Exploration · 6 min
Why Every Metric Eventually Learns to Lie
A useful metric can slowly become the job itself. Goodhart’s law explains gaming; the deeper problem is how measurement redirects attention.
A loose thread
Systems that learn the wrong lesson
What metrics, incentives, software and feedback loops learn from the signals we give them.
Exploration · 6 min
Why Do Incentives Produce the Opposite Behaviour?
Rewards and fines do more than change behaviour: they can change what an action means, what people optimize and who still wants to participate.
Exploration · 9 min
What Happens When AI Writes Code Faster Than We Can Understand It?
AI can accelerate code generation faster than review, testing and operational understanding. What happens when the assurance gap keeps growing?
Exploration · 5 min
Is Social Media Training Us—or Are We Training It?
We train recommendation systems with attention; their selections reshape what receives attention next. Preference and exposure become hard to separate.
Exploration · 5 min
Are We Prediction Machines—or Reward Machines?
Humans anticipate what happens next and learn from consequences. Predictive processing and reinforcement learning illuminate different parts of the loop.