You're absolutely correct. I read the article quite fast, and assumed they are AI, albeit not LLM powered systems as well. I'm using computers since I can read

You're absolutely correct. I read the article quite fast, and assumed they are AI, albeit not LLM powered systems as well. I'm using computers since I can read, and when somebody says "traditional OCR", I think about the older systems like Tessaract or ABBYY's FineReader which can be again automated for batch processing, albeit mostly locally. Sending huge amount of PDFs to a cloud server to get them processed is still a bit alien to me, since it can be done on-premises (or on a VPS with the said software) very efficiently from my perspective.