Notes from the workshop, not the marketing department.
Short, practical writing on AI, product strategy, architecture, and search — from a studio that builds and operates its own products. Everything here is something we've learned by shipping, occasionally at our own expense. No "digital transformation" essays; if a piece stops being useful, it comes down.
When AI is the wrong tool
A working rule for choosing between a model and a script: if you can write down the rule, write the rule. Where AI genuinely earns its keep, where it quietly fails, and why "no model needed" is a feature.
ReadThe economics of small software just changed
Why the tool, portal, or product idea you shelved because it cost too much probably doesn't anymore — a straight account of what AI-assisted development changed, and what it didn't.
ReadHow AI assistants decide which businesses to recommend
ChatGPT and Perplexity read the same web Google does — they just weight it differently. What actually moves the needle, from running automated audits on real businesses.
ReadThe smallest useful version
The first version should usually be a tenth the size you imagine. How to find the cut line, the difference between small and unfinished, and when "smaller" is the wrong advice.
ReadTechnologies change. Judgment compounds.
Eight technology eras from one vantage point: what every hype cycle looked like from inside, which bets aged well, and the small set of principles that survived all of them.
ReadThe Plain English Web Dictionary
Schema, SSL, citations, hosting, llms.txt and more — the vocabulary of the web explained without jargon, each term linkable. The reference layer behind everything else here.
BrowseExpertise should be checkable.
Any agency can claim judgment. Writing it down — specifically, with the trade-offs and the failures — lets you check ours before you ever send a message. If you read something here and disagree, that's a fine opening line: richard@ivylane.com.