Uh, no. If what you asserted were the case, then things like cortical visual prostheses would not be a possibility. I suggest you look them up. While they are still under development, they most certainly do intend on creating fine-grained control over neural activity. Same for cortical somatosensory prostheses. On the flip side, we can definitely read-out fine-grained information about neural activity, such as is used for motor system prostheses, ranging from limb prostheses to vocal chord prostheses.
On both the read-out (decoding) side, and the driving (encoding) side, we have the ability to receive and transmit information on an individual basis. Yes, there is a lot of variability, but that is part and parcel of the challenge. Just as individual variation in foot size and shape does not preclude the creation of shoes because there is an underlying structure, so individual variation in brain morphology and wiring is unlikely to preclude creation of brain/machine interfaces because again there is an underlying structure.
Any time you hear someone say it is impossible to do something, it's likely they are just not thinking in advanced enough terms to overcome whatever barrier they perceive. I myself am guilty of such mistaken proclamations.