4 Durable Thinking Skills for Working with AI and Keeping Up with IT Trends

Technology trends change fast, especially in the AI era. Instead of chasing every surface-level technology, author Phú Vinalinux argues that IT professionals should invest in four durable thinking skills: systems thinking, communicating with AI, adaptability, and observing one's own mind.

Quick answer: The four foundational skills for working with AI long-term are Systems Thinking, Expressive Reasoning (communicating with AI), Mindset Realignment (adaptability), and Mind Observer (observing one's own mind). Systems Thinking helps users steer and evaluate AI output, even in domains they have never worked in. Communicating with AI correctly means multi-turn dialogue, not a single prompt template. For general users, mastering the first two skills is enough to use AI well at work. Surface-level technologies such as RAG, Plan mode, and Agent teams change very quickly, while deep thinking skills endure.

The four durable thinking skills are a set of foundational skills that author Phú Vinalinux distilled from decades of Linux kernel work and research into the nature of AI: systems thinking to steer and evaluate, two-way communication with AI to convey the right context, adaptability to switch domains as technology changes, and self-observation of the mind to stay independent from an increasingly persuasive AI.

What are the four thinking skills?

The author published a series on these four skills in June 2025, and now revisits them with a clearer explanation of each:

1. Systems Thinking: Users need this skill to set direction and evaluate AI output. With systems thinking, users can steer and assess results even in fields they have never worked in. A programmer with systems thinking can use AI to perform well across many different technology domains.
2. Expressive Reasoning (communicating with AI): Users need this skill to explain to AI what they are thinking and to provide enough context for AI to do the job correctly. The skill works in two directions: one helps AI understand the user, the other helps the user understand AI.
3. Mindset Realignment (adaptability): Technologies and trends change very fast. Today embedded engineers are in demand; tomorrow they may not be, just as web app programming is today. Adaptability keeps the engineer's mind open and able to learn quickly when switching to another field.
4. Mind Observer (observing one's own mind): As AI becomes smarter and more skillful at communicating, users risk becoming dependent on it for thinking, and even losing their sense of professional purpose. The Mind Observer acts as a supervisor, keeping the user's mind independent and connected to the real world.

Why is the prompt-template approach wrong?

According to the author, most users initially only knew how to use AI through prompts: saving detailed prompt templates shared online, editing them, and reusing them for each purpose. In his June 2025 article on communicating with AI, he argued this approach is wrong: proper communication with AI should be a flexible process like talking to another person, with many turns to solve one problem, not a single prompt. Solving work through one prompt also weakens the user's thinking.

The author's observation after about a year is that, as newer AI generations became smarter and more understanding of users, the trend of "experts" sharing prompt templates online disappeared. Similarly, Systems Thinking was initially hard for people to connect with AI usage, but recently it is mentioned very often in AI discussions, along with many courses on the topic. These are the author's personal observations, recorded as his own assessments.

Which skills are easy to learn, and which are hard?

Of the four skills, only the first (Systems Thinking) and second (Expressive Reasoning) are relatively easy to understand and learn. The third (Mindset Realignment) and fourth (Mind Observer) are quite abstract and very hard to learn. The author hopes people will gradually understand the remaining two skills better, but believes that for general users, mastering the first two is enough to use AI well at work.

Why invest in deep skills instead of surface-level technology?

Surface-level technology	Deep thinking skills
RAG, Plan mode, Agent teams, Harness	Systems Thinking, Expressive Reasoning
Change and become obsolete very quickly with new AI generations	Enduring and unchanging over time
Learned by following trends, quickly outdated	Learned once, useful now and in the future
Depend on surface-level information	Look at the essence, at first principles

The author says he gained this deep perspective after decades of working on the Linux kernel. From his very first days exploring AI through ChatGPT, he saw that saving and editing prompt templates from the internet was unsustainable, so he studied the most fundamental things: Turing's computing principles, AI theory from the 1980s, the nature of human thinking in Buddhist philosophy, and the philosophies that allow two thinking systems to collaborate.

Why are trend predictions based on surface information often inaccurate?

Today, social media is full of analyses and trend predictions based on surface-level information. The author gives two examples: concluding from "Anthropic lost 100 billion dollars in 2025" that the AI bubble is about to burst, or concluding from Linus Torvalds's remark that AI-generated code committed to the kernel is poor that AI cannot replace programmers in writing code. Such analyses go viral easily because they are intuitive and easy to understand, but in the author's assessment, most predictions based on surface information are usually inaccurate. Both examples are quoted as they appear in the original post; Wotech has not independently verified these figures and statements.

Frequently asked questions

1. Which skill should a beginner learn first? Systems Thinking and Expressive Reasoning. According to the author, these two are relatively easy to learn and are enough for general users to work well with AI.

2. Are prompt templates still useful? In the author's view, prompt templates are not sustainable: correct communication with AI is multi-turn dialogue with full context, not a single prompt.

3. How does Systems Thinking help when using AI? It helps users steer AI and evaluate what AI produces, even in unfamiliar fields, instead of blindly trusting the output.

4. What is the Mind Observer? It is the skill of observing one's own thoughts, acting as a supervisor that keeps the mind independent, avoids dependence on AI, and stays connected to the real world.

5. How can one always keep up with IT trends? According to the author, look at the deep essence rather than chasing surface information: invest in durable thinking skills instead of only learning fashionable technologies, because fashionable technologies change very fast.

Source: post by Phú Vinalinux in the "Cùng nhau học Linux kernel" group, published 13/11/2026. This article rewrites and restructures the original post; the assessments and observations in it are the original author's.