There's a huge difference between feeling pain and reacting to demage. Humans can react to injury in a very complex ways without feeling pain. And buildings too - when you crash with an airplane into a building, the building reacts in a complex way. Does it feel pain? Feeling pain is something fundamentally different than reacting to demage.
If feel only fits your narrow, human centric definition of it, then of course machines can't feel pain. There isn't much point in having a discussion if the discussion is about your understanding of a word though.
If you go by the broader definition of feel, which includes "to have a sensation of something", then I'd argue that the line is far more blurred.
If you accept that a rat is capable of feeling pain, then surely you must accept that a functionally identical machine with a neural network functionally the same as a rat's brain can also feel pain? There's no real fundamental physical reason stopping us constructing a computer which functions in the same way as a rat's brain does. Does the computer we constructed suddenly stop being a machine?
My understanding of "feel" is purely subjective. I think it's impossible to create an objective definition. That's the problem
"To have a senseation of something" isn't really helpful definition, because "have a sensation" is just another word for "feel". It's like saying that the definition of "to buy" is "to purchase".
Physics cannot, even in theory, explain my subjective experience of pain. Why do I feel pain, when my brain's atoms are in specific positions?
