Agent-based modeling is a powerful technique that allows modeling social phenomena ab-initio or from first principles. In this paper, we review the history of agent-based models and their role in the social sciences. We review 62 papers and create a timeline of developments starting from 1759 and Adam Smith into the recent past of 2020 and efforts to model the Covid-19 pandemic. We reflect on model validation, different levels of model complexity, multi-scale models, and cognitive architectures. We identify key trends for the future use of agent-based modeling in the socials sciences.