Now
Updated October 2026A snapshot of what I’m working on, reading, and experimenting with, refreshed as things change rather than on a schedule. Inspired by Derek Sivers’ /now page convention ↗.
Most pages on this site are designed for visitors. This one’s a working log, kept honest, and useful mostly to me. Everything below stands as of the date at the top; it is not a live feed.
What I’m working on
Active work, roughly ordered by how much of my week each one takes.
In-house growth work
RoleSEO systems, AI-search visibility and the free developer utilities that earn the links.
Keeping this site current
PersonalThe 25 posts and twelve tools rebuilt in October get re-checked as the studies behind them are updated. The ranking-time figures were the first to change.
AI Citation Tracker
BuildA planned tool for the tools library: the question set, recording sheet and method for tracking when ChatGPT, Perplexity and Gemini cite a brand.
What I’m reading
In rotation: books, papers, essays. Not exhaustive, just what is actually open. Book links are Amazon affiliate links, which cost you nothing extra and earn me a small commission.
Co-Intelligence
Re-reading it. Holds up surprisingly well for a book about a field that moves this fast.
Right Kind of Wrong
For the experimentation framework. Reshaping how I think about utility-led prototyping.
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Foundational reference for how AI-citation surfaces actually retrieve. Worth the time.
Lenny’s Newsletter
Mostly for the operator interviews, skimming the SaaS posts, reading the ops ones.
Open questions
Hypotheses I am working through. No progress counters: each one gets answered when the evidence is good enough, not on a schedule.
Citation-rate decay on older content
How quickly does AI-citation surface degrade on content older than 90 days, and do minor refreshes restore it?
Entity-graph density vs. ranking velocity
Does pre-loading topical clusters with dense entity coverage speed up initial ranking compared with standard publishing?
Programmatic utility variants, fan-out economics
Past the 50-page mark, does each extra programmatic variant earn enough traffic to cover what it costs?
Founder distribution loop, writing cadence vs. inbound
How closely does this site’s publishing rhythm track inbound enquiries?
AI retrieval and SaaS distribution
How assistants that retrieve passages rather than rank pages change which SaaS companies get found. The question behind the AI search guides on the blog.
Recently shipped
What has gone live recently, newest first. The fuller history is on the about page timeline.
The working rule: depth over breadth. The temptation to publish more is constant; the discipline to publish less but better is the harder skill. Asking myself once a week: is this work compounding, or am I just performing motion? I have 3 client slots open for Q4 2026, and if you want to talk about the ideas here or a project of your own, my inbox is open.