Mac Shine and I talk about Mac’s recent intriguing opinion paper that may have radical implications for systems and clinical neuroscience. In it, the thalamus mediates between feed-forward type input from cerebellum, sensori nuclei and cortex one one hand and input from the basal ganglia that introduces an element of randomness. By projecting to the cortex in a specific manner, the thalamus can recruit these inputs to shape the attractor landscape of cortical activations. Mac develops this a theory from the cell- to the systems neuroscience level and hints at how Kahneman’s system I and II levels of thinking fast and slow could be implemented in the brain. The theory radically extends and partly opposes existing concepts such as the thalamus as a mere relay station and the model of the basal ganglia for action selection proposed by Alexander, DeLong and Strick in 1989 – so there is vast potential of this becoming transformative for deep brain stimulation, as well.