It's PRIME DAY time! In some @amazon online stores my book Building Micro-Frontends is on sale! You get the opportunity to buy it for a fraction of the retail value... for instance in the US store the paperback version is only $27 😱 Check out your local Amazon store: link.amazon/B0hokFTF5 to see if you have any promo applied! And if you don't know what changed from the 1st edition or you wanna have a deep dive into the structure of this book, fear not! I've got you covered: buildingmicrofrontends.com/b… Enjoy the Prime days! #microfrontends #discount #books #learning #primedays #amazon

310 – preview.mailerlite.io/previe… • A living map of everything we run • Managing Asynchronous APIs at Scale • The Agent Harness: What it is and Two Ways to Build One • Modernization is no longer a project @thoughtworks @IDC • Where Does Software Architect Fit in #DDD?

It's always an honour having your book on a conference of this level 🙏 thanks @GOTOcon

309 – preview.mailerlite.io/previe… • Killing the ivory tower • A Governance-First Architecture for Enterprise Personalization • Signals & Levers: Systems Thinking Tools to Unblock Software Delivery • Does it matter how you feed an agent its context? • Cracking the Code

Ready to start the first episode of a brand new live show called "What's happening in software architecture" in collaboration with @OReillyMedia Will I see you there ? 😁 learning.oreilly.com/live-ev…

308 – preview.mailerlite.io/previe… - Context Engineering at @LinkedInEng - Becoming an AI Team @PinterestEng - We turned off Pub/Sub and nobody noticed @incident_io - The 13 Multi-Agent Orchestration Patterns - System Design for the #LLM Era

YOU DON'T WANNA MISS THIS ONE. Starting next Tuesday 22 September, you can spend 30 minutes every other week catching up on what moved in software architecture during the previous fortnight. Each episode of What's Happening in Software Architecture on @OReillyMedia covers the trends worth your attention, the debates teams are having about #AI and socio-technical systems, a story from a real project and a book to add to your reading list, often with a guest who is working through these problems right now. The live Q&A is open to everyone attending, so if a decision in your own system is bothering you, bring it along and we'll talk it through together. We go live at 7am PT, which is 3pm in the UK and 4pm in Central Europe, and registration is open now. 🔗 oreilly.com/live-events/what… #SoftwareArchitecture #AI #SoftwareEngineering

The new episode of our podcast is out! We cover a lot of ground with Onur on the socio-technical aspect of every system Don't miss it! 👉 piped.video/j_paFwVqEEE

307 – preview.mailerlite.io/previe… - CircleK eMobility Journey Towards Effective Cross-functional Value Streams - Eval-driven development by @AirbnbEng - Message- and Event-Driven Systems with CQRS and Event Sourcing - Multi-Agent AI Engineering - Layered Architecture

I’ve been thinking a lot about what it actually means to be a software architect in the age of AI. We can talk endlessly about agents, models, context engineering and AI-assisted development, but there is a more fundamental skill underneath all of it: being able to design the system before you start building it. AI makes implementation cheaper and faster and THAT makes architectural thinking more valuable than ever. When you can spin up components, agents and integrations in minutes, understanding boundaries, trade-offs, constraints, failure modes and the consequences of your decisions becomes critical. That is exactly why I’m excited to run my next @InfoQ Certified Architecture Cohort, starting September 14. Over 5 weeks, we’ll work through architecture, decentralised decision-making, platform engineering, AI architectures and the role of AI in the software development lifecycle, applying everything to real problems from your own work. It is deliberately practical, with a small group of senior engineers and architects challenging each other’s thinking along the way. If you’re trying to make sense of where architecture fits in this rapidly changing AI world, I’d love to have you in the cohort. You can save $30 with code LucaARCH082630. Check out the cohort, link in the comments 👇

and if you enjoyed this issue, subscribe to Dear Architects, your weekly dose of #softwarearchitecture deararchitects.xyz/

#305 – preview.mailerlite.io/previe… - #Architecture as a Socio-Technical Craft - You can’t govern what you can’t see by @lucamezzalira - Making Your Data Ready for #Agentic #AI @martinfowler - The end of #programming - Running a Software Factory Efficiently at @UberEng Scale

What if there was something you could listen to the way you listen to a podcast, short enough to fit into a week, except you came out of it knowing something you did not know before? What if you could put your hand up in it, the way you would at a conference, ask the thing you are genuinely stuck on, and spend a few minutes on it there and then? And what if people from across the industry came in with it, and you can interact with them? That is what I have been building with @OReillyMedia, check it out: oreilly.com/live-events/what…

If #AI is generating code at speed, who is checking that your design decisions are still viable a few months later? We shrank the time it takes to write code, which finally leaves room for the part that always mattered more: understanding the context properly and turning that understanding into a system that works. What we lose is visibility, and we lose it fast. Several teams moving at that speed, shipping to production dozens of times a day, and the shape of the system drifts away from the shape you designed while every check stays green. We have used deterministic tools for years, to bump dependencies across a dozen repositories, to check whether the code quality was holding, to flag when complexity climbed somewhere it should not have. They are fast, they are cheap, and they are never wrong about the thing they measure. That has not changed, and it is still where most of this belongs. What they cannot tell you is whether any of it matters. A rising number is progress if you meant to build a hub and rot if you did not, and no tool knows which one you were doing. That is the half a model is actually for: reading across every signal at once and telling you what the combination means, while the numbers underneath stay something anyone can recompute by hand and argue with. The article is the whole build: the signals I chose for this system and why, where the model earned its place, and where it told me everything was fine when it was not. Enjoy the read and please share your thoughts!