It's Searle's argument, which I did not fully appreciate as an AI pup. Consciousness, whatever it is, is a real, physical phenomenon, and therefore must arise out of real-world physics.
Interpretive symbol pushing is not such a process. We don't know what part of the brain's physics gives rise to consciousness as a real, physical phenomenon, but it isn't from an abstract interpretation of electrons or chemicals.
Note I am not saying you couldn't simulate consciousness, but said simulatiou would not actually be conscious in the sense of experiencing the greenness of green. I also don't think a perfect simulation of physics as we know it would give rise to simulated consciousness even. There is something else we are missing.