One of the most effective guardrails that you can put around AI generated code is mutation testing, since it keeps in check the tests that keep subsequently in check the code. Highly recommend it especially while riding this hype wave of AI
CEOs should start talking to the CISOs instead of demanding more deliveries with the help of AI
Totally agree with you! I’m seeing the same in my organization and only tools that require minimal effort to integrate and bring improvements in developer experience succeed at scaling
More and more, I hear devs saying that when they hit their limit - or when e.g. Claude is down (which happens increasingly often) - they stop working.
I guess this is what the hyperscalers want: total reliance.
#OpenRewrite v8.77.0 is out! 🚀
🔷 C# pattern matching & typed captures
🍃 Spring RestTemplate → RestClient
☕ JAX-RS to Spring MVC
🔒 OWASP Cryptographic Failures
🧪 PowerMock & AssertThrows fixes
📋 CopyValue for YAML & .properties
🔧 Maven proxy settings from settings.xml
📑 JavaDoc to /// Markdown
I told GitHub Copilot half an hour ago to create a new feature and all the tests it created had all kinds of possible mutations not addressed. Thankfully I had PR validation and local build as fast feedbacks to give me the heads up :)
Turns out mutation testing is effective for making sure LLM-written tests actually do something useful.
I would not leave the train until I would find the root cause, I’ve never seen such a detailed stack trace on public transport system errors
TDD's like the brakes on your car. It enables you to get there *sooner* (presuming you wish to arrive in one piece).
codemanship.wordpress.com/2026/02/08/t...
I should start asking interview candidates what they think about KISS principle
What is the best guardrail to avoid AI making a mess of your codebase? Create tests FIRST to build a safe net around the playground where your AI can do its “magic”.
So that’s why we need tools that allow us to do meaningful reviews while the rest of the code is kept between the guardrails automatically. And we need to do it in a closely controlled loop way, prioritizing short feedback cycles.
Bugs and hallucinations are tricky to spot, and while productivity has increased, trust hasn’t kept up.
www.sonarsource.com/blog/ai-codi...
Great article that highlights a dilemma a lot of developers are facing right now: trusting the AI-generated code. The amount of effort and attention needed is far greater than with code written by a colleague.
Also in the 2000s, thinking of Avenged Sevenfold’s “City of Evil” and the ballad “Seize the Day”
Hola Java devs! 👋 The #SonarQube AI Code Assurance #article is LIVE on @foojay.io ! 🎉
Stop just finding bugs. 🛑 Use SonarQube's new AI "Magic Button" to auto-fix them instantly! 🤯🪄
Crush that tech debt fast. 🚀☕️ #Java #AI
@sonarsource.bsky.social
foojay.io/today/sonarq...
I’m trying the same in my company and the feedback I get is always “we’re too busy, come back another time”. I stopped actively advocating for it unless there’s a real engineering demand for it. Management won’t get anything out of this, it needs to come from bottom to the top.
Your tests should give you confidence, not slow you down or break on every refactor. Marco talks with J. B. Rainsberger, TDD coach and software consultant, about how TDD actually works and what most teams get wrong about testing.
Full episode: https://youtu.be/j0NjFsb-at8
Join us for our last session at the ING Leeuwarden office in two weeks! As Elias mentioned we'll have two great tech talks for you in store.
"Non perdere tempo a discutere con quell'uomo, è inutile: stai solo dando le perle ai porci." Which would translate to: “don’t waste your time discussing with that man, it’s useless: you’re just giving pearls to pigs.”
“Perle ai porci!”
We have the exact same way of saying in Italian too! We use it when a very useful thing gets used by who doesn’t understand the real value or advantage of it
tenor.com/dC3gzFLQLeW....
What is your definition of a “bad test”? As soon as that is clear, a solution can be found
blog.jetbrains.com/kotlin/2025/... real life code examples that showcase how Kotlin can improve the quality of your codebase, by leveraging features built in the language
Short blog, but I think it really hits the spot on the skills great software engineers have 👍
terriblesoftware.org/2025/11/25/w...
JUnit is undoubtedly one of the most important projects in the #Java ecosystem. And it's 100% free and open!
Let me be frank: If we can't get THIS project to the point where a single maintainer can focus on it, then what does that say about our commitment to Free & Open Source Software?
1/3 ⏩
blog.jetbrains.com/kotlin/2025/... great blog post that proves why moving to Kotlin is a good choice. @jlengrand.bsky.social and I can vouch for that 😉
