JavaOne is over. And it was a really good one. Huge thanks to the Java DevRel team at Oracle. Chad Arimura, Sharat Chander, Heather Stephens, Nicolai Parlog, Lize Raes, Jim Grisanzio, Billy Korando, Ana-Maria Mihalceanu, José Paumard, Melissa Jacobus, Denys Makogon, David Delabassee, Crystal Sheldon You notice when people care about what they’re building here. For me, the best part was the conversations. Not the talks. The hallway track. Talking directly to the people working on Java. Getting their unfiltered view. Hearing how things actually work and where things are going. That’s the stuff you don’t get online. Also - very well organized. And yes, the food was great (not comparable to JavaOne from 10 years) I gave two talks around the Java ecosystem: The core idea was simple: Java can do a lot more than many people think. I showed how tools like OpenRewrite, Vaadin and EclipseStore help you move faster and get more out of your applications without throwing everything away. And then there was the ribbon game. Luiz Real and Ivar Grimstad took it seriously. Really seriously. Just look at their badges. Next stop is JCON Europe in April: europe.jcon.one/ The Oracle team will be there. If you want to have these kinds of conversations, that’s the place.

Day 2 at JavaOne: What I didn’t expect was how consistent the message of the day would be. In one of the sessions, “Just-In-Time Compilation for Java Performance: Recent and Ongoing Improvements,” Roberto Castañeda Lozano walked us through what has changed in the JVM. Not over the last decade. Just from Java 21 to 26. And that alone was enough to make me pause. He showed how much effort is going into the JIT compiler. How the JVM is getting better at understanding our code, optimizing it, removing allocations, and using modern CPUs more effectively. Even fundamental things like subtype checks are being reworked, leading to real-world gains—Netflix is seeing up to 3.5x improvements in certain scenarios. What makes this different is not a single breakthrough. It’s the accumulation. Each release adds something small. A better optimization here, less overhead there, smarter decisions in the runtime. None of it feels dramatic on its own. But together, over time, it compounds into something significant. Sitting there, you start to realize that the system gets faster not because we rewrite everything, but because the platform keeps moving forward underneath us. Java rewards patience. Later that day, Brian Goetz took the stage and told a story from his college days. He struggled with writing an essay and, after receiving a bad grade, argued that writing was hard. His professor responded: “Writing is not hard; thinking is hard.” That line reframes what we’re seeing. If the platform keeps improving performance for us, and tools increasingly help us write code faster, then the real leverage shifts. Not into syntax. Not into micro-optimizations. But into how we think about problems, systems, and decisions. And this is where it connects to a broader theme. Java doesn’t chase hype cycles. It evolves. Quietly, steadily, and with a focus on long-term stability. That can make it easy to underestimate. Especially if your mental model is stuck a few versions behind. But days like this make it clear: if you stay with the platform, it keeps paying you back. Before you optimize your code, try upgrading to Java 26 and measure first. There’s a good chance the JVM already did the work for you.