Growing the capability is one accomplishment and owning the business that forms around it is another, and the industry keeps paying for the first while reporting the second.
Norman gave us the psychology of things, but we meet two thousand interfaces a day and forty doors, so the difference between his subject and ours is not the object but the dose.
When software costs almost nothing to make, the scarce thing is knowing what to build and proving it worked.
Software is getting cheaper. Taste is getting more expensive.
The harder the AI problem, the more it needs somebody the client already trusts.
Enterprises need their lovable software too.
Revenue stopped being a signal about the product the moment we learned to book it four years ahead.
The model was never the hard part. The context it was sent to act in was.
Four inputs decide an AI business, and a start-up can afford an edge in none of them.
A city survives the bomb, and a company does not survive a bad decade.
Twenty years of transformation run like bridge building, for a job that was a playground.
The future of the user interface is a stone.
We spent twenty years teaching computers a universal business language, and then it stopped being necessary.
Salesforce is an Oracle database with a few thousand buttons stuck to it.
Peak AI hype has buried the more interesting story — the cost of innovation is collapsing, and language may finally become the interface that makes work software serve the people using it.
Every jump in this history began with people saying something to each other that could not be said before.
