Shiroe comments on Brain-over-body biases, and the embodied value problem in AI alignment