The Loss Function
Organisations are optimising for the appearance of AI progress while ignoring what's essential. The electricity era made the same mistake for thirty years.
Archive
Organisations are optimising for the appearance of AI progress while ignoring what's essential. The electricity era made the same mistake for thirty years.
Nobody ever gave their Terraform pipeline a job title. So why are we handing out name badges to system prompts?
The distinction between working with AI and working through it isn't semantics. It's the difference between using a tool and becoming the steering layer for one.
The shift from working with AI to working through it changes everything. Specs gave way to skills, and now building AI-native workflow is the capability that matters.
There are two competing visions of how AI will reshape work. Most organisations are starting with one. I think there's a better sequence.
We're at an interesting point with LLMs. The engine has just been invented. Why are people already attaching wings?