Field Notes
Thinking, in progress.
Notes from the lab. Things I'm seeing in the field, trends I'm exploring, arguments I'm still having with myself. One idea each, no tidy endings promised.
Read the NotesWhy these exist
Most writing about AI and strategy comes from people commenting on the work. These come from inside it. A Field Note is me working something out in public: one idea, triggered by something real, followed as far as I can honestly take it. If a Signal Page is the reference answer, a Field Note is the lab notebook.
I write these with AI and say so at the bottom of each one. The ideas and judgment calls are mine. The AI helps with structure, phrasing, and arguing against me.
The Notes
Published Field Notes
Five patterns from 19 AUTONOMOUS 2026 sessions, spanning verification, proprietary context, ownership, pricing, and the shift from making to judging.
Read →The side project didn't die. It moved inside the org. The builders who thrive are the ones shipping useful internal tools every week.
Read →Prompts are disposable instructions. Skill docs are standing policy. The real shift in working with AI is from telling it what to do once to teaching it how you work permanently.
Read →A widely shared post argued new software will behave like media: made fast, shared in feeds, discarded by dinner. When building gets cheap, the feed becomes the distribution surface, and timing replaces durability.
Read →A site that looks like 2012 is either dead or quietly doing $50M a year. Design quality is a demand signal, just not the one most strategists read.
Read →Anyone can generate working code now. The scarce skill isn't prompting โ it's knowing when the output is confidently wrong. Vibe coding is a judgment problem.
Read →Your organic traffic is dropping while your content works harder than ever. In the answer-engine era, the metric that matters isn't the click โ it's the citation.
Read →