The Essence of Learning: Abstract Models, Philosophy, and Computer Architecture

What to Learn

Zhuangzi said that life is finite while knowledge is infinite, making it foolish to try to learn everything; therefore the ultimate learning goal is not knowledge itself, which is transient, but rather enduring abstract models that act like universal keys.

Mathematical formulas are perfect embodiments of these abstract models, allowing us to understand the laws governing the universe and nature.

Every discipline has its own abstract models, akin to stars in the sky—some similar, many different—so expanding our cognitive structure means expanding the boundaries of these models.

For computers, what is the unchanging essence?

Computer Model

Physically, a transistor’s capacitance has two states (on/off); electrically, voltage has high/low—these correspond to binary 0 and 1 (excluding quantum computing). Adding more capacitors or lines yields 2⁴, 2⁸, 2¹⁶, 2⁶⁴ states, and with advances in nanotech, multi‑core CPUs, and 5G bandwidth, the number of representable states keeps growing.

Regardless of how complex the virtual world becomes, tracing back to the physical origin starts with capacitance; mathematically with binary; philosophically with yin‑yang.

The powering‑up of a computer resembles the Big Bang, after which bits embark on a journey: disk → bus → memory → CPU, constantly moving and multiplying.

Thus the world is built on a stable philosophical foundation, a mathematical representation of infinite states, and a wave‑particle‑based ultra‑efficient substrate.

Von Neumann System

The classic computer pyramid—CPU, memory, controller, I/O—remains stable across PCs, mobiles, and emerging IoT architectures, with improvements only in performance and power consumption.

Compilation Principles

Understanding any programming language’s low‑level workings requires knowledge of lexical analysis, syntax analysis, semantic analysis, regular expressions, and finite state machines; these concepts have changed little over time.

Whether Go, Rust, Java, C/C++, Python, JavaScript, or C#, the focus should be on the underlying compilation process rather than superficial syntax differences.

All languages, regardless of being scripted or compiled, construct syntax trees, perform lexical and semantic analysis, and translate the tree into binary code.

Distributed Principles

Distributed storage systems use identical data‑replication methods, first described in Lamport’s 1978 paper “The Implementation of Reliable Distributed Multiprocess Systems.”

Even after decades, these principles still underpin modern databases (MySQL, SQL Server, Redis, MongoDB) and messaging systems (Kafka, RabbitMQ).

Methodology

Agnosticism: Assume we know nothing, then re‑examine our thoughts; this mirrors Einstein’s view of the universe as a watch whose interior is unknown.

Universal doubt: Question all established beliefs, separating thought from self, akin to discarding rotten apples to discover new truths.

Independent thinking: Recognize that thought is not identical to the self; avoid emotional reactions that conflate criticism of ideas with personal attack.

Independent thinking in computing means adopting mathematical, model‑driven, rational analysis rather than mere intuition.

Summary

Therefore, learning should target abstract knowledge models—universal keys that unlock various domains—while employing agnosticism, universal doubt, and the separation of thought from self to continuously challenge and refine our understanding.