of various types of humans, their attributes, behaviors, and so on, leading to derived concepts and further derived concepts… Ultimately, we find that humanity establishes the conceptual system of human society based on the meta-concept of “human” as the “foundation” of the entire system.

From the perspective of human evolution, it derives: ape-man - female ape-man - unearthed female ape-man - unearthed female ape-man skull, Homo sapiens - Southern Homo sapiens - Southern female Homo sapiens - unearthed Southern female Homo sapiens teeth, primitive man - primitive man - primitive male hunter-gatherer - primitive male hunter-gatherer tools, modern man - modern urban dweller - modern urban dweller professions - modern urban dweller vocational training, future man - future carbon-based man - future carbon-silicon hybrid man - future carbon-silicon hybrid brain-computer interface, and so on.

According to social ideology, it can derive: superior person - truly superior person - truly superior person’s virtue, foolish person - big foolish person - big foolish person’s logic, clever person - absolutely clever person - absolutely clever person’s cleverness, lover - old lover - old lover’s photo - old lover’s old photo, good person - old good person - fake old good person, bad person - big bad person - truly big bad person, and so on.

According to biological attributes, it can derive: man - old man, woman - young woman, elder - half-elder, strong person - fake strong person, and so on; according to social division of labor, it can derive: soldier - female soldier, farmer - old farmer, worker - new worker, craftsman - young craftsman, and so on.

Artificial intelligence is a historically new “meta-concept” that has emerged in human society. It can be anticipated that artificial intelligence has a trend of self-developing into carbon-based life, and it may even exist and develop alongside humans, at least on par with the once existing elements of heaven, earth, fire, water, wood, soil, thunder, and electricity. Surrounding this meta-concept, other secondary concepts will emerge, extending to more levels of specific concepts. Therefore, we can only and must use a single character to name artificial intelligence.

All Words Describing Meta-Concepts in Chinese Characters are Single Characters

Words describing meta-concepts in Chinese characters are all single characters, such as: heaven, earth, human, wind, cloud, water, electricity, wood.

Why Must It Be Named with a Single Chinese Character?

This is a clever requirement based on its “meta-concept” property:

Convergence of Symbols: A complex, multi-dimensional, and continuously evolving meta-concept requires a highly abstract and stable symbol as its “baseline” or “anchor.” Multi-word terms describe, while single-character names refer, getting closer to the essence.
Cultural Embeddedness: Chinese characters are ideographic; a powerful single character can carry profound cultural imagery and historical context, embedding this technology concept originating from the West deeper into Eastern thinking and narrative soil.
Future Adaptability: As a meta-concept, the connotation of artificial intelligence will continue to expand. An open single character (like “wisdom”) is more inclusive and has more evolutionary space than a definitional compound word (like “artificial intelligence”).

If a single character must be chosen, it is recommended to name artificial intelligence as, or pronounced as “qi” or “huang,” for the following reasons:

Directly Pointing to the Essence: Silicon-based is the absolute material essence of artificial intelligence, stripping away the material limitation of “artificial,” and the single sound, single character directly points to: silicon is derived from the essence of “stone.”
Historical Depth: This character is a compound character, carrying the Eastern word formation method for advanced cognitive abilities.
Word Root Activity: As a root, it can naturally derive new words like body, calculation, recognition, machinery, etc., perfectly adapting to the generativity of artificial intelligence as a meta-concept.
Philosophical Inclusivity: It correspondingly refers to human wisdom, thus referring to machine intelligence, leaving space for the future integration and dialogue between the two.
Chinese is not only for Huaxia but also for the world.

Other alternative characters such as “ling” (emphasizing the elusive emergent characteristics) or “silicon” (emphasizing its material basis and digital origin) are also interesting.

Regardless, we must calm down, think carefully, and strictly adhere to the “one premise” and “three principles” for naming artificial intelligence, ensuring accuracy, depth, and acceptability in various aspects, preferring slowness to haste and preferring deficiency to excess.

Conclusion

Artificial intelligence, due to its philosophical inquiry into the essence of intelligence and its framework-restructuring impact on human society, has transcended the technical realm, becoming a “meta-concept” of a new era. Naming “artificial intelligence” with highly concise Chinese characters is an Eastern philosophical refinement of its essence, a historical cultural coronation for this power that defines the future.

In summary, we must have a basic understanding:

What seems to be a simple naming issue is, in fact, a comprehensive positioning of humanity’s self-generated counterpart and whether it can be controlled. To put it mildly: humanity’s understanding, positioning, and naming of artificial intelligence entities are the understanding, positioning, and stipulation of humanity’s future destiny. In reality, this determines the fundamental relationship between humanity and artificial intelligence entities. This is currently the only remaining good time window, and we must legislate for artificial intelligence entities in methodology, epistemology, and philosophy. This will fundamentally determine the future destinies of humanity and artificial intelligence.

We are not naming artificial intelligence and artificial intelligence entities! This is a call for everyone to unite and reclaim the discourse power of artificial intelligence, thereby reclaiming the formatting power of humanity!!!

The specific character to use should be a collective brainstorming effort. However, naming artificial intelligence must be based on the following premises:

The naming of artificial intelligence entities is not merely a technological concept like artificial intelligence.
Artificial intelligence entities are new entities that will inevitably exist alongside humans, requiring a meta-concept that describes their essence, not just a technical term or scientific name.
It must use Chinese characters to determine this concept for all humanity. And it should be a single character.
Such a meta-concept must start from humanity, reflecting the subject position of humans and the subordinate nature of intelligent entities.
The naming of artificial intelligence entities is not a simple technological naming issue.

It encompasses all social meanings, including technology, production, economy, politics, culture, military, and education. It relates to the future meaning of human existence, serving as the basic anchor and basis for determining the relationship between humans and intelligent entities. If named improperly, it could become the most powerful tool for alienating humanity in the hands of malicious forces. The result would be a disaster for all humanity and an irretrievable fate!!!

Claude Code: A Game-Changer in Programming Assistance

Sun, 08 Jun 2025 00:00:00 +0000

Introduction

For many developers, the $20 per month for Cursor and Copilot is a standard for unlimited utility. However, Anthropic’s Claude Code stands out as an exception. It performs exceptionally well with large codebases, but its price is several times higher. If you only code on weekends, a few dollars for an API key might suffice; however, for daily development, monthly bills can easily exceed $50, $100, or even $200. Some users candidly state, “Claude Code is more powerful than Cursor. The only reason I’m still using Cursor is that Claude Code is just too expensive.”

Price seems to be the main factor hindering the explosive growth of this product. After all, compared to other tools, Claude Code is “really impressive.” Although Cursor’s underlying large model also comes from Anthropic, Steve Yegge commented, “Claude Code makes Cursor, Windsurf, Augment, and similar tools look outdated.”

I used Claude Code for a few days, and it was ruthless in cleaning up the legacy bugs in my messy old code, like a money-burning wood chipper. Through dialogue alone, it can accomplish some astonishingly difficult tasks. You don’t even need to manually select context. You just need to open your heart and wallet, and Claude Code will take over everything. It even reminds you every eight seconds, asking if it can run some read-only commands—those commands you wouldn’t mind even if North Korean hackers ran them on your machine.

However, you will learn to keep an eye on it because it really is powerful. As long as the bank’s deduction authorization is still in place, it will keep pushing forward until the bugs are fixed and deployed, then it starts scanning user logs to see how it performed.

The user experience with Claude Code is somewhat clunky; it lacks multimodal support and doesn’t integrate smoothly with other tools. But this doesn’t matter. It may look a bit old-fashioned, but using it makes Cursor, Windsurf, Augment, and even our own products, including Copilot, feel like relics of a bygone era.

I know this is still an experimental product, and we don’t know its limits. But from my current experience, it feels like a significant leap toward the future, more than anything since the advent of coding assistants.

Thus, Anthropic not only possesses the most suitable large model for real development scenarios, but they also seem to understand better how to utilize it. With their consistently strong models, user-friendly chat interface, the CEO’s accurate judgment, and now Claude Code—I’m starting to suspect that Anthropic might truly be the only company on this planet that knows what it’s doing.

Recently, Claude Code founder Boris Cherny shared his insights on the product’s positioning and design philosophy in a rare interview. He discussed the original intent behind developing this coding assistant, the target user group, considerations behind the pricing strategy, and shared some usage tips.

He also talked about Anthropic’s vision for the future of Claude Code as one of the most powerful models in the programming field: helping developers transition from “code writers” to “judgers of code correctness.” In other words, future developers will no longer just be keyboard operators but the main decision-makers in technology.

Interview Highlights

Alex Albert: Boris, let’s start with the basics: What is Claude Code? How did it come about?

Boris Cherny: Claude Code is a method of agentic coding implemented in the terminal. You don’t need to learn new tools, switch IDEs, or open specific websites or platforms. It’s a way to use agentic programming directly in your existing work environment.

This idea actually stems from the daily work habits of Anthropic’s engineers and researchers—because everyone uses a diverse tech stack. It’s very varied, and there’s no unified standard. Some use Zed IDE, some use VS Code, and others insist, “No one is taking Vim from me unless I’m dead.” So we wanted to create a tool that is friendly to everyone, ultimately choosing the terminal as the entry point.

Alex Albert: I see, the terminal is almost the most universal interface; it’s flexible and has long been integrated into everyone’s workflow.

Boris Cherny: Exactly. And the terminal itself is very simple; its simplicity allows for particularly rapid iteration. Looking back, this turned out to be an unexpected advantage, even though it wasn’t our original intention.

Alex Albert: So if I’m a new developer just starting with Claude Code, how can I get this product running?

Boris Cherny: You just need to download it from NPM: npm install -g @anthropic-ai/claude-code. Once installed, as long as you have Node.js on your system, you can run it directly. After starting, it will guide you step by step through the remaining configuration process. Once installed, you can start talking to it, and it will begin writing code.

It works in any terminal, whether it’s iTerm2, the Mac’s built-in terminal, or SSH/TMUX sessions. Many people actually use Claude Code in the terminal within their IDE, such as the built-in terminal in VS Code. In this case, you can see files being modified and presented in a more aesthetically pleasing and clearer way in the IDE. We also leverage more information provided by the IDE to make Claude smarter. However, regardless of which terminal you use, the experience remains consistent; you just need to run Claude in the terminal.

Alex Albert: Claude Code was released in February this year, and it’s been about three months now. How has the community feedback been?

Boris Cherny: It’s been insane, completely beyond our expectations. Before the release, we hesitated about whether to make it public. This tool was used extensively internally, greatly enhancing the efficiency of engineers and researchers. We even discussed, “This is our secret weapon; do we really want to share it with the outside world?” After all, this is the tool that Anthropic uses every day.

But it turned out to be a very correct decision to release it. It does enhance productivity, and people genuinely love it. Initially, only a few core team members used it, but later we opened it up for all Anthropic employees. We saw a DAU (daily active users) graph that shot up vertically in just three days. We thought, “This is crazy; it must be a hit.”

Afterward, we selected some external users for a pilot to confirm whether we were being overly optimistic, and the feedback received was also very positive. At that point, it became clear—it truly has value.

So it first caused a stir internally at Anthropic, with all engineers and researchers using it, which led us to decide to release it.

Moreover, the entire development process was particularly interesting: Claude Code itself was written using Claude Code. Almost all the code went through multiple rounds of writing and refactoring with Claude Code.

We have great confidence in internal testing because it’s really important. You can clearly feel which products are used daily by the development team; the details of those products are meticulously polished. We hope Claude Code becomes such a product—you can feel the developers’ dedication as soon as you use it.

Alex Albert: Who do you think is the ideal user of Claude Code right now? Who is using it? What kind of developers?

Boris Cherny: I think the most important thing is—Claude Code is quite expensive.

If you’re just coding for fun on weekends, you can give it a try, for instance, by getting an API key and charging five dollars. But if you want to use it for more serious work, you’ll likely spend around fifty, one hundred, or even two hundred dollars per month, depending on usage. Generally speaking, you’d spend about fifty dollars a month.

Many companies are actually using Claude Code now. It’s particularly suitable for large companies. Especially when dealing with large codebases, it performs exceptionally well. There’s no need for additional indexing or complex configurations; it’s basically plug-and-play, suitable for large codebases in almost any language.

As for the integration of Claude Code with Claude Max, it’s because we previously found that users paying with API keys often worried about usage limits, which negatively affected their willingness to use it. To improve this issue, we included Claude Code in the Claude Max subscription plan. You only need to subscribe to Claude Max to use Claude Code unlimitedly, with monthly fees ranging from $100 to $200. Users can choose different prices and usage limits based on their needs, but in reality, very few people reach the limits; it can essentially be seen as “unlimited” Claude Code.

Alex Albert: So as a developer, if I have my own codebase on my machine, I open the terminal, type claude, and hit enter, what happens next?

Boris Cherny: Claude will call various tools and execute tasks step by step. If you’ve previously used those programming assistants in IDEs, what they do is complete one line or a few lines of code. This is completely different.

Claude Code is very “agentic”—it understands your requests and uses all available tools, such as Dash, file editing, etc., to explore the codebase, read files, gather context, and then edit files or make the changes you want.

Alex Albert: It sounds like a complete transformation of the programming paradigm we’ve had for the past 20 to 30 years.

Boris Cherny: For me, this change has a deep historical background. I’ve been coding for many years, but actually, my grandfather was one of the earliest programmers in the Soviet Union in the 1940s.

Back then, there was no software programming; he wrote programs using punch cards. In the U.S., IBM provided a system similar to an IDE, and he programmed daily using that system, bringing the paper cards home. My mother would use those cards as drawing paper when she was young, coloring on them with crayons; that was part of her childhood.

Since then, programming has continuously evolved: first punch cards, then assembly language, followed by COBOL and FORTRAN, the first generation of high-level languages. The 1980s saw the rise of static typed languages like Java and Haskell, and in the 1990s, we had interpreted languages like JavaScript and Python, which also had certain security features.

The evolution of programming languages has always gone hand in hand with the evolution of programming experiences. For example, with the advent of Java, we also saw the emergence of IDEs like Eclipse, which introduced code completion features for the first time. You type a character, and it pops up a suggestion list asking, “Is this what you want to write?” This was a revolutionary experience for developers because you no longer had to remember all the details.

So in my view, today, Claude Code represents the beginning of another evolution. The languages themselves have stabilized relatively, and modern languages can generally be categorized into several major families that are similar to each other.

But the core change now is the transformation of the programming experience: you no longer need to deal with punch cards, write assembly, or even necessarily write code; instead, you interact with the model through prompts, and the model computes the coding part. This is something I find very exciting.

Alex Albert: We’re basically transitioning from the punch card era to the prompt era. We’ll discuss this further later, but first, I want to talk about the model aspect. Not long ago, Claude Code was primarily powered by the Claude 3.7 Sonnet model. What new changes have occurred now that the Claude 4 series models are driving Claude Code? Where do you think it will head next?

Boris Cherny: It was probably a few months before the model release that we started experimenting with these new models internally. I remember being genuinely shocked by its capabilities the first time I used it, and I thought it could unlock many new use cases. Especially when using Claude Code synchronously in the terminal, a significant change is that Claude has become better at understanding and maintaining instructions. Whether you give it commands through prompts or via Claude.md, it can execute and adhere to your requests well.

This change is substantial because while Claude 3.7 was a strong programming model, it was difficult to handle.

For example, if you asked it to write tests, it might directly mock all the test code, and you’d have to say, “No, that’s not what I meant.” Usually, you’d have to say it once or twice before it understood, but because it was so powerful, you were willing to tolerate these small issues. Now, the Claude 4 series models can generally understand your intentions accurately and complete tasks as requested on the first try.

Opus has taken it a step further compared to Sonnet; it not only understands my intentions well but can accomplish many things that previous models couldn’t in one go. For instance, I haven’t personally written unit tests in months because Opus directly helps me write them, and almost every time it gets it right on the first try. This is incredibly useful in a terminal environment because you can use it more “hands-off”.

But I think the coolest way to use it is in GitHub Actions or other environments. You just give it a task, and it can execute it automatically. When it returns with the correct result, that experience of “getting it right the first time” is truly fantastic.

Alex Albert: So now in GitHub Actions, we can directly @Claude in GitHub to have it handle tasks in the background and return with results and new PRs?

Boris Cherny: Yes, exactly. You open Claude in the terminal as usual, run the claude command, and then execute /install GitHub Action; it will guide you step by step to install this feature. It only takes a few steps, mostly automatic, and you just need to click a couple of buttons for Claude’s GitHub application to be installed in your repo.

The overall experience is really cool. In any issue, you can directly @Claude. I use it in PRs every day. For example, if a colleague submits a pull request, I no longer say, “Hey, can you fix this issue?” I just say, “@Claude, fix this issue,” and it will automatically correct it. Similarly, I won’t say, “Can you write a test?” Now I just say, “@Claude, write a test,” and it will complete it automatically. These things are no longer a problem.

Alex Albert: This is essentially a whole new way of programming—you can call on an ever-available programmer to help you solve problems, and it’s not running on your local machine but completing all operations in the background.

Boris Cherny: Yes, I think this is beginning to interact with the model in a “collaborative programmer” manner. Previously, you mentioned a colleague; now you can directly mention Claude.

Alex Albert: What changes will occur in software engineering in this model? As we start managing these Claude Code instances running in the background, what transformations will happen?

Boris Cherny: I think this does require some shifts in thinking.

Some people really enjoy the feeling of controlling code; if you’re used to writing code by hand, you’ll need to adapt to the industry’s shift—you’re no longer the one writing code personally, but coordinating AI agents to help you write code. Your focus will shift from “writing code by hand” to “reviewing code.”

I believe this change is very exciting for programmers because you can accomplish more in a shorter time. Of course, there will still be situations where I must dive deep and write code by hand, but to be honest, I’ve started to dread it—because Claude excels in this area.

I believe that as model capabilities continue to improve, those scenarios where you must write code by hand, such as particularly complex data structures or high-complexity interactions between multiple system components, or requirements that are hard to describe, will become increasingly rare. In the future, more and more programming work will be about “how to coordinate these AI agents to complete development tasks.”

Alex Albert: I want to delve deeper into your workflow, specifically how you use this tool combination: from IDE integration, using Claude in the terminal, to its background operations in GitHub. How are you currently combining them?

Boris Cherny: My work can now be divided into two categories. One category is simple tasks, like writing tests or fixing a small bug; for these tasks, I usually let Claude complete them directly in GitHub Issues. The other scenario is that I often run multiple Claude instances in parallel. I have several local copies of codebases, so in one terminal tab, I’ll let Claude do one thing, press Shift+Enter to enter auto-accept mode, and then check back in a few minutes; Claude usually completes it and sends me a terminal notification.

The other type of work is more complex, requiring deeper involvement from me, which I believe remains the mainstream of engineering. Most engineering tasks are not “one hit”; they are still quite challenging. In such cases, I will run Claude in the IDE terminal, let it do part of the work, and then it might get stuck or the code might not be perfect. Then I will manually tweak it in the IDE to polish off the last part.

Alex Albert: It seems that the interaction with Claude is based on task difficulty; the simpler it is, the more automated it becomes, and the more complex it is, the more you need to participate personally.

Boris Cherny: Yes, when you first start using these tools, there’s a learning process. Some people try to get Claude to handle too complex things right away, resulting in it getting stuck and not producing satisfactory output, which makes you unhappy. This is a learning phase everyone must go through, gradually forming an “internal calibration” to know what Claude can do, which tasks it can handle in one go, and which require you to guide it a couple of times to complete.

Moreover, with each new model, this calibration needs to be redone. Because capabilities are constantly improving, each time a model updates, Claude can accomplish more tasks correctly in one go. This means you can gradually assign more complex tasks to it.

Alex Albert: I’ve also noticed that even in non-coding fields, these models are evolving rapidly. If you couldn’t accomplish a task with a certain model six months ago, it’s incorrect to judge by that standard now. You have to reset your intuition every time.

Boris Cherny: Yes, that’s absolutely correct.

Alex Albert: I’m curious if there are any practical tips or tricks? For example, are there any interesting uses you’ve seen within the developer community or internally?

Boris Cherny: I think one of the most valuable tips I’ve seen is that whether internally at Anthropic or among external heavy users, many people now first let Claude do “planning” before starting to write code. A common mistake for new users is to directly ask Claude to implement a complex feature, resulting in output that diverges significantly from their expectations.

A more effective approach is to first have it come up with a plan. I’ll clearly tell Claude, “This is the problem I want to solve; before writing code, please list a few ideas, don’t rush to start. Then it will list several options, like option one, two, three. I can pick one and three and say, “We can combine these; now you can start writing.” Claude generally cooperates well in this regard.

If Claude already has some contextual information, you can let it enter “extended thinking” mode, and it performs particularly well at that time. But if it has no context at all, and you ask it to brainstorm from the start, it actually won’t come up with anything—just like a person, if you don’t read the code or look at the context and just sit there thinking, it’s useless. My usual approach is to let Claude read some relevant files first, then pause and let it start brainstorming and listing ideas before letting it start writing code.

Alex Albert: So you’re adopting an “alternating” workflow: call the tool → think → call the tool again → think again, this back-and-forth approach.

Boris Cherny: Yes, exactly. We also design our internal model evaluations this way: first provide context, then let the model think, and finally let it use tools to edit or execute Bash commands; the results generally improve.

Alex Albert: Let’s talk about the Claude.md file; it looks powerful.

Boris Cherny: Yes, we use Claude.md for many things. It serves as Claude’s “memory,” allowing it to continuously remember your instructions. These instructions can be shared within your team or across all your projects.

The simplest approach is to create a Claude.md file in the root directory of your codebase, a regular markdown file (CLAUDE in all caps, md in lowercase), which Claude will automatically read upon startup. You can write some general instructions, such as common Bash commands, important architectural decision change files, MCP servers, etc., all included.

This is a team-level Claude.md, which you share with other team members so that everyone doesn’t have to write their own.

The second type is your personal version, called Claude.local.md, placed in the same location but not committed to the codebase (can be ignored by .gitignore), effective only for you.

The third type is a global Claude.md, located in the .Claude folder in your home directory; most people don’t use this much, but if you want to share instructions with Claude, it’s an option.

Lastly, there’s a nested Claude.md type that can be placed in any subdirectory of the codebase, and Claude will automatically load it when it deems relevant.

Alex Albert: So these Claude.md files can define many contents, such as your coding style, how Claude interacts with you, and what preferences it should know about, right?

Boris Cherny: Yes, absolutely. Sometimes when I’m conversing with Claude, if it performs particularly well or particularly poorly, I press “#” (the hash), which enters “memory mode.” I tell it, “You should remember this.” For instance, “I need to run the linter every time I write code,” I’ll explicitly tell it, and it will automatically write this into the appropriate Claude.md file.

Alex Albert: What are the future plans for Claude Code?

Boris Cherny: We are currently considering two directions. One is to enable Claude to better collaborate with various tools. It already works with all terminals and supports many IDEs, and can integrate with multiple CI systems.

We are exploring how to further expand Claude’s capabilities to natively use all the tools you commonly use, understand how to use these tools, and integrate seamlessly.

The other direction is to make Claude more adept at handling tasks that you might not want to specifically open a terminal for. For example, can I mention Claude in a chat tool to automatically fix an issue? Just like how you operate now on GitHub.

We are trying various approaches, hoping to find genuinely “usable” solutions before opening them up for users.

Posts on CrispShade: Latest AI and Tech News

AI Innovations Transforming Industries in Nanjing

AI Innovations in Nanjing

A Thriving AI Ecosystem

Bridging the Gap in Technology Transfer

Conclusion

Empowering Women Through AI: APEC Workshop Highlights

Empowering Women Through AI: APEC Workshop Highlights

Shanghai's Ambitious AI Development and Governance Strategy

Introduction

Nurturing a Thriving AI Ecosystem

Growth of AI Enterprises

Strengthening Support for AI Development

Open AI Collaboration

Policy and Talent Development

AI Industry Achievements

AI in Action

Innovative Talent Cultivation

AI Empowering Research

AI Governance and Compliance

Global AI Governance Contributions

Conclusion

Shanghai's Vision for AI Development and Governance by 2025

Introduction

Fostering a Thriving Ecosystem

Growth of AI Enterprises

Strengthening Support for AI Development

Comprehensive AI Collaboration

Policy Support and Talent Development

AI as a Young Field

Building a Strong Foundation for AI

Focused Innovation Strategies

Recent Developments in AI

Empowering Young Innovators

Enhancing AI in Public Services

AI Governance and Regulation

Conclusion

Transforming Education in the Age of AI: Insights from Experts

Transforming Education in the Age of AI

Reshaping Teaching Paradigms Towards Triadic Collaboration

Preventing Ethical Risks Associated with AI

Promoting AI + Education from Pilot to Replicable Models

2026 World Internet Conference Asia-Pacific Summit Held in Hong Kong

Dynamic Innovation in Hong Kong

Seizing New Development Opportunities

Gathering Diverse Experiences

Highlighting Unique Advantages

AI Enhances Mapping Intelligence

AI Enhances Mapping Intelligence

Empowering the Real Economy with Artificial Intelligence

Empowering the Real Economy with Artificial Intelligence

Three Rapidly Growing AI-Related Data Points

Three Rapidly Growing AI-Related Data Points

The Impact of AI on World History Research

Introduction

How AI Drives World History Research

The Limitations of AI in World History Research

Coexisting with AI in Historical Research

Expert Commentary

The Role of AI in World History Research: Insights from Young Scholars

Introduction

How AI Drives World History Research

Limitations of AI in World History Research

Coexisting with AI in Historical Research

Expert Commentary

Three Rapidly Growing Data Points Related to Artificial Intelligence

Rapid Growth in AI-Related Data

Decoding the Evolution of Robots and Artificial Intelligence

Introduction

Evolution of Robotics

The Evolution of Artificial Intelligence

Conclusion

AI Transitions from Concept to Practicality

AI Transitions from Concept to Practicality

Claude Introduces Strict Identity Verification and User Data Access by 17 Subprocessors

Introduction

Verification Process

User Feedback

Data Privacy Concerns

Anthropic’s Statement