The Slovene Ruby meetup, organized by Krištof Črnivec, is transitioning from seasonal meetup titles to a quarterly format, while keeping its now well-established structure of short show-and-tell talks interwoven with open-floor discussion.
The Q1 meetup at the Computer History Museum opened with a presentation by Damjan Rems, who demonstrated an elegant approach to working with .docx files in Ruby, focusing especially on safely modifying documents using user input from web application forms. His talk highlighted practical techniques for handling structured document generation in real-world Rails workflows.
This was followed by Martin Artnik’s energetic summary of Mario Zecher’s widely-shared article Thoughts on slowing the fuck down. The article argues that while modern AI coding agents can be powerful and useful tools, the industry is adopting them too quickly and too carelessly, resulting in increasingly brittle software systems, rising complexity, and declining code quality.
Artnik’s presentation sparked a lively open discussion, with many attendees sharing similar experiences and concerns. One recurring theme was the shifting role of senior engineers. Instead of primarily designing and building systems, they are increasingly expected to review large volumes of AI-generated code submitted by junior and mid-level engineers, or even by non-engineering colleagues. Participants noted that this trend not only consumes significant time and energy, but also introduces new sources of stress and responsibility, as engineers are asked to vouch for code they cannot realistically verify in depth. Several attendees pointed out the risk that this dynamic could contribute to higher burnout rates if it continues unchecked.
Continuing along the meetup’s central AI theme (though in a more optimistic direction), the final talk was given by Oto Brglez, who presented his AI-powered personal running coach application. The app features a simple chat interface, integration with Strava, a RAG system backed by a curated expert-written running knowledge base, and the ability to search the web for running events and related information.
Alongside a live demo, Brglez structured his talk as a practical “AI engineering 101” introduction aimed at developers coming from non-AI backgrounds. As a seasoned developer who only recently began exploring AI engineering himself, he shared his surprise at how approachable many of these tools and techniques actually are in practice.
At the same time, he raised an important question: when so many components of modern AI applications depend on proprietary APIs and hosted services, what does it really mean for such an application to be open source?
That question carried the discussion well beyond the formal program and into the nearby pub, where conversations continued late into the evening, helped along by delicious local craft beer.
Photos
