By 2026, generative AI (GenAI) has gradually evolved and materialized into what is known as intelligent agents, entering our lives at an astonishing pace (regardless of the specific concept, this article will refer to it as AI for ease of description). From content creation and information retrieval to daily office work, more and more people are relying on AI to improve efficiency and assist in decision-making. With the help of AI, many things that were once only in our imagination can now quickly become reality. Take software development as an example. In the past, many people had a wealth of creative ideas and product concepts, but lacked the programming skills to bring them to fruition. However, with the rise of AI Coding, the barriers to entry for the entire process, from requirements description to code generation and deployment, have been significantly lowered. Today, a practitioner can quickly build product prototypes or even complete applications based solely on an idea.
This is precisely why the differences between people are amplified. For some, it’s an accelerator of efficiency; for others, it’s more like a shortcut to bypass the learning and accumulation process. AI enables many people to accomplish things that were previously beyond their capabilities, but it also accelerates the exposure of problems that were previously difficult to detect.
Ownership: The strong get busier, the weak get idler.
Laziness is human nature, which is why ownership is so important. People who lack ownership often just “complete tasks”; but in the AI era, what is truly scarce is not the act of completing tasks, but judgment, critical thinking, and an attitude of taking responsibility for the results.
For example, your boss asks you to research an industry or collect data. Some people would delegate the task to AI, which would then polish it to produce a well-structured, beautifully formatted, and professionally worded report. It would look high-quality, even better than many people could write themselves, so they would simply skim through it and submit it.
But is such a result truly valuable? The content provided by AI may be mostly correct, but it often lacks the most crucial element — thought. AI doesn’t know why the boss needs this information, what problems the team is currently facing, or what the company’s real priorities are. Its answers might be comprehensive, objective, and logically clear, but they could also be a bunch of “correct but meaningless statements.” This is because AI excels at answering questions but not at solving them (similar to…).Stay away from those who only talk the talk(The talker mentioned in the article?). Many people see the results provided by AI as flawless or even far exceeding their expectations, but in reality, they haven’t even figured out what the real problem is.
This feeling arises because the content remains at the informational level without addressing the actual problems. A person with true ownership, upon receiving a task, doesn’t first consider “how to complete it,” but rather “why do it.”
- Why does the boss need this information?
- What problems is the team currently facing?
- What decisions will this information ultimately influence?
- What content is most important, and what content is actually unimportant?
Only by clearly understanding these issues can one know from what angle to conduct research and which areas require focused attention. He breaks down the problems into multiple parts, forming his own analytical framework, and then uses AI to verify, supplement, and challenge his judgments. This process is far more time-consuming than directly asking AI questions, and naturally, the final value produced is completely different.
This explains an interesting phenomenon in the AI era: the strong are getting busier, while the weak are getting more idle. The strong leverage AI to restructure their work methods, breaking down, optimizing, and improving tasks that were previously impossible due to time and cost constraints. They don’t reduce their thinking because of AI; on the contrary, AI’s existence empowers them to contemplate more and deeper questions. The weak, however, tend to remain superficial, treating AI as an answer machine. While they may appear to complete more tasks, they essentially reduce their own thinking process. For them, the time saved by AI simply allows them to finish work faster, shifting the responsibility of thoughtful reflection to others.
AI hasn’t reduced workload; it has simply shifted the focus from execution to thought. This will become the standard for selecting and eliminating talent in the next era.
AI-generated works are actually obvious at a glance, but that’s not the problem.
Many people are confused: Why does my boss get annoyed when he sees a polished report generated by AI? Is it because my boss doesn’t like me using AI?
In most cases, bosses don’t care whether it’s written by AI or not. What really displeases them is a lack of effort on their part.
For you, this might be a high-quality output. The pages are beautifully designed, the structure is complete, the logic is clear, and it even includes charts and data analysis. But for the boss, he spent ten or twenty minutes reading through dozens of pages, yet still couldn’t find any truly valuable information.
- What is the conclusion?
- Why is this happening?
- Where is the most critical risk?
- What should we do next?
This content is often buried beneath a large amount of seemingly technical text. AI excels at generating content, but it’s not good at judging which content is important and which methods are practical and feasible. Therefore, people who lack critical thinking skills can easily fall into a trap: mistaking fancy content for value, and complete expression for deep understanding. The end result is a report that “looks very professional” but is full of “nonsense.”
Everyone knows Amazon has a concept called “One Page,” which is often required for a project or report. Essentially, it requires you to explain everything on one page. If a single page cannot distill the core ideas, then any amount of content is just a pile of information and a waste of time. (This is similar to why a resume should ideally be one page.)
This is why truly outstanding individuals, when using AI, often produce shorter deliverables. They leverage AI to gather information, validate hypotheses, and expand their thinking, but ultimately present decision-makers with only the most crucial conclusions.
The gap only begins to appear when the threshold disappears.
AI is breaking down barriers of knowledge and skills at an unprecedented pace, but this is not equality; rather, it’s widening the wealth gap. In the past, developing software required programmers; analyzing finances required a financial background; and designing marketing plans required market experience. Today, with the help of AI, a financial professional can develop applications, a product manager can write code, and an engineer can quickly learn about investment, law, and even marketing. The cost of acquiring cross-disciplinary skills is decreasing dramatically. Does this mean that industry experience and professional skills are no longer important?
On the contrary.
In the past, the differences between people largely stemmed from the barriers to accessing information, knowledge, and skills. Now that AI has significantly weakened these barriers, the truly decisive factors are emerging — judgment, depth of understanding, and quality of thought. In other words,AI has narrowed the gap in capabilities, but amplified the gap in cognition.Because AI can help you write code, but it can’t help you judge whether a product is worth making; it can help you generate a business plan, but it can’t help you judge whether there is real market demand; it can help you complete the analysis, but it can’t help you decide which data is important or which conclusions are credible.
Take AI coding as an example. Today, product managers, operations personnel, and marketing staff can quickly develop a website, a system, or even deploy the product to a production environment using tools like Claude Code and Codex. This is an exciting change. But the real challenge is never just building a product, but building a product that people want to use. AI will help you implement functions, but it won’t actively question your needs; it will help you solve problems, but it won’t tell you whether the problem is worth solving; it will constantly provide answers, but it rarely points out that the problem itself might be wrong. It’s better at answering questions than defining them.
AI coding lowers the development threshold, but it doesn’t reduce the complexity of the system itself. A demo that runs doesn’t mean it can support real users; a feature that goes live doesn’t guarantee long-term stability. From access control, data security, and performance optimization to monitoring and alerting, disaster recovery, cost control, compliance requirements, and the various edge cases and anomalies that may arise with user growth, these problems won’t automatically disappear with the advent of AI. Often, the truly difficult part isn’t even writing code, but anticipating problems, designing solutions, and making trade-offs under various constraints.
While using AI, some people will become more and more like professionals, while others may not even know what they need, let alone find keywords to convert into prompts for AI to do.
AI will enable more and more people to get started, but advanced experience and knowledge will remain scarce. As more and more people can enter the same field, the real difference will begin to shift from “whether you can do it” to “whether you understand it,” and from “whether you can do it” to “whether you can do it correctly.”
soul
The reason for concluding with “soul” is that it has become more important than ever in the AI era. Modern AI essentially learns and summarizes the knowledge and experience accumulated by humanity in the past; it can imitate, integrate, and generate, but it cannot truly possess its own values, beliefs, and pursuits. What truly gives a work its soul is never the tool itself, but the person using it.
AI is making knowledge readily available and execution more efficient than ever before, but it cannot replace the experience and insights accumulated by a professional over many years. More importantly, a sense of responsibility, judgment, values, and accountability for results are not only irreplaceable but will also be amplified.
