CONCEPT
From Transistors to Tokens
The fundamental shift in what is being scaled: from physical objects that can be counted, photographed, and fabricated to statistical artifacts manipulated through matrix multiplication — a change in unit that alters the character of the scaling law, the nature of its limits, and the economics of its sustenance.
In 1965, the unit of measure was the transistor — a physical object, a switch made of doped silicon, occupying defined area on a die, governed by quantum mechanics, fabricated through photolithography. The transistor could be counted, photographed, measured. Moore's Law scaled a real thing. The new scaling laws measure something different. The unit of the AI era is the token: a fragment of language, typically a word or piece of a word, processed by a neural network during training or inference. Tokens are not physical objects. They do not occupy space on a die. They are statistical artifacts manipulated through matrix multiplications on hardware whose relationship to any individual transistor is so remote as to be meaningless.
In The You On AI Field Guide
The shift from manufacturing to metabolism changes the economics fundamentally. A transistor, once fabricated, generates value for