Last month I produced a proof of concept zero shot categoriser for images using OpenAI’s GPT-4 Vision-Language model (VLM) capabilities. However this had the disadvantage…