A statement to (current) limits in ML models of tissue examination Surprising for me because I thought they were improving for xray imaging, and assumed the same was true for histology.