principles

Last updated 2026-09-04

My primary goals as a creative person are creative freedom and sustainability above all else. I am not motivated by rapid growth, but rather a small invested audience of patrons, fans & collaborators.

I try to stick to the following principles with all my creative work in service of the above:

  1. Do no harm, do not enable harm
  2. Minimize service lock-in
  3. Minimize money given to services/platforms misaligned with my values
  4. Keep costs & maintenance time as low as possible
  5. Strength in numbers: support & collaborate with similarly-minded artists & service providers
  6. Hold myself accountable by building & sharing in public

When I can, assuming my projects are self-sustaining, I try to support causes that I care about. So far I’ve raised $485.10 this way:

release raised benefiting
REMIXED CASTLE $231.10 a struggling trans creator in the Music Threads community
THREADS ON MY ART, vol. 2 $100.00 ACLU
THREADS ON MY ART, vol. 1 $106.00 the Trevor Project, World Central Kitchen
hive mind(s) $48.00 a struggling trans creator in the Music Threads community

stance on music streaming

I generally consider the streaming business model hostile to the ability for the average music artist to sustain a living, given the marketplace dynamics it enables. I think of it as having democratized radio (warts & all), and not music discovery.

I aim to minimize the presence of my music on streaming to only:

  • single releases, aimed to promote Kid Lightbulbs and tease my other releases as offered on my website, Bandcamp and other more artist-friendly platforms
  • a singular album, ideally never my most recent one, to highlight the fact that I make albums without commoditizing them entirely

As of July 2025, two of my albums (RUINED CASTLE and REMIXED CASTLE) are available on streaming services, and a few singles meant to promote my other albums. I will not release albums on streaming platforms moving forward, and am considering withdrawing the rest of my catalog.

I take significant issue with Spotify’s dominance over the habits of casual listeners, exploitation of artists, and use of its profits. I neither pay for any of Spotify’s services nor promote it as a destination for my music. However, at this time I have not withdrawn my catalog from it, mainly to not restrict access for potential listeners (see principle 4 above). However, if over time Spotify’s market share decreases, I will likely remove my music from Spotify’s platform fully.

stance on generative AI

Like many artists today, I grapple with what to do about “generative AI”. This is where I’ve landed as of this moment:

music & art

  • I consider generative AI in art a form of cheap commission or direction of that art, not creation, which exploits the work of billions of human minds across time. I recognize the appeal of generative AI for music, though I consider it culturally destructive and largely extracts money away from artists & art supporters and toward wealthy capitalists.
  • There is absolutely no use of generative AI in any of my artistic output. This includes: all music I write & release, any writing you read on this site, and any content you see on my social & marketing channels.
    • The one exception to this is the cover art for my first album, THROW MYSELF INTO THE BAY, which was generated using an early version of the Stable Diffusion model as an experiment with using LLMs on local devices. I liked the output, but do not intend to use it again.
  • Given the above statements, and my stance on music streaming, I do not collaborate with artists who rely on generative AI to “create” music or visual art.
  • I do sometimes use two predictive & machine learning-based tools in the production of music, each available in Logic Pro:
    • Session Drummer (to stage drum patterns, which I usually then manipulate by hand in MIDI format)
    • Flex Pitch (to correct odd inflections in my voice)

at work / coding

  • My above statement on generative AI in art holds here. That said, I recognize that the ethos of the Internet, software development, and open-source software & programming languages means that the general perception of AI in these contexts is somewhat more acceptable to the creators of these things. Therefore I feel comfortable using generative AI to write code, but I do not consider myself a software developer.
    • I appreciate the vast amount that I’ve learned about software development via my usage of AI.
  • I use Anthropic’s Claude exclusively in a software development & product management context for my day job, the cost of which is covered by my employer. I use Claude in the following ways:
    • To rapidly prototype and/or build refinements & features to improve the Buffer product, with the goal of saving time, minimizing distraction for my teammates and building clarity
    • I occasionally use it as a research and sparring partner – that is, I use it to capture & summarize information about the social products landscape and bounce ideas related to product strategy.
    • I also used Claude to build the initial version of Unstream, and occasionally plan out larger features for it, though neither Claude (nor any AI) is implemented in the app’s functionality. It is strictly not an “AI product” per se.
  • I have begun using open-source models via Ollama to replace my use of Claude for use cases outside of Buffer. I primarily use the cloud-hosted versions of GLM-5.1/5.2 via a lightweight research & coding agent, and Muse Glimmer 30b-mlx locally to periodically review Unstream code and tinker with other small side coding projects. I would like to eventually train my own models for specialized use cases.