AI is getting smart and everyone is talking about vibe coding⦠but a piece I read recently ended with a line that I liked a lot! - āJava is here to stay, and so are folks who know itā. That felt super encouraging.
Apparently, first version of JUnit was written on a plane š. Kent Beck and Erich Gamma pair-programmed it on a flight from Zurich to the 1997 OOPSLA in Atlanta. What else were two geeks to do on a long flight but program? š
martinfowler.com/bliki/Xunitā¦#JUnit
I used to think I had to learn and figure out every single detail before starting a project.
Now I just start with what I already know and learn the rest in small iterations. And always keeping my final goal in mind.
This way I see progress and results(good or bad) much faster.
When I look back, I see how many mistakes Iāve made in code.āØAnd instead of being embarrassed, I use them as proof that Iāve been learning all along.āØThose mistakes are part of why I can solve problems faster now.
GitHub has moved fully under Microsoftās CoreAI, and Copilot is center stage. Feels like dev workflows are now AI-first by design.
windowscentral.com/microsoftā¦
Iāve realized that my environment affects how I think about code.
A noisy place makes me write quick fixes.
A calm space makes me write more thoughtful solutions.
āDevin, the AI engineer,ā just raised nearly half a billion dollars.
Impressive but someone still has to know how to prompt it well, review its output, and fit that code into the real system.
I still remember how lost I felt the first time I tried to write Unit tests.
Now I can't imagine a good project without them.
Progress doesn't always feel fast, but it's always there when we look back.
----
Java developer? Check my video courses: udemy.com/user/sergeykargopoā¦
Iāve realized that saying yes to too many side projects quietly eats up my energy, even if they all seem interesting.
I always feel happier and less burned out when I finish one thing before moving on to the next.
Anthropic shipped OpusāÆ4.1 for tougher coding and agent tasks. I like that theyāre chasing reliability more than showy demos.
anthropic.com/news/claude-opā¦
Every time when I write too much clever code is usually a trap. Itās fun in the moment and feels cool. But later, Iām the one stuck trying to understand what I wrote š¬.
Simpler code with good comments is always better.
When I'm working on something tricky, I often talk through the problem out loud. Especially when no one is around. Somehow, it helps me understand it better š.
JetBrainsā Kineto is another tool that lets us build AI apps without coding every line. I wonder if no-code tools like this might make us lose touch with the craft š¤. Still, itās tempting to spin up an AI app in a day and see what sticks.
youtu.be/2Xwu5aVvg84
Iāve noticed that when I stay in a rush for too long, my creativity drops.
My code and my lectures start to feel mechanical and plain.
Iāve learned itās important to slow down and not feel guilty about it, because that slowdown is what actually improves the quality of my work.
When Iām working on personal projects, there are days when I try to write the most elegant code I can.
While other times, I just focus on making the feature work and leave refactoring for the second iteration.
And Iām fine with that because it keeps my projects moving.
A little practical tip. Keep a short ānext stepsā list in your project folder.
When you return to the code, youāll know exactly where to start.
Itās a small habit that removes friction every day.
Some of the best developers I have worked with were not the fastest.
They were the ones who cared about leaving the codebase better than they found it.
That habit quietly builds your reputation over time.