What Should I Teach My Son? A Technologist's Deep Reflection

Tech entrepreneur Steve Krouse published an article titled "What shall I teach my son?" on his personal blog, sparking a heated discussion with over 300 comments on Hacker News. This article is not a simple parenting checklist, but rather an exploration—from the perspective of a tech professional and educational thinker—of what core capabilities and values parents should pass on to the next generation in an era of rapid technological iteration.

Why This Question Deserves Serious Consideration

Krouse's framing itself carries an engineer's characteristic prudence: not "what do I want to teach" but "what should I teach." The former concerns personal preference; the latter demands a judgment about the child's future needs. In a context where AI tools are increasingly ubiquitous and the cost of acquiring knowledge approaches zero, the value of traditional "knowledge transmission" is being reexamined. If the world the child will face offers answers at their fingertips, then perhaps the most valuable thing parents can give is not specific skills but rather a more fundamental way of thinking and an intrinsic drive.

This topic garnered 315 upvotes and 378 comments on Hacker News, indicating it struck a shared anxiety among the many parents in the tech community: we don't want our children shaped by the exam-oriented system, yet we worry that letting them chart their own course will leave them without a competitive edge.

Hacker News is a tech startup and programming community operated by the venture capital firm Y Combinator, with a user base primarily composed of engineers, entrepreneurs, and tech professionals. Its content ranking algorithm (based on votes and time decay) determines the visibility of popular discussions, so a post receiving hundreds of comments typically means it has touched on concerns shared by community members. The fact that Krouse's article generated discussion on this scale shows that "how to educate the next generation" has a strong resonance within the tech community—it's not merely a personal parenting topic.

Core Argument: Teach "How to Learn" Rather Than "What to Learn"

Judging from the arc of the article's discussion, Krouse leans toward shifting education's focus from content to method. In an environment of information overload, the ability to filter, verify, and integrate information matters more than memorizing facts. This means parents need to model how to ask good questions, how to face uncertainty, and how to adjust strategies after failure—capabilities that cannot be transmitted through lectures but can only be cultivated through shared practice.

A recurring argument in the comments is that children's most natural way of learning is imitation. How parents solve problems day to day, how they treat mistakes, how they allocate attention—all of this has far more influence than verbal preaching. This also explains why many parents with technical backgrounds deliberately let their children see them debugging code, reading documentation, and debating with peers.

"Metacognition" refers to an individual's awareness and regulation of their own cognitive processes, including planning learning paths, monitoring comprehension, and evaluating strategy effectiveness. In educational psychology, metacognitive ability is widely regarded as a key variable distinguishing effective learners from ineffective ones. What Krouse emphasizes—"how to ask good questions" and "how to adjust strategies after failure"—essentially falls within the metacognitive domain. Such abilities are difficult to acquire through direct instruction and typically require repeated practice in real tasks, supplemented by timely questioning and feedback from adults.

Balancing Technical Literacy with a Humanistic Foundation

Krouse's article does not fall into the utilitarian narrative of "everyone should learn to code." Many voices in the comments remind us that technology is merely a tool, and what truly determines a person's long-term development is judgment and values. When AI can generate code, write copy, and even make preliminary decisions, a person's unique value is increasingly reflected in asking the right questions, setting reasonable goals, and making trade-offs in ambiguous situations.

This raises a deeper question: if the child's future involves collaborating with AI rather than competing against it, then empathy, aesthetic judgment, and cross-domain associative thinking may be more worth cultivating than any single technical skill. The reason Krouse's question resonates is precisely that it forces readers to answer "why teach" before "what to teach."

The phrase "collaborating with AI rather than competing" involves a key distinction: AI excels at efficiently generating candidate solutions under clearly defined goals, while human value is more reflected in goal-setting, weighing constraints, and evaluating outcomes. This means education should focus on cultivating "evaluation" rather than "generation." For example, children need to learn to judge whether a piece of AI-generated code is safe or whether AI-written copy is appropriate—judgments that depend on domain knowledge, ethical awareness, and aesthetic experience rather than mere tool proficiency.

Consensus and Divergence Across Multiple Perspectives

In the Hacker News discussion, participants supplemented Krouse's thinking from various angles. Areas of consensus include: the importance of reading habits, giving children opportunities for self-directed projects, and space to tolerate boredom and failure. Divergence centers on the timing and intensity of technical education—some argue that early exposure to programming is essential, while others maintain that before adolescence, priority should be given to protecting children's imagination and physical activity.

These discussions offer no standard answer, but they present a healthy picture: educational decisions are inherently contextual, and there is no one-size-fits-all formula. The value of Krouse's article lies not in providing conclusions but in modeling an honest attitude toward educational questions—acknowledging uncertainty, then continuing to think.

Implications for Tech Professional Parents

For parents in the tech industry, this article offers an opportunity for self-calibration. We tend to overvalue technical skills because they have indeed mattered in our own careers; but we also tend to undervalue qualities that cannot be quantified—patience, curiosity, and the ability to connect with others. Krouse's question reminds us that the goal of education is not to replicate another version of ourselves but to help children become who they are.

In today's world where AI tools are readily available, teaching children how to use tools is certainly important, but even more important is teaching them to judge when not to rely on tools and when to face difficulties alone. This kind of judgment cannot be programmed—it can only be cultivated slowly through day-to-day companionship and conversation.

One-line Summary

A tech entrepreneur explores what to teach children in the AI era: learning methods and values matter more than specific skills.

Paragraph Summary

Steve Krouse poses the question "What should I teach my son?" on his blog, sparking over 300 comments on Hacker News. The article argues that in a context where AI makes knowledge acquisition nearly free, education should shift its focus from specific knowledge to learning methods and thinking habits. Parents' behavioral modeling is more effective than verbal instruction—children learn by imitating how their parents solve problems. Technical literacy must develop in balance with judgment, empathy, and aesthetic ability, avoiding utilitarian coding education. Consensus in the discussion includes reading habits, self-directed projects, and space to tolerate failure, while divergence centers on the timing and intensity of technical education. The article's value lies in modeling an honest attitude that acknowledges uncertainty and continues to think.