1. Spring and Java

>> New Java Benchmark for Coding LLMs puts GPT-5 at the top [foojay.io]

Brokk’s new Java-heavy benchmark (93 real-world tasks) ranks GPT-5 first across performance tiers and price points. The catch: it’s slow—and build noise, flaky tests, and quantization can seriously sway outcomes.

>> Infusing AI into Your Java applications [infoq.com]

Plug LLMs into Java with familiar tools—LangChain4j and Quarkus—so you can prototype fast and scale cleanly.

>> Debugging Transactions? Let Spring Debugger Do the Heavy Lifting [youtube.com/@intellijideau]

The plugin can help you visualize and understand how transactions behave and how entity states change as you debug. Quite useful.

Also worth reading:

Webinars and presentations:

Time to upgrade:

2. Technical & Musings

>> Team OKRs in Action [martinfowler.com]

Make OKRs a team sport: co-create objectives, own outcomes, and align by conversation—not top-down cascades. Solid weekend read as always.

Also worth reading:

3. Pick of the Week

>> New trend: extreme hours at AI startups [pragmaticengineer.com]

Next »
Java Weekly, Issue 609
« Previous
Java Weekly, Issue 607