COTRAX covers artificial intelligence the way it actually moves: fast, uneven and full of claims that need checking. Our writers ship production ML, run agent stacks in anger, and benchmark tools before recommending them.
The gap kept getting wider
COTRAX started because the distance between what AI can demo and what AI can ship kept growing, and almost nobody was writing honestly about the difference.
Launch posts described capabilities. Nobody described what happened in week three, when the model changed underneath you, or the tool returned HTML instead of JSON, or the cost per task quietly multiplied by forty.
So we started publishing what we found in our own systems — including the parts that made us look slow. That turned out to be the thing people actually needed.
What we write about
Six running threads, updated as the field moves rather than when a press release lands.
Heading
Six running threads, updated as the field moves rather than when a press release lands.
Heading
Six running threads, updated as the field moves rather than when a press release lands.
Editorial standards
These are commitments, not aspirations. If we break one, tell us and we will publish the correction.
Editorial standards
These are commitments, not aspirations. If we break one, tell us and we will publish the correction.
Small team, deep benches
Bio
Five years, briefly
Timeline Title
Timelien Description
Where we stand today
Got a story worth telling?
Tips, corrections, pitches and partnership questions all reach a human. We read everything, even when we cannot reply to everything.
Contact →