Redundancy
Human language contains a surprising amount of built-in safety: our spoken communication is approximately 60% redundant, while written language maintains about 40% redundancy. This isn't accidental—it's essential.
Unlike the programming language APL, where a single letter change completely transforms meaning, human communication requires buffer space for error. "Almost no one can write dialogue so that it sounds right," notes computer scientist Richard Hamming, highlighting the often-overlooked distinction between written and spoken forms. This gap reveals something fundamental about our nature. We are inherently unreliable processors of information, prone to mistakes and misinterpretations. Systems with low redundancy—like APL—lead to undetected errors, while higher redundancy creates natural error-checking mechanisms.
Our languages evolved this way not because of inefficiency, but because of necessity—they accommodate the beautiful imperfection of human cognition.
Craft can be defined by the amount of redundancy you remove from your work. Speech to text makes people speak more, brain barfing onto a page. AI can take that, polish it, and turn into a viable idea. But that reinforces a habit of in precise words.