While true, it is also irrelevant. All known antibiotics belong to a very small number of classes by their mode of action. Resistance genes often confers resistance to not just one antibiotic, but an entire class, for example mecA in MRSA. Therefore it is possible for a bacterium to develop resistance to all therapeutic antibiotics without ever having encountered the exact types.
What's more alarming is that bacteria frequently exchange genetic material between species. Just because the resistance genes have not been discovered in a pathogen does not mean that the gene cannot be easily and instantly transferred from another species, and the chance of that happening is greatly increased if the environment is flooded with resistance genes due to selection by human use of antibiotics.
Just because it hasn't happened yet is not a reason not to do something. A tornado hasn't wiped out my house yet, should I move because it might happen some day? Of course there is a risk associated with their use, but the estimated risk is miniscule. It's not even measurable yet. Let's worry about problems we can identify and solve rather than wasting time fretting about what may be.
If by "risk" you meant the horizontal transfer of resistance, then I point you at genomic studies of horizontal gene transfer across bacteria. It happens with such regularity that the concept of a species cannot be naively applied to bacteria.
You're correct, feed grade antibiotics aren't used for humans. The use of antibiotics that are used in human therapy for feed grade antibiotics is already outlawed by the FDA.
