My formal education was in mechanical engineering, not computer science or artificial intelligence. Most of what I understand about the field has therefore come through self-teaching.
I have learned by following questions beyond the point at which the easy explanations end—reading, building, getting things wrong, and returning to the fundamentals with a clearer sense of what I did not understand. This has made learning feel less like following a fixed curriculum and more like drawing a map while walking through unfamiliar territory.
The pull beneath the abstraction
When something interests me, I find it difficult to remain at the surface. I want to understand what is happening underneath the abstraction: why a method works, which assumptions support it, where its limits come from, and which parts of the accepted explanation are simply being repeated.
This tendency is a double-edged sword. It helps me develop a deep and independent understanding, but it can also lead me down rabbit holes when I should be building, testing, or moving forward.
Depth is useful only when it eventually returns to reality.
Understanding and action
I am learning to balance depth with action: to understand enough to build, let reality reveal the gaps in my thinking, and then return with better questions.
Building turns vague uncertainty into specific failure. A system is slow in a particular place. A user misunderstands a particular interaction. An assumption about the data breaks in a way no diagram revealed. Those failures create a better curriculum than abstraction alone can provide.
The goal is not to stop going deep. It is to give depth a rhythm: study, build, observe, return. Each pass should make the map more accurate and the next action more deliberate.
Why publish the unfinished path?
I do not want this website to turn learning into performance or present every note as a conclusion. I want it to preserve the development of the thinking itself—what I currently believe, where I remain uncertain, and how an idea changes when it meets better evidence.
Some thoughts here will become essays. Others may remain learning notes or unfinished research directions. Both are valuable if they make the next question sharper.