Things I've been reading this past week or so
From angry Tudors to AI tutors
Claude is not a senior engineer (yet) by Leila Clark
He wrote normal, bug-prone code, like you and me. He was good at coding because he was a gardener. Every time he walked into his codebase, he picked up his shears and manicured a bit of stray code. Over time, he rewrote every line of code over and over, tightening it down to only the perfectly-abstracted essentials.
Age of Invention: Tudor Trade War by Anton Howes
Having usurped the crown by invading England and taking it by force, Henry VII was willing to sacrifice everything, including the wellbeing of his new subjects, to prevent someone else from doing the same to him.
How to read ‘evidence pyramids’ by Tim Morris
A horizontal line does not indicate a “floor” that a design lives on but a ceiling for the design named below.
Given this “quality could be anywhere below the ceiling” reading, these pyramids/hierarchies are not particularly useful for judging the study in front of you. Well, maybe if it’s right near the bottom.
Maybe LLM tutors might be able to work... by Daisy Christodoulou
LLMs are very good at explanations, but explanations are over-rated as a means of learning. Sets of really good questions are better.
I built a Chrome Extension to audit my Substack newsletter’s health by Karen Spinner and Sharyph
To more systematically map Substack’s API, I asked Claude to write a script that would intercept and log every API call as I navigated the dashboard. This let me capture dozens of endpoints I would have missed otherwise, including ones triggered by lazy-loading, infinite scroll pagination, background polling, and UI components that only render under certain conditions (like having paid subscribers enabled, or viewing a specific date range).
It was not a super flu year by Christina Pagel
If the government and NHS messaging was about trying to change the public’s behaviour (and behaviour might have changed), then I don’t think the approach is justified. Getting the public to get to a behaviour through misleading presentation of data from the people and organisations who are meant to be trusted experts is counterproductive and wrong.
The Toxicity Firewall: How to Survive Office Villains Without Becoming One by Louise Deason
The most important part of the Firewall is internal. In racing, when a car crashes in front of you, you don’t stare at it. You look for the gap.
If you stare at the toxic colleague - if you complain about them at dinner, if you let them live rent-free in your head - you have already crashed.
Was the green lantern a good chap? by Dan Davies
It is incredibly fragile, but it’s obviously fragile, and in this way achieves a sort of paradoxical antifragility. In a “good chap” system, when a bad chap shows up, all the good chaps know that they have to band together and oppose or get rid of them, because they know that there are no systemic constraints on badchappery.
It’s about a deeper question. What matters more: to be able to get hold of something cheaply or to be able to make it domestically? That’s not an easy question to answer.
Stop posting AI slop 🛑 by Sam Illingworth and Daniel Ionescu
When words arrive instantly, it becomes easy to publish without deciding whether something should exist at all.
Drafts appear finished before anyone has asked what they are for or who they are for. Voice flattens because nothing has been weighed.
This is slop because no one stood behind it.


Adam, thank you for sharing again. There are a lot of interesting topics in this list that I can learn from.