By combining artificial intelligence (AI), machine learning, and microfluidics, researchers have designed an autonomous decision-making synthesis system that chooses optimal reagents, synthetic routes, and reaction conditions to make valuable products with targeted properties. Dubbed “Artificial Chemist,” the system uses tiny quantities of precursors and does not require experienced users to monitor and guide experiment choices. The system, which learns from successive experiments, can reduce the time, cost, and quantity of waste associated with traditional R&D synthesis (Adv. Mater. 2020, DOI: 10.1002/adma.202001626). In a proof-of-concept study, Robert W. Epps and Milad Abolhasani of North Carolina State University and coworkers used the system to make a series of uniform colloidal perovskite quantum dots with preselected emission profiles and other electronic properties. These materials are used in solar cells, light-emitting diodes, and electronic displays. Fitted with syringe pumps, micromixers, and built-in spectroscopy probes, Artificial Chemist reacted cesium lead bromide, zinc halides, and organic capping ligands, then analyzed the products, modified the conditions, and repeated the experiment. The team found that at the outset, the untrained system took about 1.5 h, or 25 experiment runs, to identify optimum synthesis conditions. After the initial run, the system cut the prep time for making new products with other target properties to roughly 15 min.