Nearly every major bacterial lineage studied to date possesses strains that produce protein-based toxins that are thought to function as anti-competitor compounds. Toxin production occurs stochastically and often involves lysis of the producing cell. Such lysis can be seen as a form of indirect altruism because while the producer dies, its latent clones are immune to the released toxin and benefit from having other susceptible competitors killed. As with other forms of altruism, free riders exist in this system as well—cells that are resistant to the toxin, but do not produce it. In this talk, I will explore various aspects of the ecology and evolution of these toxin-producing communities. I will start with a general model to investigate conditions where altruism (or spite) can be maintained. We will then move to a specific case involving the eco-evolutionary dynamics of a community of toxin-producing Escherichia coli. I will review some older work on non-transitive (rock-paper-scissors) coexistence and discuss the evolution of competitive restraint (another form of altruism). I will then turn to some current work on the origins of novel toxins using a directed evolution approach. This pursuit of origins has led to a new evolutionary hypothesis, as well as some unusual perspectives on the molecular underpinnings of this system.