Binary data in computers
Inside a computer, everything is bits. Sensors turn real-world signals into binary before any AI runs.
A microphone waveform, a camera image, and a text message all end up as sequences of 0s and 1s. How you pack those bits (integers, floats, character codes, pixel grids) decides what algorithms can do next.
Bit
A single binary value: 0 or 1. Groups of bits encode numbers, characters, pixels, samples, …
From sound to numbers
An analogue sound pressure wave is sampled thousands of times per second. Each sample becomes an integer or float. Your speech recognizer never hears “sound” — it sees a long vector of those numbers.
Key takeaways
- Binary is the universal substrate.
- Representation choices constrain learning.
- Digitization always loses some information.