Field Notes
Practical writing for product leaders and boutique advisors – on deciding what to build, answering the hard questions, and running a sharper practice.
Consultant or Fractional CPO? You're Asking the Wrong Question
Founders compare a fractional CPO and a consultant on price and hours – but the real question is whether your gap is a decision or an operating system.
The Most Valuable Thing Research Can Tell You Is "No"
Before you build for a new market, the research that pays for itself is usually the one that returns a defensible no, not a green light.
KPMG Published a Report on Agentic AI. 40 of Its 45 Citations Were Fabricated.
I reported a bounce rate that was real and wrong, which is the small version of what just happened to four of the world's largest advisory firms.
Writing for People Is Persuasion Now. Writing for Agents Is Instruction.
Most documentation readers are agents now, which splits the job in two: writing for a person is persuasion, and writing for an agent is instruction.
Agents Refactored 26% of Commits. Design Smells Fell by Exactly Zero.
AI agents refactored 26% of commits in a 15,451-instance study, and the median change in design smell counts was exactly zero.
The Two Companies That Know Models Best Just Went Into Consulting
OpenAI and Anthropic both stood up billion-dollar AI implementation firms this year, which tells you who actually gets paid for AI work.
Everyone Can Run Agents Now: the Edge Is the Opinion You Hand Them
Everyone can orchestrate agents now, which makes orchestration the floor; what stays scarce is the conviction you hand the agent.
Syntax Pass Rates Climbed From 50% to 95%. Security Pass Rates Never Moved.
Veracode tested more than 100 models: syntax pass rates now near 100%, security flat at 56% for a year, and the gap is the real finding.
An Estimated 146,932 Fake Citations Entered the Scholarly Record in a Single Year
An audit of 111 million references estimates 146,932 fake citations entered the literature in 2025, and peer review is not catching them.