Introduction

This statement aims to provide a transparent overview of the use of collaborative AI tools in the production of this website, and its content. It aims to serve as a guide that enables curious visitors to quickly understand the boundary between my own efforts and the that of technology.

The Transformative Power of LLMs

I embrace the transformative power of Large Language Models (LLMs) and other Machine Learning (ML) tools. These technologies represent a paradigm shift in how we interact with information, and I believe in integrating them thoughtfully into my creative and technical workflows.

Acknowledgment of Limitations

While these tools are powerful, they are not infallible. I acknowledge that AI-generated outputs carry an error rate, including the potential for factual inaccuracies or biased reasoning. Consequently, no AI-generated output is ever published or deployed here without human approval; I remain responsible for all content and code.

Editorial & Content Creation

I make use of LLM's and other ML tools to a varying degree during content creation, and strive for a correct balance between human output and productivity.

  • Articles & Essays: These are written to consolidate my own thinking and learning. The words are my own. I may use an LLM to assist with "writer's block" or to research specific topics, but the core narrative remains a personal reflection of my opinions.
  • Technical Assistance: I utilise ML tools (such as Apple Intelligence) for essential grammar and spelling checks.
  • Non-Article Pages: For structural pages, LLMs are often used to assist with scaffolding content, which I then edit and refine to ensure the tone and facts are correct. The scaffolding process typically involves supplying a base narrative, headings and raw thoughts as input, and requests to elicit further refinement through a series of Question and Answer rounds whereby the LLM challenges my thinking.
  • Commitment to Transparency: For article pages if a specific post deviates from this standard, I commit to providing a transparency statement at the end of that post.
  • Translations: The German translations on this site are AI-assisted. As my personal German language proficiency is at a B1 level, I review all translations for correctness. For words or phrases I am unsure of, I consult tools like Google Translate and online dictionaries for verification.

Engineering & Site Development

The codebase of this site is heavily AI-augmented.

  • Design & Brainstorming: LLMs are used to discuss requirements, outline architectural designs, and manage GitHub issues.
  • Code Generation: While I oversee the entire architecture, LLMs generate the vast majority (>85%) of the raw code. I use a combination of Gemini, skills, and worktrees to manage this flow. It is acknowledged that this is only possible because of the efforts of the human maintainers of the modules this project depends on; to whom I am truly thankful.
  • Testing & Verification: LLMs help create the testing suites for this site. However, every test and every line of code is manually checked, altered, and verified by me before being signed off.

Updates to this Statement

LLM's, ML and the practices surrounding the use of these technologies are in a continual state of progression and flux. I continue to test new ideas and alterations to my workflows. The fundamental principals in this document remain, and any major deviation from them will be recorded in updates to this document.

Questions or Concerns

If you have specific questions or concerns please get in touch with me at hello@charliewoolford.com