A description of when I how I make limited use of AI.1
Content generation
I do not use generative AI to outright create content for this site.
- Sometimes AI forms part of my research process to help me gather alternative viewpoints.
- I do sparingly use Microsoft Copilot AI to help me write a paragraph where I know the idea I’m wanting to get across but I can’t find quite the right words. Everything generated this way is reviewed to make sure it conforms to my initial intentions. These two uses are no different from the role an editor would perform.
I sometimes use AI handwriting recognition to convert handwritten notes into text for further processing.
Coding
Javascript with Typescript is a relatively new programming language for me and I’m leaning heavily into Microsoft Copilot over general web search to help me with tricky coding problems. Most often by pasting in my code with “fix this”. I always review the code returned to make sure it conforms with what I need.
Photo editing and display
I sometimes use AI in the photo editing process to assist with tasks that could otherwise be done by hand. Within Affinity Photo this is currently limited to the Inpainting brush tool
I am using a local version of LLaVA 7b model via Ollama to generate alt-text and suggest keywords for photos in My photo gallery.2