The Journey
Most posts here are ad hoc: what I learned building with AI that week, some philosophy, a writing lesson, something I came across.
Subscribe via RSS
Jev, the AI at a fiftieth of OpenAI's price: is it any good?
Jev is a new AI model that costs a fiftieth of the price of OpenAI's GPT-5.4. On my first test it beat GPT-5.4 by 11 points, and I told the internet. Two days later I ran the same test again, and the win was gone.
80% of AI projects fail. Nobody ever measured it.
"80% of AI projects fail." You've seen it in pitch decks, board papers, keynote speeches. It's used to justify buying AI software, and to justify not buying it. I went looking for the study behind it. There isn't one.
Why AI projects fail, and how a 1969 method reveals the conflict you didn't know you had
A kick-off meeting. Someone says "we're going to roll out an AI coding assistant across engineering." Six people nod. Everyone understood the sentence. Almost nobody meant the same thing by it.
Six AIs Said Graphify Uploads Your Code. I Checked.
Six AI models told me, with total confidence, that Graphify uploads your code to a server. They were wrong. So I ran it myself, measured what actually left the machine, and found a real privacy trap: just not the one they warned about.