Yanwen’s blog

7/15/2026

On July 13, 2026, Harici published an article reporting on an EY-Parthenon study of the costs of reducing Western dependence on China across strategic supply chains.

The study, originally reported by the Financial Times, estimates that reducing reliance on China could cost the United States, the eurozone, and the United Kingdom more than $23 trillion over the next quarter century.

Upon reading it, three questions came to my mind.

The first question is: Is it technically possible?

Almost certainly yes. There is no economic law that says manufacturing ecosystems cannot be built elsewhere.

The second question is: Is it economically expensive?

Again, yes, very much so. The study's estimate of $23.6 trillion illustrates just how costly such a transformation could be. Building factories is only part of the challenge. Entire industrial ecosystems—including suppliers, logistics, infrastructure, skilled workers, and refining capacity—take decades and more to develop.

The third question is: Is there sufficient political will?

That's a trillion dollar question. Democracies tend to struggle with long time horizons. The costs are immediate and visible—higher public spending, higher prices, and difficult budget choices—while the benefits, such as economic independence, may not become apparent for years or even decades.

History offers many examples in which what once seemed impossible became possible through sustained political will and collective efforts. One example is modern China. 

Before 1949, China was burdened by the "Three Mountains"—imperialism, feudalism, and bureaucratic capitalism. Overcoming these forces appeared an overwhelming task at the time. Yet history shows that objectives once considered unattainable can become achievable when there is sufficient determination, organization, and persistence.

Very often, the limiting factor is not capability but determination. Individuals, organizations, and nations frequently possess the resources and knowledge to achieve difficult goals. 

Like the book review that I did a few days ago on Why Was the PLA Able to Win? by Xu Yan. It's the same question asked again and again, why do some people and institutions successfully pursue long-term goals while others do not?

What ultimately determines success is whether they have the vision, the determination, the discipline, and the political will to sustain the effort over many years and decades.

A reader commented, “ 还有一个因素，现在的西方政客更像职业经理人，而不是政治家，很难制定超过任期的蓝图，更难保证不人走茶凉。” Make sense.

7/14/2026

A few weeks ago I read some research about the functions of human hippocampus. I made a note of the research and wanted to come back to this, because one of my college classmates lost about 90% of the volume of her hippocampus due to some unknown reason. This research suggests that the hippocampus plays a much broader role than neuroscientists once believed.

Neuroscientists have discovered that the hippocampus is involved not only in forming memories but also in imagining the future. Whether we plan tomorrow's activities, picture how an event might unfold, or visualize a place we have never seen, many of these mental simulations depend on the hippocampus. In other words, remembering the past and imagining the future rely on many of the same neural machinery.

Modern neuroscience increasingly suggests that the hippocampus is a simulation engine—a brain system that uses the past to construct possible futures.

From an evolutionary perspective, this makes perfect sense. Remembering the past is valuable because it helps us anticipate the future. Animals that can predict danger, locate food, or choose better actions are more likely to survive and reproduce.

Memory evolved not to preserve yesterday, but to prepare for tomorrow.

A growing body of research suggests that regular physical exercise, mentally stimulating activities such as reading and language learning, music, and rich social engagement are all associated with better hippocampal health and a lower risk of cognitive decline.

Finally, writing serves another valuable purpose. By recording our experiences, we create an external memory that works better at preserving our memory, and with that years later, these memories allow us to revisit the past, reflect on it, and use it to imagine the future once again. 

Perhaps that is one reason human beings have always kept diaries, written memoirs, and told stories. Memory allows us to learn from yesterday; writing allows us to preserve those memories long after they begin to fade.

7/13/2026

A few days ago a college classmate shared this with me, "University of Chicago Law to ban phones, laptops in classroom for first-year students in new AI plan."

I read the statement issued by the university. It is much more thoughtful than the news headlines suggest. The document focuses deeper questions than laptop or phone ban.

The university's AI strategy rests on three principles.

First, instead of banning AI or pretending it doesn't exist,  "our pedagogy and assessment should be designed to ensure our students learn how to think critically and solve legal problems with sound professional judgment."

AI can provide quick answers, but relying on those shortcuts too early may hinder intellectual development. As a result, classrooms, assignments, and assessments should be designed to promote genuine learning rather than dependence on AI.

Second, "AI may transform the legal profession, but many aspects of legal practice are likely to remain the domain of humans, not merely because humans are good at them, but because clients, employers, judges, and society will want humans to perform them."

Skills such as oral advocacy, strategic judgment, critical thinking, and building relationships with clients are likely to remain fundamentally human responsibilities. Legal education should therefore place renewed emphasis on developing these uniquely human capabilities. So they ask, "What are the essentially human skills that AI can't replace?"

Third, because AI will inevitably become part of legal practice, students must also learn how to use it effectively. However, the goal is not merely to master today's AI tools, which may soon become obsolete, but to develop the analytical thinking, adaptability, and ethical judgment needed to work with whatever technologies emerge in the future.

If one of America's top law schools believes that students must first learn to think before relying on AI, what does that imply for education more broadly?

In the age of AI, thinking becomes more valuable than knowing. Facts are now available almost instantly, but judgment, reasoning, creativity, and wisdom cannot simply be downloaded from a machine.

The human brain develops and maintains through active engagement. We learn by working with difficult problems, making mistakes, and thinking them through. If we rely on AI too early or too often, we risk outsourcing the very mental work that education is meant to develop. That concern applies not only to students but also to children, whose minds are still developing, and to older adults, who benefit from continued cognitive engagement to keep their brains sharp.

Finally, human behavior is shaped by the environments we create. Rather than simply telling students, "Don't overuse AI," the University of Chicago Law School is redesigning the learning environment so that students first develop the habit of independent thinking.

7/12/2026

I read the news that the European Union formally accused Meta on July 10, 2026, of violating the Digital Services Act by designing Facebook and Instagram in ways that the EU says encourage addictive use, particularly among teenagers.

According to the European Commission, two features are of particular concern: infinite scrolling and autoplay videos. The EU argues that these are nothing but psychological hooks that encourage users to keep scrolling almost automatically, putting their brains on "autopilot." The concern is especially great for children and teenagers, whose self-control and decision-making abilities are still developing.

The Commission also argues that Meta has not done enough to protect minors. It says parental controls are too complicated to use, can be easily bypassed by teenagers, and that stronger measures are needed to prevent children under 13 from creating accounts.

Meta disagrees. The company says it has already introduced "Teen Accounts" with built-in protections, including nighttime restrictions and daily screen-time limits, and says it will continue to work with regulators.

Whatever the legal outcome, the case raises a much larger question for all parents.

The issue is no longer simply how much time children spend online, but how digital platforms are intentionally designed to keep them there. Infinite scrolling and autoplay are designed to remove the natural stopping points that once existed when reading a book, watching television, or even browsing the early internet. Instead of asking ourselves, "Should I keep going?" we simply continue.

Perhaps the most significant aspect of this case is the precedent the EU hopes to establish. Rather than treating excessive screen time solely as a matter of individual self-control or parental responsibility, regulators are asking whether business models built around maximizing users' attention should also bear responsibility for the effects they have on children.

For parents, the lesson is clear. The most important question may no longer be whether children use social media, but whether adults understand how these platforms are designed to capture and hold a child's attention infinitely. Recognizing that design is the first step toward helping children use technology wisely rather than being used by it. 

Finally, people do not make choices in a vacuum; institutions and environments influence those choices. Our responsibility is to help children choose wisely.

7/11/2026

Two days ago I read these words by UN Secretary General Antonio Guterres:

"AI is impacting our children’s learning, their friendships, and how they view the world. And we are already seeing the cost:

Children deceived by machines posing as friends. Children steered toward self-harm. Children violated by abuse images made at the touch of a button. No child should be a guinea pig for unregulated AI."

Like all technologies before AI, television, smartphones, and social media, they all have the power to change human childhood. Also like all these technologies, there's no way that children can get away from it. Already children are using AI for homework, entertainment, and other activities.

Unlike earlier technologies, conversational AI can convincingly imitate human conversation and empathy. Younger children may therefore be more likely to perceive it as a real friend.

Childhood is a critical period for forming emotional attachments that lay the foundation for healthy relationships throughout life. If children spend too much time with AI companions without proper guidance, they may become emotionally dependent on them.

Young children may easily mistake an AI for a trustworthy friend. They often lack the maturity to distinguish good advice from poor advice or to recognize the limitations of AI. Once they begin to trust an AI companion, they may also be more willing to share highly personal information, putting their privacy and safety at risk.

The range of potential risks becomes much broader than many parents realize.

Ideally, companies design products with children's safety in mind. Profit never comes at the expense of child welfare. But we don't live in an ideal world.

I worry for parents everywhere, and that is not groundless. The risks are real, but as always, panic is not the answer.

What can parents do effectively? Here are a few practical things that they can try.

• Don't let AI replace human relationships.

• Use AI together with children rather than leaving them alone with it.

• Teach children that AI is a tool created by humans, not a friend.

• Carry one plenty of real life conversations—with parents, siblings, friends, grandparents, teachers.

• Set age-appropriate boundaries on AI use.

• Teach children to question what AI tells them, just as they would information from the internet.

• Make sure children know they should tell a trusted adult if an AI conversation makes them uncomfortable or frightened.

Parents cannot shield children from every new technology, but they must be present with the children and provide something no machine can: love, guidance, judgment, and genuine human companionship. Children who grow up surrounded by caring relationships are far less likely to seek emotional substitutes elsewhere.

Like what I've been writing about, human well-being depends on real relationships. AI adds a new dimension to that timeless truth. The technology is new, but our deepest human needs—for family, belonging, emotional attachment, and authentic connection—have not changed. With the arrival of AI, parents can ensure that no AI companion ever replaces the warmth, guidance, and security that only a family can provide.