Most of the wild animals in the world live in the ocean. But they are hidden from sight, and hard to sample scientifically, which has made it difficult to estimate their distribution and abundance. In this talk I will discuss how the intense sampling of the ocean by industrial fisheries offers an unparalleled resource to provide a global picture of marine fish abundance. By comparing with process models of fish production and dynamic fisheries, satellite observations and catch records can constrain global maps of fish growth and biomass. I will show how these provide new insights for the role of fish in ocean biogeochemistry and carbon storage, as well as prompting the idea that the growth of marine fish can be limited by iron availability. They also help understand the full ocean size spectrum, confirming the theoretical Sheldon spectrum, and showing the depth of the impact that industrial fishing has on the ocean.