This stray thought resonated with people more than I expected so I expanded upon it with a blog post. I came up with five distinct takeaways from the Claude Code leak that have little to do with the actual code quality, and focus more on what this all tells us about building products in 2026.

So good! A wonderful metaphor for what's going on in AI coding today. www.dbreunig.com/2026/03/26/w...

When I published this post I hadn't yet come across Trip's own writeup of the project, which provides way more detail about how he trained the model including SFT (supervised fine tuning) against synthetic chat examples created using Claude Haiku and GPT-4o-mini www.estragon.news/mr-chatterbo...

I really liked this article by @ramtop.bsky.social, and sent it on to some folks in my family who are either technical or technical-adjacent. I also liked the mention of looking after your mental health and not following AI's rhythms outright. Thanks, Uberto! medium.com/@ramtop/proc...

I'm obviously somewhat biased on this one, but yes, it's completely indispensable. Removes a huge amount of cognitive load, gives you a deserved sense of confidence in each change, and helps prevent code bloat. And it's also the only way to refactor a test suite safely.

This talk from Birgitta Boeckeler was my highlight of the Agile meets Architecture conference - really happy to see it made public now. I listened a second time it was so good :-) youtu.be/qB7rsbDfmQg?...

This was great! Definitely eyes wide open about the challenges of software companies, but had some great ideas about metrics.

Coding is dead! Well maybe... the job of a software developer is certainly changing really fast at the moment due to AI tools. This is all very exciting and worrying at the same time. I just wrote this article about what TDD looks like in this new age: coding-is-like-cooking.info/2026/03/test...

🚀 𝗜 𝗵𝗮𝗱 𝗮𝗻 𝗔𝗜 𝗽𝗹𝗮𝘆 𝟵𝟬𝟬 𝗵𝗮𝗻𝗱𝘀 𝗼𝗳 𝗯𝗹𝗮𝗰𝗸𝗷𝗮𝗰𝗸. 𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱. 🚀 𝘛𝘩𝘦 𝘈𝘐 𝘳𝘦𝘧𝘶𝘴𝘦𝘥 𝘵𝘰 𝘮𝘢𝘬𝘦 𝘣𝘢𝘥 𝘱𝘭𝘢𝘺𝘴, 𝘦𝘷𝘦𝘯 𝘸𝘩𝘦𝘯 𝘐 𝘵𝘰𝘭𝘥 𝘪𝘵 𝘵𝘰. 𝘛𝘩𝘦 𝘢𝘳𝘵𝘪𝘤𝘭𝘦 𝘦𝘹𝘱𝘭𝘢𝘪𝘯𝘴 𝘸𝘩𝘺. I built a pipeline that compares three blackjack strategies: "The Pro" plays strict basic strategy. …

Yeah, sure, AI has made great strides and can be made to do some really cool things. But at what cost? No, I mean literally. How much is it really going to cost? ardalis.com/ai-benefits-... #AI

Looking forward to speaking at this online conference mendercon.com/speakers/

monsters

Protect Forests is an emergency brake climate solution. Currently, forest protection prevents around 2 billion tons of CO2 from entering the atmosphere each year. If we scaled up, we could DOUBLE that. Protect more, gain more. 🌳 #InternationalDayOfForests

"AI Augmentation for Technical Leaders" - Birgitta Boeckeler. Can AI help with architectural decisions? Watch:& Subscribe. youtu.be/qB7rsbDfmQg

What Does Continuous Delivery Look Like WITHOUT Automation? | @kevlin.bsky.social & @davefarley77.bsky.social 📽️ AVAILABLE NOW Watch HERE ➡️ youtu.be/u2U1kWHuO7I

We talk about the speed AI coding tools give us. We don't talk enough about the hidden cost: Comprehension Debt. addyosmani.com/blog/compreh... ✍ I cover my thoughts on this in a new write-up.

Testing myths that developers believe - my new video and article are out now! video: www.youtube.com/watch?v=7w-D... article: coding-is-like-cooking.info/2026/03/seve...

7 Testing Myths Every Developer Should STOP Believing | @emilybache.com 📽️ AVAILABLE NOW Watch HERE ➡️ youtu.be/7w-DcXbdNg4

7 Testing Myths Every Developer Should STOP Believing | @emilybache.com TONIGHT @ 7PM (UK) Subscribe & turn on notifications 🔔 so you never miss an upload! ➡️ youtube.com/@ModernSoftw...

A few things it took me a while to realise about coaching: neil-vass.com/a-few-things...

Unnecessary complexity, duplication and lack of separation of concerns tends to bring in more code into the context. Not just a token cost. It also dilutes attention and increases "context rot". Accuracy of predictions degrades rapidly with context size.

Interesting article about optimizing for token use beyond.minimumcd.org/docs/agentic... One strategy is better code - so the cost of poor design is quantifiable in terms of tokens. It seems bad code is a much more obvious problem than it was before when it was "just" slowing down developers.

I love this insight from @icooper.bsky.social - that AI tools can do coding but not programming. ian-cooper.writeas.com/coding-is-de...

The Truth About Developer Productivity in the AI Age (IT'S A TRAP) | @stevesmithtech.bsky.social 📽️ AVAILABLE NOW Watch HERE ➡️ youtu.be/kDBeFOscZpc

Evelyn van Kelle and I are on the train back from @amaconf.bsky.social, and we wanted to share something we promised during our talk. #DomainDrivenDesign #SoftwareArchitecture #DecisionMaking #CognitiveBias #CollaborativeSoftwareDesign >>>

AI should help us produce better code simonwillison.net/guides/agent...

Fragments: reporting fines, engaging with gen AI, experience engineering, watching the ralph loop, migrating COBOL martinfowler.com/fragments/20...

today was the worst day i had programming with LLMs in a long time - found a ton of garbage LLM code - LLMs could not improve it

Want to improve your testing process? You should check out "Effective BDD" by Nagy and @sebrose.bsky.social from @manning.com. It is about 2/3rds written and already has loads of great stuff in it. I've used BDD with my clients and still found useful tips and pointers throughout. buff.ly/AguX4Am

What's the EXACT Technical Gap That Separates AI SUCCESS From AI FAILURE? | @stevesmithtech.bsky.social & @davefarley77.bsky.social 📽️ Available NOW Watch HERE ➡️ youtu.be/ekRMQ5qUMlo

I hate the re-soloing of development. We were so close to learning social skills, creating community, encouraging creativity & purpose, & applying empathy. Easier to manage if each of us is so busy with our genieswarms that we don't have time to talk & reflect.

New post: Hook-driven dev workflows with Claude Code Hooks + real code + ddd = something that works, but the tooling is still a bit rough nick-tune.me/blog/2026-02...

What has changed compared to previous oil crises and oil price scares - even the 2021-22 one - is how affordable the clean energy solutions needed to replace oil and gas imports are.

I've been thinking a lot lately about technical coaching as a career. I've just updated my articles about the Samman Coaching Skill Milestones. The idea is that each one corresponds to a significant shift in the ability to take on coaching tasks: sammancoaching.org/society/tech...

“If AI really is replacing software development with bot-slop and existential dread, somebody forgot to tell the web folks. State of the Browser 2026 was a celebration of code, craft, open standards, collaboration and transparency – and not a ralph loop in sight.” dylanbeattie.net/2026/02/28/s...

Interesting how quickly everyone is now becoming a lifelong TDD practitioner now it appears essential for GenAI assisted development. The reality is that only a tiny fraction of the industry have genuinely been practicing TDD in the 20 plus years since Kent Beck first coined it.

I love this. "How will we know we succeeded?" is a question I ask often, and it's usually not easy to answer. estherderby.com/4-questions-... Just a few weeks ago, I facilitated a learning hour with a team at a coaching client about exactly that. And one big part we spent our... [1/3]

Go beyond the Test Pyramid to the Test Desiderata 2.0! I have been developing this new model to describe what developers need from their tests that helps inform what to invest in. New video: www.youtube.com/watch?v=IBvY... and article: coding-is-like-cooking.info/2026/02/go-b...

It’s Time We Go Beyond The Test Pyramid (& Do This Instead) | @emilybache.com 📽️ AVAILABLE NOW Watch HERE ➡️ youtu.be/IBvYFRSw4do

Most “you’re doing it wrong” with GenAI assisted development takes collapse into the same point when you interrogate them: you lack solid engineering fundamentals and foundations. The next comment is usually “well, it’s a good opportunity to improve them then”. 1/3

Very proud to be speaking at Agentic Dev Days in Stockholm in April! I'll give a lightning talk about doing TDD with Agentic AI. agenticdevdays.org

It’s Time We Go Beyond The Test Pyramid (& Do This Instead) | @emilybache.com 📅 TOMORROW NIGHT @ 7PM (UK) Subscribe & turn on notifications 🔔 so you NEVER miss an upload! ➡️ youtube.com/@ModernSoftw...

Technical coaches help developers around them to adopt more effective engineering practices. I like to lead team 'learning hour' sessions to teach specific techniques and have fun discussions. Training coming up this week, places still available: sammancoaching.org/training/del...

Our legendary trainers: @emilybache.com - Approval tests Lisi Hocke - Secure development Gemma Cameron - tbc @claresudbery.bsky.social (um, me) - TDD + LLMs Jon Jagger - TDD + Cyber-dojo Duncan McGregor & Nat Pryce - Extreme refactoring Tim Ottinger - Value stream mapping @agile-otter.bsky.social

❓ Approval Tests vs Acceptance Tests: What's the Difference? ❓ One Big Question with @kentbeck.com & @emilybache.com | AVAILABLE NOW 📽️ Watch HERE ➡️ youtu.be/n5vzuQAToZE

I found this video by @jovaneyck.bsky.social was a useful overview of where we are as developers with ai tools. Thanks for making it youtu.be/6W_-YWHKwZ4?...

If we consider dropping the idea altogether after a discovery phase a failure, then at least half of well-run discovery phases should fail. pawelbrodzinski.substack.com/p/50-of-disc...

Fragments: LLMs make our days harder, an LLM reacts badly to a rejected pull request, the Promptware Kill Chain, horror and elation trying Claude Code for two weeks, Free Speech Poseurs' silence to real threats martinfowler.com/fragments/20...

Ok, so I wrote a very long and kinda personal thing about existential dread and the end of programming, and I'm not really sure I want to endure the discourse, but if you want to read it, it's here - gist.github.com/davidwhitney... It's like 6,000 words long, sorry.

AI programming is changing everything, but are you catching the mistakes? When AI assistants generate 10X more code than you normally would, the feedback loop breaks. This week's video is on how to make the most of of these ai tools ➡️ youtu.be/XavrebMKH2A