Loiane Groner

🚀 I’ve started a new YouTube course: Modern Angular From zero → advanced, using how Angular is meant to be written today: • Standalone components • Signals-first patterns • Modern templates & tooling • Real projects, no legacy fluff ▶️

🚀 DevFest season is upon us! I'll be speaking at #DevFestFL in Orlando, Nov 7–8, about Specs-Driven Development (SDD). Joining the hackathon? Put the concepts into practice and use them in your project right away! 💡 Come join us! 👇

Loving Opus 5.5! 🥰🥰🥰

Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.

RT @simas_ch: People are confusing software engineering with coding. The later will be replaced by AI but the first is more important than ever.

Loop engineering is easier than you think: it's just scheduling, but the thing running on the schedule is an AI agent instead of a fixed script. 4 practical examples in my latest blog post, along with the concepts so you can build your own loops. #LoopEngineering #AI #SDD

🔗

Having the option to work with different AI tools and models is one of the best things about AI right now. Until last week, Claude was my favorite. This week, it’s been really slow for me, so I switched to Devin with Gemini and also Codex. Every week brings a new experience. The AI tooling landscape is moving fast.

100% test coverage on an Angular pipe. Stryker still found two real bugs: a redundant branch coverage happily marked "tested," and a missing null check that throws in prod. Wrote up mutation testing for Angular with Stryker, the other half of my PIT post.

An AI agent builds your Spring endpoint and Angular service in one session, then renames a DTO field and forgets the client that reads it. Both compile. Both test suites stay green. Production breaks. Contract testing catches it.

🔗

RT @simas_ch: The Java Story | The Official Documentary

The power of AI. I had back-to-back calls and needed to resolve a problem to achieve a certain metric in a legacy codebase. I had the AI agent running in a loop until the metric was achieved. I joined my calls, and 3 hours later the issue was resolved, all I had to do was configure the prompt and loop.

It’s #mvpbuzz day! Thanks again for the opportunity to be part of this program for the 9th consecutive year! Looking forward to continuing contributing!

Correction: 10th year! 🎉

Architecture diagrams are where design intentions go to die, especially when AI writes your code. 💀 AI agents optimize for immediate features, not structural boundaries. They’ll happily bypass layers and inject a repository straight into a controller for a fast path to a green test. They aren't careless; they just lack structural feedback. To protect your codebase from rotting, you must treat architecture rules exactly like unit tests. Enter ArchUnit for Java & Spring Boot. 🛠️ Learn how to build automated "architecture fitness functions" to enforce layer boundaries, block leaked database entities, and safely baseline legacy debt. Stop trusting your diagram. Start testing it.

👇 Full guide here:

RT @ladyleet: We're #hiring! @ThisDotLabs is looking for Engineering Leads, Architects, and Senior Software Engineers to join our team. If you're an experienced engineer who enjoys solving complex technical problems, leading projects, and building modern software with #AI as part of your workflow, we'd love to hear from you. These are fully remote roles with overlap during U.S. working hours. 📩 jobs@thisdot.co

RT @angular: With generative UI, you can create truly dynamic, user-centric experiences. You really don't want to miss our latest blog post on A2UI. We dive into using the new A2UI protocol from the Angular developer's perspective. 🚀 Catch up here:

Got my first Claude loop running at work today, connected to a few MCP servers! I'd only experimented with it on my personal machine before, so seeing it work in a real workflow was awesome. It saved me a ton of time and let me focus on other tasks while the loop was running. This feels like a real game changer.

Stop chasing 100% line coverage. It’s a vanity metric. Line coverage tells you what your tests execute, not what they verify. It’s time to move to Mutation Testing. I just wrote a deep dive on setting up PITest in Spring Boot + Maven. #Java #SpringBoot #Testing

Stop testing for numbers. Start testing for quality. Read it here: