Writings
Writings on AI, data, and the organizational realities behind success and failure.
The Scarcity Upstream of the Answer
As AI makes answers cheap, from a recalculated ratio to a diagnostic score, the scarce work does not disappear. It moves upstream: from accepting the number, to choosing what it should optimize, to remembering why the threshold was set.
When a Model Learns to Speak
A statistical prediction does not become more accurate when it begins speaking in complete sentences. What changes is the kind of authority institutions may quietly grant it.
Where the Easy Arithmetic Runs Out
Organizations readily compare what is visible: benchmarks, competitor rollouts, projected savings. The harder question is whether any of those measures correspond to the value they were supposed to create.
The Donkey That Does Not Starve
As AI systems become increasingly capable, many discussions quietly assume that motivation comes bundled with intelligence. What distinguishes humans may not be reasoning itself, but the urgency that comes from having something to lose.
The Cost of Knowing What to Believe
Computation is no longer software’s scarce resource. Judgment is. As AI makes answers increasingly cheap, the harder institutional problem becomes deciding which answers deserve to be trusted, acted upon, and defended.