When you add two numbers in your head you naturally use base 10, almost certainly because human beings have ten fingers. Yet not a single computer on Earth uses base 10 for its core arithmetic. They all use base 2. Why?
The answer is engineering, not aesthetics. Transistors are easy to build with two reliable states (off and fully on), and brutally hard to build with more. Imagine trying to make a base-4 chip where each wire can carry four distinct voltage levels: say 0 V, 1 V, 2 V, and 3 V. Now your circuits have to reliably distinguish four thresholds despite manufacturing variation, temperature drift, power-supply noise, and crosstalk from neighboring wires. Each gate would need an analog-to-digital comparator with three reference voltages. The noise margin on each level shrinks. A neighboring clock edge couples a few hundred millivolts of noise onto the line and you cannot tell whether you are looking at "2" or "3."
Two states, on the other hand, are radically forgiving. You set the threshold halfway between rail-to-rail. Anything above is a 1, anything below is a 0. A 5 V chip can tolerate hundreds of millivolts of noise and still get the answer right. A 1.2 V modern CPU has thinner margins, but the principle holds: two values per signal is the most noise-tolerant scheme nature offers when your underlying device is a switch.
Each signal is a bit, short for binary digit, with values 0 and 1. (You will also see "false" and "true," "low" and "high," "set" and "clear." All the same thing.) Eight bits make a byte, the universal unit of memory addressing. The choice of 8, rather than 6 or 9 or 12, is partly historical: IBM's System/360 standardized on 8-bit bytes in 1964, and almost every architecture since has followed. It also happens to be a clean power of 2, divides nicely, and fits two hex digits.
Stadium-scoreboard analogy. Picture an enormous scoreboard at a stadium with 100,000 light bulbs. Each bulb is on or off. Even though one bulb carries only a single bit of information, the scoreboard can collectively spell out player names, scores, animations, even simple video. The information lives in the pattern of which bulbs are on, not in any individual bulb's brightness. Your computer works the same way: trillions of bits, each one trivial, encoding everything from this paragraph to a movie to the memory of a self-driving car's neural network.
This abstraction is also what makes hardware security attacks possible at all. If a chip's internal logic transitions between clean digital states, then power consumption, electromagnetic emissions, and timing all correlate with which bits are flipping. Power analysis attacks (Chapter 24) sit precisely on this fact: the attacker peers at the analog underbelly of the digital abstraction and recovers secrets the abstraction was supposed to hide. The digital world is a useful lie. Lies leak.