As this is my first newsletter of 2026, I would like to wish you all a Happy New Year!
For me, this year will be the Year of AI. For the first time in my 25-year career in the IT industry, I feel this year I will change the way I write software. Over the past six months, I tried various AI chatbots for questions about the code I was writing. The results pushed me away. I got wrong answers and hallucinations.
In November 2025, I used a chatbot to convert numeric data into a CSV file and generate a bar chart. I had a number of commits on the Y-axis, and time on the X-axis. At first, I got a nice-looking chart, but a second glance revealed it skipped all the months with no commits, which turned the X-axis into a non-linear time flow. Mildly expressed, this is unintuitive. It was my first positive experience, which already highlights two aspects: you need to prompt clearly, and you need to review the result for soundness.
Over the holidays, I tried agentic coding, and I got excited. I started reviewing my prompt with the AI, asking it where it needed clarification, until I let it run. I learned that I get better results by making larger changes all at once. I also learned that I can let it run tests and execute bisecting sessions to identify a commit that contains a bug.
By now, I am convinced that agentic coding is a new tool that a software developer should be aware of. Now is the time to learn to use it for the right tasks. On Linux Weekly News, I read that others use LLMs for patch reviews. I have also rolled this practice out to the LINBIT team.
A statement I have heard about agenting coding that resonates with me is: Give it a proficient developer, and they will get faster. Do not give it to a newbie, as they will be tempted to take the output without sufficient scrutiny.
Let’s review at the end of the year whether the stock markets will see an AI bubble burst, as the dot-com bubble did in the early 2000s, or whether players’ revenues go through the roof as they promised to their investors. I stopped counting the 100s of billions of announced investments into AI. I am glad I am not in the investor’s shoes now. Instead, I can lean back and learn what the new shiny tools can do for me and LINBIT.
Moving on to the latest LINBIT content, we recently published Using the LINSTOR Python API to Interface with LINSTOR Clusters in Kubernetes by LINBIT Solutions Architect Matt Kereczman. LINSTOR APIs are a core part of the LINSTOR ecosystem, not an afterthought, and you can learn more by enjoying the blog post.
We held our Q4 Community Meeting in December. One of the segments was led by Moritz, our lead developer for all things Kubernetes, whose presentation covered the big news of our LINSTOR operator 2.10.0 update.
Regarding software updates, we have released LINSTOR GUI v2.3.0, and LINSTOR Operator 2.10.4 since I last wrote.
