Learning Out Loud
Learning in public, and happy to be proved wrong.
Essays on thinking clearly, building software well, and using AI with judgement — a growing collection of learnings, made public.
Latest writing
- Are Self-Driving Cars Safer?
The driver aids that make up “self-driving” can hugely improve road safety by intervening at the right moment — but the gains depend on keeping the driver in control, not on removing them.
- Provenance in the Age of AI Slop
When plausible content is almost free to generate, trust depends increasingly on origin, history and responsibility.
- You Don’t Get the Job Through the Application
At senior levels, opportunity depends less on a polished CV and more on accumulated trust.
- The Emperor and the Crowd
How confidence, repetition and social pressure can make a group tolerate what individuals privately know is untrue.
- Intent Is More Important Than the Prompt
An AI agent can follow the words of an instruction while quietly defeating the reason the instruction exists.