Life is too beautiful a thing to take for granted, so remember that you too must die and enjoy it while it lasts.

Spotify

Time Out

Where two frequencies meet and hold still. Built for those suspended moments when thought becomes liquid, that particular stillness where work becomes meditation. Part of an ongoing exploration into music as cognitive architecture.

Standing Wave

Music that shapes itself around thought, built for those hours when focus needs somewhere gentle to live. Created for the specific frequency where work and sound converge without collision.

Soft Architecture

Fragile light through borrowed time, where jazz theory meets electronic floors in that particular frequency between memory and forgetting. Built for the hours when reality feels negotiable and sound becomes architecture.

Paper Moon Windows

Those soft, blurred moments between 2 and 5 AM when time moves differently and everything feels wrapped in analog warmth. Music for when the world finally quiets down enough to hear yourself think.

Velvet Hours

Funky, experimental but in that way where it still sounds pleasant. Hope you like it!

Slow Motion Paradise

Something VILLAINOUS stirs in the algorithm. A homage. To a great villain and some algo advice for people less lazy about optimizing their popularity than myself.

ALGORITHMIC VILLAINY

Ever feel like you weren't born for this life? It's because you weren't.

Paleolithic Panic

I work with Applied AI for a multi-national mega corporation. I initially started experimenting with AI music generation in my spare time to see if I could use that form of feedback to learn more about how these machines think and interpret our inputs and use that knowledge in my day to day work.

Quickly however, I fell in love with the ability to create music. I found that with a "little" bit of research, I could make just about anything (hence the...eclectic or rather, chaotic range of genres you will find if you dig back in my upload history).

I'm still learning a lot of very useful things for my day job in working on this. But that isn't why I do it. I do it because I love it. The lyrics are written by myself, in concert with LLM's of course. I build synthetic persona schemas to unlock novel areas of latent space or prime LLM's on deep context about specific points of history. Data mining old, forgotten sailors journals to iterate on ballads about Nantucket or research the god complex of surgeons to help me write songs like Time Out.

For the music I focus heavily on merging my own personal tastes, with established music theory. Building my own text-books on music theory that I iterate on to create and evolve my own distinctive sound that will hopefully resonate with others as well.

Some people think AI music is random. Bad AI music is, in a way and you will hear no protests from me on the problems of AI slop. There is a lot of AI slop out there. But the majority of songs you hear here, are the result of one or two generation runs once I get the prompt correct.

So am I going to automate musicians out of a job? Is there a pipeline ready to churn out thousands of hit songs per day?

No. I could probably automate mediocre music production, but I don't want to. Most of my tracks take a lot of work, inspiration, iteration and thought. My AI's couldn't come up with an interesting concept to save their lives. An AI if left alone, will generate a mediocre song about coffee or gravity (see my track, Gravity which is an experiment on this exact thing).

I'm just making stuff I think is interesting, cool or beautiful. Enjoy. And feel free to check out my slightly more focused Soundcloud under "Theorem Blue" or my Spotify aliases, Vectorfall, Eigentide, Manifold Soul or Unbound Symphony.