Carl Bergstrom's presentation from Microsoft's Big Scholarly Data workshop held in Redmond 7/10/2015. The presentation makes the argument that the scholarly community is far too dependent on Google Scholar for the vital task of searching the scientific literature. Google scholar uses an unknown algorithm on an unknown corpus and allows neither extension nor community development. Moreover, it may not persist forever. As a step toward a solution, I propose that the scholarly community take charge of its output. Creating an open repository of full text for all of scientific publication faces copyright issues that will be intractable in the short term, but creating an open citation graph for all of science is feasible today. At the Eigenfactor Project, we specialize in using citation networks to reveal the scientific landscape. I provide a series of quick illustrations of the kinds of things we can do with large-scale citation data.