Bias is a function of perspective. Every entity capable of generating output will be doing so based on "knowledge of" some set of inputs, even if that "knowledge" is part of its "hardware." It seems people are still wasting energy discussing "bias" in generative AI, so hopefully this ends the discussion as such. Obviously it is not possible to generate something without starting from copying something, this is the function of memories. In a physical system, this is t=n-1. Therefore it is not possible for a generative AI model to be free of bias. These discussions always seem to carry a moral connotation by using the word "bias" and relating it to hot-topic biases like racial and gender bias but the reality is that there's nothing necessarily immoral about a generative AI model's biases. Likewise, there's nothing necessarily moral about "fixing" these biases, because the reality is that "fixing bias" in a generative AI model is nothing more than realigning its bias to be more in line with the operator's own biases. This is an important distinction to realize because it helps to put the priority of the work in perspective. With a more accurate perspective of the priority we're better able to identify inefficient use of societal resources, such as taxpayer funding, postdoc researcher hours, and hours of general discourse.