Learning-based algorithmic agents increasingly mediate major markets and online platforms, reshaping their dynamics and economic outcomes. This transition raises foundational questions about how decision makers should interact with such systems and how learning agents affect market behavior. In this talk, I will discuss some of these questions and present results from the paper Markets with Heterogeneous Agents: Dynamics and Survival of Bayesian vs. No-Regret Learners. The paper studies competition between heterogeneous learning agents in asset markets with stochastic payoffs. We compare Bayesian learning and no-regret learning and analyze how these different approaches interact in a competitive environment. A central finding is that regret plays a key role in market selection, but low regret alone does not guarantee survival: an agent may succeed in achieving even logarithmic regret and yet be driven out of the market when competing against a Bayesian learner with a finite prior that assigns positive probability to the correct model. At the same time, we show that Bayesian learning is highly fragile, while no-regret learning requires less knowledge of the environment and is therefore more robust. Motivated by this contrast, we demonstrate a simple balanced strategy that incorporates Bayesian updates while improving robustness and adaptability to distribution shifts, taking a step toward a best-of-both-worlds learning approach.