I’ve been thinking recently about what it actually means to learn.
While implementing an AI model, I noticed an interesting connection between how AI processes knowledge and how humans learn. It also made me reflect on my own experience.
During university, I often learned better from reading textbooks by myself than from attending lectures. This doesn’t mean lectures are ineffective. A good lecturer can explain difficult concepts efficiently, and I especially enjoy one-on-one conversations with lecturers and tutors.
For me, reading provides the space to stop, question, decompose, and reconstruct knowledge in my own way. A conversation allows me to actively test my understanding.
This became even more interesting while working with RAG AI. It retrieves information, identifies what is relevant, organizes it, and produces an output. In some ways, this resembles part of human learning:
retrieve → know → understand → apply → evaluate
And I think evaluation is particularly important.
We can retrieve information without truly knowing it. We can know something without understanding it. We can understand something without being able to apply it. And applying something successfully doesn't necessarily mean we know when it should be applied.
Language learning is a simple example. We may comprehend a foreign language long before we can express ourselves fluently. The gap between comprehension and expression reveals what we can actually produce, rather than what we can merely recognize.
This matters even more as AI makes obtaining answers increasingly easy. As AI becomes better at retrieving and generating information, our ability to understand, evaluate, and apply that information becomes increasingly important.
We need to ask:
Is this correct? Why do I believe it? Can I apply it? What happens when reality contradicts my understanding?
And perhaps most importantly, we need to learn to discern what is good from what is almost good. Because almost good is not the same as good.
I also used to be quite good at navigating the education system. I passed all my courses, but sometimes I was more focused on surviving the system than genuinely engaging with learning.
My perspective has changed.
Learning becomes much more interesting when I can take an idea and build something with it. Practice is a test of understanding.
Perhaps the real learning curve is not:
learn → learn → learn
but:
learn → learn how to learn → learn how to learn how to learn.
Every new piece of knowledge can change not only what we know, but also how we learn the next thing.
Perhaps lifelong learning is not simply about accumulating knowledge. It is about continuously improving our ability to understand, evaluate, apply, discern, and update what we know, while making thoughtful use of the tools that help us do so.
Perhaps what we need most is not simply more information, but more curiosity about the world—a curiosity that motivates us to keep learning.