The question "How do we use AI?" has shifted from the zeitgeist to a deafening noise where all answers feel equally valid and equally absurd. "It is a tool." "It is a partner." "It is cheating." "It is devaluing work." We are exhausted by the debate because we are focusing on the wrong side of the equation: we are looking at what the technology can do, rather than what we can do with the technology. To the instrumentalist, this is a matter of utility: How do we make it faster? To the cognitivist, this is a question of ontology: What does it do to the thinker?

Systems are engineered. Thinking must be Architected.

For the past quarter, I have used the term "Context Engineering" to describe the work of optimising AI inputs.

I am formally deprecating that term today.