GP's point is that the goalposts of what counts as "AI" have been constantly shifted for decades, and right not it seems that even many HN commenters seem to have somehow gotten the idea that "AI" strictly means "deep learning neural networks" (or something even more specific like "generative transformers"). GP just went a bit too far in the opposite direction, being too inclusive in their definition.

FWIW, my working definition of "AI" is "anything discussed in the latest edition of Artificial Intelligence: A Modern Approach". It seems to work well.