Honest Take — Before You Begin
Companion to COMPUTERSYSTEMSMASTERYCURRICULUM.md Written April 30, 2026 — honest reflections, no sugar coating.
Computer Architecture — From Gates to CPU covers: Nand2Tetris, Computer System Architecture, Write Great Code Vol 2. Understand how a computer is built from the ground up: logic gates, combinational circuits, sequential circuits, ALU, CPU, memory hierarchy, pipelining, caching. Build a working computer in Nand2Tetris.
When you write User.find(1), Rails generates SQL, sends it to PostgreSQL, which parses it, creates a query plan, reads from disk (or cache), and returns results. The performance of every step depends on architecture: the CPU's branch predictor guessing whether your WHERE clause will match. The L1/L2/L3 cache determining whether the index page is in cache or must be fetched from RAM. The disk controller's DMA transferring data without CPU intervention. After this module, you understand why "reading from cache is 100x faster than reading from disk" — because you know what a cache line is, what a cache miss costs, and what happens in the memory hierarchy.
This course unlocks once you've finished its prerequisite. Open prerequisite →
Companion to COMPUTERSYSTEMSMASTERYCURRICULUM.md Written April 30, 2026 — honest reflections, no sugar coating.
There is one logic gate — NAND, false only when both inputs are true — from which every other digital component can be built. NOT, AND, OR, XOR, multiplexers, adders, registers, R…
Somewhere in project 2 of Nand2Tetris, you probably built a chip that worked but felt bloated — more gates than the problem seemed to deserve. You were right. There is a formal me…
Paste this into godbolt.org (Compiler Explorer) as C, with -O2:
Work through each item before the checkpoint.
Build a working computer from NAND gates up: 31 chips in HDL, two Hack assembly programs, and an assembler. Every chip is verified against the official Nand2Tetris test scripts; y…
6 lessons. Read in order; spiral back when you need to. By the end you'll have used the core ideas twice — once on the abstract, once on something you'll meet at work next week.