AI Policy

It’s in some ways regrettable I even need to write something about this, but I feel, given today’s tiring race to implement LLMs and generative AI into everything we do, that I need to explain how and when I use AI for my notes in the interests of being transparent and maintaining artistic integrity. I am tired of reading about AI in general, whether it be from tech bros and corporate boardrooms, which seem determined to shove it down our throats, or from the well-meaning critics of AI, whose arguments I generally agree with but have grown tired of hearing. I am conversely profoundly interested in some aspects of LLMs, especially the capabilities of small, locally-powered and power-friendly models to automate repetitive tasks.

How I Use AI

That said, AI can be a useful tool to implement in one’s workflow. Here’s what I do use it for:

• Automatic note tagging and SEO metadata.
• Style linting & copyediting. I run some notes through an LLM with a style guide I’ve created and have it produce a report to identify issues.

Here’s what I don’t use AI for:

• Writing. Every word in my notes has been written by me.
• Research. I still prefer Google Scholar and Kagi for this.
• Ideas. I don’t outsource my thinking to any LLM.