Backblazes own analysis of their testing mostly debunked the Seagate stuff that the internet latched onto from their raw results. Backblaze doesn't do scientific testing in controlled environments, they just provide data from their datacentre use, which often has various generations of server racks & enclosures, uses the cheapest price/GB drives possible and submits them to an environment the drives were rarely intended for. IIRC they generally used consumer Seagate drives more than other manufacturers due to the price disparity.