ljómi: 10 years in the making
A little over 10 years ago, I had the fortune of witnessing the jaw-dropping spectacle that is the Aurora Borealis. My partner had treated us to a week in Iceland in mid-January, with the sole purpose of getting a look at this natural wonder. And to see first-hand the alien landscape that is Iceland’s southern flanks.
After a couple of days of gawping at lava fields, black beaches, and magnificent waterfalls, we ended up spending two nights at a ranch in the middle of nowhere, surrounded by nothing but deep snow. It was on the first night, while sharing a cigarette outside with a fellow guest (I’ve long since quit), that I noticed a green spear-like entity appear high above our heads. I dashed indoors to fetch my partner. We rapidly donned as many layers as we could, and dashed outside; the air temperature, with wind-chill, was around -25C. The Lights did not disappoint; shifting horizon-to-horizon arcs of glowing green ribbons; twisting cords of purple; it was all too much, and tears of wonder were shed.
It was a couple of years later, when recounting my Iceland experience with a close friend, when an idea began to form. We agreed that the Northern Lights, witnessed hundreds of years ago when their origin was a mystery, must have not only awed people, but struck fear into them also. And how better to convey that fear than if the Aurora made a God-like, gargantuan sound emanating from the entire sky?
As an electronic music nerd, my creative juices were set flowing. Alas, despite having developed software for many years hence, developing a system that could transcribe, in real-time, the Northern Lights into music, was beyond my capabilities. Until now.
AI software development tools, such as Cursor and Windsurf, Claude and Deepseek, have hit the scene, and with gusto. By the beginning of 2025, the idea of an AI assisting with software development became a reality. Today, it’s a norm. AI tools such as these, enable people to take an idea and realise it as a software system in an incredibly short time.
Needles to say, I jumped in feet first. And after several months of product development, specification writing, testing, and iterative feature development, Ljómi was born.
