Our Policies
DRAFT ARTICLE
In the interests of transparency, New Quaker Ministries drafts all of it’s
articles in the open. We are a non-profit, independent ministry, funded entirely through small individual donations
and contributions. While an article is in DRAFT, it is unfinished and is likely to change rapidly.
Committment to the Testimonies #
At all times, and as many ways that we can reasonably do so, we will stand true to the Quaker testimonies; those of simplicity, peace, integrity, community, equality, and stewardship.
- Simplicity:
- We will reach for common and simple language that is easy for others to understand.
- We will provide multiple pathways to learn the concepts we teach.
- We will make engaging with our content accessible, and intuitive.
- We will avoid unnessesary embelishments, staying focused on the main topic and points.
- We will avoid elitist or devisive language, which desprives others of the respect they deserve.
- Peace: We will not advocate for the use of violence, war or bloodshed for any reason.
- Integrity: We will act in the interest of truth, honesty, respect, and integrity at all times.
- Community: We will honor the need for community.
- Equality: Jesus was “woke”, and so are we. We honor that of God in everyone. We love all equally. We seek to end inequality in all its forms. Stop asking what your idols would do. Open the Bible. Read Jesus’ teachings and then ask “What would Jesus Do?”. Destroy your idols.
- Stewardship: We feel a responsibility towards everyone to present the truth in the best possible way. Sometimes presenting uncomfortable truths to others can make them feel threatened, and we feel that. We do. That is why we seek to teach without judgement towards others.
Research Methods #
I am committed to truth and integrity in our articles, and in all our dealings.
The purpose of New Quaker Ministries is to present the truth and light to others, to invite new people into Quakerism, and to do so while honoring the Quaker testimonies. We are here as educators, and teachers in the Spirit. We take great care to draft our articles using the best research methods we know. Thus, our articles are the results of intense focused independent research based on critical exegesis, the historical method, the scientific method, and verifiable, published, and peer-reviewed archaeology. We fact check and cross reference everything stated in our articles, and provide a robust bibligoraphy, so that you can verify our claims. We won’t always get it right, but we try our best.
Transparency & Publications #
TODO: Section on how DRAFT ARTICLES work.
AI Use Policy #
Understanding Raymond’s Background #
Before becoming a public minister, I worked in technology for over 18 years, some time of which I’ve spent working in machine learning and artificial intelligence. I worked in the “AI Apps” team at DataRobot for about a year and a half as a Staff Machine Learning Engineer. I’ve given talks on reinforcement learning at conventions. I currently lead a small cybersecurity team at a major bank.
I am an expert in software development, technical leadership, cybersecurity, generative AI, reinforcement learning, evolutionary algorithms, statistical modelling, machine learning (classification, regression, clustering) among many other related topics one would stumble across in an 18 year career in software, technology and startups.
I have published two tabletop board games, both of which use extensive handmade artworks created by amazing people. I understand the need for copyright holders to protect their works from copyright laundering and I understand the need for artists and other storytellers to make an honorable wage from their work.
We do not use AI in these ways: #
- We do not use it to generate content for articles.
- We do not use it to generate ideas for content for articles.
- We do not use it as a way to mass format content; to edit and “fix” content in-place.
- We do not use it to perform fact-checking activities for our articles.
- We do not use it to generate images without also disclosing how they are generated.
- We do not use it to generate images for any use other than storyboarding, prototyping, or placeholder illustrations.
- We do not use it for placeholder illustrations intended for articles unless an associated Bounty is also placed.
- We do not use closed source models or proprietary vendors like OpenAI, CoPilot, or Anthropic.
We do use AI in these other ways: #
- We do use it as a general information search engine, no different than Google. Just like with Google, these results are cross-referenced and critically examined.
- We do use it to summarize our own articles to present summary text on list pages on the website. These summaries are then edited by a human.
- We do use it to generate a list of tags for articles we’ve written to make them more searchable for others in the future. These tags are also edited by a human.
- We do use it as a “junior peer reviewer” for articles, to check if the article is aligned with the Quaker testimonies. We then examine the output and discern if we should listen to it or discard it. Sometimes it surprises us.
- We do use it as a “junior copyeditor” to make editorial suggestions for improvements to articles we have written, aligned to the Chicago Manual of Style. Again, sometimes it surprises us. These are merely a list of suggestions. We do not ask AIs to “rewrite” or “edit” articles in place.
- We do use it to generate images for the purposes of storyboarding, prototyping, or for creating placeholder illustrations in our articles until the associated Bounty is filled.
- We only use open source and open weight models.
Raymond’s AI Tooling: #
-
I run all my software on Ubuntu 24.04 LTS, except for the Affinity Suite of desktop publishing tools, which I run on Windows 11. This website is primarily edited inside the vim editor, which is a terminal based application that developers use to write code. You might find something like VSCode more accessible. Since I am familiar and productive in this tool for my everyday career, I also use it when writing articles for this website.
-
I have a 16-core AMD Ryzen 9 5950X CPU, 128 GBs of RAM, and a single MSI Ventus Black 8GB RTX-4060 graphics card. This is a commodity “gamer grade” desktop computer. I built my desktop by hand. It cost me around $2,000 in parts and a weekend of effort to put it together. My point is, you can run these tools at home on a roughly similar machine.
-
I use Ollama to serve my LLM models locally. At the moment I prefer the output I get from Gemma3:27b and from Qwen3:32b. I am playing around with GPT-OSS:20b. I use these models to summarize articles, generate tags, and make editiorial suggestions at the most basic level (e.g. spelling, grammar, and manual of style).
-
I use ComfyUI as my workbench for working with generative models. At the moment I mostly use HiDream-I1-DEV, and Flux.DEV to generate images, either to use as placeholder illustrations, to storyboard ideas, or to create prototypes. All placeholder illustrations that I create with AI will also have an Art Bounty placed on them. You can read more about our Bounty program here.
-
I actively avoid the use of AI systems that I do not have direct control over. I avoid products that “integrate AI” into them. For example Notion, or AI suggestions in Google Docs, or CoPilot suggestions in office applications. It is part of the reason why I stick to vim as a text editor. I do not trust that AI platforms have my best interests or the best interests of society at heart. I see them as a terrible evil and a tyrranical entity at scale. I actively resist the use of any algorithm (AI or otherwise) that isn’t free (as in beer and as in freedom).
I do not pay for any AI subscription services, I do not purchase AI generated artwork, and I do not use any APIs provided by generative AI companies for the purposes of this ministry. I’m not down with the vibes.