Author: lethuphuong

📌 AI agents open up powerful automation capabilities but simultaneously create a new layer of technical debt more complex than previous microservices. With 7 infrastructure blocks from integrations to orchestration, enterprises may have to spend up to 50% of resources just to control the agent system. Without building a foundation early, risks such as data leaks, production errors, and uncontrollable AI costs will emerge rapidly as the number of agents grows to many times the number of personnel.

📌 Generative AI is not just a productivity tool but also changes the structure of work: increasing speed while lengthening working hours, expanding responsibilities, and increasing cognitive pressure. From UC Berkeley research to historical examples like email, the general trend is that as productivity rises, the feeling of overload increases accordingly. Without proper job design and AI usage discipline, a 10x gain in speed could be traded for a significantly higher risk of burnout.

📌 Cowork marks AI’s transition from a tool for programmers to an assistant for the entire workforce. By reaching 95% of employees and growing faster than Claude Code ($2 billion/year), Anthropic is aggressively expanding its market. However, legal risks and operational errors show that the increasing speed of AI development also comes with major challenges in governance and global competition.

📌 Physical AI is ushering in the era of “dark factories” where machines operate autonomously in mines, construction sites, and battlefields. With a valuation of $15 billion and applications ranging from self-driving trucks to military drones, this technology reduces labor dependency, especially as the transport industry faces a shortage of 1.2 million drivers. The military drone market is projected to reach $260 billion by 2035, showing that AI is not only changing production but also reshaping the global economy and defense.

📌 The battle between the federal government and US states over AI regulation is escalating, with over 100 state laws despite opposition from the central government. While the White House prioritizes global competition, states focus on protecting citizens from AI risks. This lack of consensus could shape how the US controls AI in the future, while significantly impacting innovation, safety, and privacy in the new technological era.

📌 Singapore is proactively preparing its workforce for the AI era with plans to train 10,000 personnel and support 10,000 businesses. Instead of only measuring efficiency through technology, the country focuses on jobs and income to ensure sustainable growth. However, challenges regarding skill standards, costs, and AI safety remain significant, requiring a clear verification system for AI to truly deliver economic and social value.

📌 Medical AI is growing fast with tens of millions of queries per day, opening opportunities to improve healthcare access. However, current evidence is insufficient to ensure safety, particularly in diagnosis and emergency handling. Without large-scale independent testing, AI can be both a solution and a risk. The future depends on balancing rapid deployment with ensuring reliability in sensitive medical environments.

📌 China now accounts for 51% of the world’s top AI talent, surpassing the combined total of the US, Europe, and other regions (outside China), with its university system leading (9/10 top schools) while the US share fell from 20% to 12%, indicating a clear shift in talent power. Despite its technological strength, the US relies heavily on ethnic Chinese talent (35%). Notably, domestic talent retention and attraction have surged, with 68% staying in China and 28% returning home. If this momentum continues, by 2028, the number of AI experts in China could double those in the US, altering the…

📌 Leaders often mismeasure AI effectiveness by relying on usage frequency rather than quality and impact. Approximately 90% of employees have used AI frequently, but only about 5% achieve a sophisticated level of usage, proving that tool adoption does not equate to effectiveness. Power users tend to write long prompts, engage in multi-turn interactions, switch flexibly between models, and use AI at high frequencies but with clear objectives. The solution is to build “AI-first” standards, provide scenario-based hands-on training, and clearly define expectations for each job role.

📌 The new US AI legal framework not only focuses on child protection and scam prevention but also draws attention with a proposal to reduce legal liability for developers. This aims to foster innovation and attract investment but also sparks debate over AI control risks. With 7 main pillars and a drive toward unified federal law, the US is attempting to both accelerate technology and maintain its global leadership role.