About us

The man behind Haxbli, Karl Schultheisz, as interviewed by his foil, Claude Opus 5.5.

How did you start making music, and when did you first open up gear to change its sound?

I was about 14 when I got Sonic Foundry Acid 3.0 and began making beats with Acid Loops. Around the same time, I got my first hardware synth, a Novation K-Station, which I thought was aptly named. My first gear hack was a few years later, changing the tone stack of my Fender Blues Jr.

What was the Blues Jr. doing that you didn’t like, what did you change, and did it work?

I seem to remember the mids were a bit too scooped, not flat enough, and I wanted a flatter, thicker sound. At the time, I was really into the band Engine Down, and I liked their guitar tone, which was unconventional for using the neck pickup.

I’d say I was experimenting. Not necessarily chasing a particular sound, but beginning my journey of understanding the relationship between technology and art.

Did music lead you into electrical engineering, or the other way around?

They were both interests from the beginning. My parents met in a band—mom was a singer, dad a keyboardist. I was tearing down electronics by age six, and not putting them back together, of course. The convergence of the two into electronic music happened in my teens, after my dad exposed me to Dig Your Own Hole by the Chemical Brothers. I asked my dad how these sounds were made, and he said “with circuitry.”

I didn’t like high school much, so I chose to spend as little time as possible there by enrolling in an electronic technology vo-tech program. I didn’t realize it wasn’t meant to be a path to college. But it gave me an environment, for half of the school day, to explore something I cared about. I asked my instructor about music tech, such as voltage-controlled filters, that was beyond his knowledge.

Did you get your answer about VCFs? How did you get from vo-tech to engineering?

My tech instructor, Mr. Krug, was a good guy, but his best answer was to try JFETs, since they can have a voltage-controlled resistance. He didn’t know about OTAs or transistor ladders. I was psyched when I finally built a four-pole VCF with LM13700’s in his class.

After high school, I went to Penn State for electrical engineering. But my first-year coursework, except for calculus, was uninteresting. I dropped out and started a music production program at the Art Institute of Seattle. There, I didn’t fit in socially and didn’t want to be behind a console or interacting with artists. I wanted to design the gear. I wanted to engineer stuff.

When that finally sunk in, I decided that finishing my Penn State degree was my best shot at a decent life. So I did it.

After Penn State, what did you work on? Where did software come in?

I emerged from five years of engineering college convinced that I had no future in music technology. I got a job in automotive system engineering, working with audio-adjacent systems, but doing nothing creative with sound.

I had become a Linux user in college. I needed to escape my automotive job, and when the opportunity came, my Linux knowledge allowed me to switch to software engineering, which was a lot more fun, and became a passion.

After a few years, I got laid off from my software engineering job, and I began to think of starting my own business. Haxbli was legally formed in 2021, but didn’t emerge as a creative audio tech project until 2026, amid the LLM boom.

Let’s back up a little bit. r4’s precursor was written in 2020. What were you doing with audio beside your engineering career, and how did r4 start?

Since switching to Linux, I had given up on using Windows for music production, including making music. I was, and still am, a Linux die-hard. I couldn’t run FL Studio on Linux, but I could engineer my own music tech. I started playing around with audio programming languages; making my own, and exploring existing ones. I liked functional programming, and Faust makes it easy to write a plugin, so it was a good match.

What did AI change that made the business seem possible? What do you hand to AI, and what do you keep?

I was never a “10x” programmer. I tried very hard to be productive, but I eventually realized my strength is different. I’d get bogged down after a project reaches a certain size. And my mind was always flooded with new ideas, making it hard to finish my previous ones.

Using AI, for me, resembles being the leader of a team that I can delegate to. I’m then able to spend my time on what I’m best at and most passionate about: product design in the functionalist sense, how a product works in the hands of the user; excellence in engineering; and my creative vision.

Tell me about a decision in r4 you made by listening.

The instantaneous attack is part of the plugin’s quirkiness, what makes it unique. R4 has an identity crisis. It has a limiter’s architecture, but it’s tuned for distortion. The precursor was called “envelope flattener” and wasn’t framed as a limiter. It was just a way to smash sound, and a fast attack suits that.

What music do you make now, and what do you reach for r4 to do in it?

My favorite music is dark drum & bass from the late 1990’s. I’ve been teaching myself to DJ, and recently made a mix of my favorite tracks from that era. Since I’m a novice, my mix level was uneven in some spots. I threw r4 on it with a long release to smooth out the levels, and it worked. Its distortion is low enough for that application, where absolute fidelity isn’t necessary. A little distortion suits the music well.

Aside from that, I reach for r4 when distorting subtractive basses or adding oompf to drums. It works well with the production techniques from that era, which often used hard clipping from a mixer or inside of a sampler. But r4 goes beyond hard clipping, because of its variable knee. I wanted to find a balance between simplicity and generality.