Weekly Roundup
Things that caught my attention!
What I Read This Week
Week ending 04 March 2026
A week of big questions dressed up as industry news. The usual mix of AI headlines — but a few things stuck with me. Here’s what I’d actually recommend reading.
1. AI tooling for software engineers in 2026
Solid data from 900+ engineers on how AI tools are actually being used in the wild. Claude Code is dominating. But the more interesting finding: it’s the senior engineers — staff+ level — who are leaning into AI agents hardest. Not the juniors. That flips a lot of assumptions about where AI disrupts first. Worth reading if you lead or work in any kind of engineering team.
2. Dario’s choice and Anthropic’s future
Anthropic’s reported clash with the Pentagon has been the story everyone’s been talking about this week. Big Technology frames it well — is this a Tim Cook/FBI moment, or something messier? I don’t think there’s a clean answer. But the question of how you build “safe AI” while operating inside geopolitical reality is one every AI organisation will have to answer eventually. Worth forming your own view on this one.
3. The design process is dead. Here’s what’s replacing it.
Jenny Wen, Anthropic’s design lead, on how the traditional design process is breaking down under AI pressure. Engineers are moving faster than the process can keep up with. This isn’t just a design story — it’s a preview of how almost every knowledge profession gets reshaped. What does your craft look like when the tools can do the rote work?
4. How to debug a team that isn’t working: the Waterline Model
Nothing to do with AI. Just a really practical framework for when a team isn’t clicking and you’re not sure why. The Waterline Model says: before you reach for the personnel lever, look below the surface — structure, incentives, context. One of those reads that’s immediately useful. I’ve bookmarked this one.
5. F1
How did a chaotic, dangerous, gloriously dysfunctional European racing series become one of the greatest business reinvention stories in sport? Great piece tracing the arc from near-irrelevance to global cultural phenomenon. The lessons about brand, patience, and community building apply well beyond motorsport. Save it for the weekend.
💭 One thing I keep thinking about
The Anthropic/Pentagon story ran through multiple newsletters this week from different angles — ethics, commerce, geopolitics. What struck me is how quickly the “responsible AI” conversation is colliding with questions about power. There’s no comfortable position here. But I think the organisations — and people — who’ll navigate it best are the ones who’ve already done the values work. Who know what they stand for before the pressure arrives. That’s the work worth doing now, not later.
👋 One more thing
A few people asked how I manage to keep up with this many newsletters without losing my mind. Honest answer: I have help.
I recently set up Percy — my own AI assistant, running on hardware at home in Wellington. Percy scans my inbox, flags what’s worth reading, and drafted most of this bulletin. Named after the barnacles on the Atlantic rocks of A Coruña — if you’ve read anything I’ve written about Galicia, you’ll get it.
I wrote about how it works (and why I built it this way) here: Meet Percy – My AI Sidekick. Worth a read if you’re thinking about what a genuine AI working partnership actually looks like — not the hype version.
See you next week.