← SpoonGrid

AI Transparency Statement

How I use “AI” for SpoonGrid

To support you in making informed decisions about what technologies you use in your life, I want to be transparent about how I do and do not use large language models (LLMs, also commonly known as AI) for SpoonGrid.

Things I do not, and will not, use LLMs for:

  • My original app concept, architectural direction, product or UX designs
  • Artwork, graphic design, and iconography of any kind
  • Branding assets (logo, etc) for SpoonGrid, SpoonOMatic or related projects
  • Copy in the user interface
  • Editing/rewriting my own writing directly (though I may have LLMs review my work, I make my own edits)
  • Handling, analyzing or storing user data
    • I commit to never using LLMs to ingest user data in SpoonGrid, except as part of future features that would be strictly opt-in with detailed disclosure and informed consent from the user.
    • I have no current plans to develop any feature where LLMs ingest user data, but I include the possibility here since some users may eventually want it, and I haven’t ruled out the possibility of developing such a feature.

LLM tools I currently use and their purposes (use does not constitute endorsement):

  • Claude / Claude Code (paid)
    • validating and refining my designs (both UX and code architecture)
    • scaffolding implementation plans
    • building documentation
    • validating and refining ideas, pitches and concepts, which I then write by hand
    • Code review and refactoring, e.g. for quality and performance
  • Cursor (including Claude’s Sonnet 4.6) (paid)
    • planning and executing code
    • may include minor copywriting, e.g. accessibility hints
  • CodeRabbit (paid)
    • reviewing code for bugs and quality gaps
    • generating fixes and docstrings
  • Other uses I’m currently considering using LLMs for:
    • Screening for security vulnerabilities in my code
    • Other infrastructural and code-deployment purposes

My choices to use LLMs

I have been a software developer for over a decade; I entered the field out of a combination of interest and economic necessity after being unable to find work in public policy research after getting my Master’s degree in that field. Developing software has never been a morally neutral enterprise; software has always had replacing human labor as one of its goals. The business environment around software has always had a number of serious ethical problems, from the amount of waste (energy and otherwise) it generates, to the way the venture capital system shapes how businesses are run and whom they benefit. I do not defend software as an industry in of itself, and I do not defend the use of LLMs generally.

However, I still have to make decisions based on my own abilities and constraints, and at this time continuing to work in software is my only viable option for making my own income. Software developers these days are expected to understand and use LLM tooling in their workflows; it is part of almost all job requirements. When beginning development on SpoonGrid, I made the decision to learn and use LLM tooling, so that my experience creating it can serve as job experience using tools and skills that are sought-after by employers.

I am benefiting personally from using LLMs because it helps add skills to my resume. I also benefit directly from my LLM use because of the speed at which it has allowed me to build my software and get my app to market. Coding with LLMs is vastly more efficient, which is both more convenient for me, and more accessible for me as I have cognitive fatigue and other chronic illness symptoms that make the assistance really helpful. I fully acknowledge that my use of LLMs and other harmful softwares (social media and other major consumers of data centers, e.g.) is not on a purely “necessary” or survival basis; however, at this time, I believe I am making the best decisions available to me, and I will continually learn and re-evaluate my decisions as my app and conditions evolve.

The one bright line I feel I can draw with my use of LLMs (“AI”) is that I do not use it to replace or supplement my own (or others’) artistic output. That includes my app concept, product and UX design, app and marketing copy, and any visual media such as iconography or brand imagery that I produce. The same is true for any creative writing I produce on Substack or elsewhere (and the text you are reading now). The icon sets I’ve used in the app are Huge Icons and Health Icons. Several of my icons I have modified by hand, as allowed by the icon license, using SVG software. I can’t guarantee that no AI was used in producing those icon sets, but they have both been around for a while and Health Icons has a substantial open-source ecosystem behind the design and deployment of their icon set.

I understand that not everyone will find the choices I make about using LLMs justifiable, which is why I am choosing to be transparent. There are many valid reasons to boycott LLMs and technologies that were built using them; I want to make sure that people can be informed about the decisions I have made to use LLMs so that they can decide whether SpoonGrid is something they are comfortable using.

Thanks so much for your time in reading this, and for considering using the SpoonGrid app. If you have any questions or concerns, you can reach me at contact@spoonomatic.com .