I used to easily mistake organizing knowledge itself for an output.
After reading an article, I would write a summary; after finishing a book, I would organize the key points into notes; when I encountered a new topic, I would collect more information and build a fuller structure. These things are still useful, but with AI, organizing information itself is no longer as scarce. Given enough material, AI can quickly summarize key points, compare different viewpoints, and even write an article that looks complete.
The harder question gradually became: what in this information matters to me?
AI can infer preferences from my materials and conversations, but it does not naturally know which experience relates to my current situation, which viewpoint changed my thinking, or which choices I am willing to take responsibility for. Unless I make those judgments explicit, even very complete organized content may only preserve every point equally.
Knowledge Still Matters, but It Needs Human Judgment
Emphasizing perspective does not mean that having an opinion is enough.
Opinions without knowledge, evidence, and real experience can easily become hollow. The problem is not that knowledge has lost its value; it is that simply relaying or rearranging knowledge is increasingly less likely to count as a person’s unique contribution.
What usually creates the difference is how we choose, connect, and use that knowledge: Why keep this viewpoint rather than another? What experience does it connect to? What judgment did it change? What did it lead us to do?
I now use a more concrete test when I look at my own content: Which judgments in this piece only make sense in the context of my experience and tradeoffs? If I remove those parts, is all that remains a summary that anyone—or AI—could have organized?
This is not about proving that someone is irreplaceable. It is about checking whether I am only restating information. An article that only lists five productivity methods may not change much when the author changes. But if it explains when a method fails, how the author chose, and how the result changed their later judgment, the reader gets more than information that has merely been organized.
Perspectives Form Through Repeated Choices
Perspectives do not appear from nowhere. They often begin with small reactions: this is well said; I do not quite agree here; this reminds me of an experience; or this concept might change what I am doing.
AI can help summarize, compare, ask follow-up questions, and organize, but I still need to bring these reactions into the process. Each time I keep, discard, revise, or connect something, information that originally came from outside gradually becomes my own understanding.
If a judgment is still vague, I can let AI ask questions back and help different options surface. If the direction is already clear, the next step is to turn that understanding into concrete output that can be checked, used, or shared. Both are more useful for seeing whether a perspective is really mine than continuing to collect more information.
Expertise Is Not Going Away, but Expertise Alone Is Becoming Less and Less Enough
I used to push this question toward a distant hypothetical: if AI kept improving, would “expertise” eventually disappear, leaving only interest?
Now I think a more accurate description is this: expertise still matters, but the threshold for gaining access to expert knowledge and producing expert-looking output is getting lower. Knowing how to write, design, or build a prototype still has value. Those abilities may simply not be enough to answer what is worth doing, what quality is sufficient, or who will be affected by the result.
This is when interest, curiosity, experience, and responsibility become more visible. Not because they replace expertise, but because they determine what we explore over the long term, how we use our abilities, and where we are willing to invest our time.
So when knowledge is no longer scarce, what remains is not a new answer that can be separated from knowledge. It is a fuller combination: knowledge provides the foundation, experience provides context, judgment sets the direction, and action lets the perspective be tested against reality.
AI can help me process more information, but it also makes one question harder to avoid:
Among all these possibilities, what really matters to me?
The answer may not become clear all at once, but I need to start answering it.
If your judgment has not formed yet, you can continue with “Don’t Just Ask AI—Let AI Ask You Questions Too”.
If you already know what matters, the next step is “From Input to Concrete Output Is a Skill You Can Train”.