!brute_force utilises computation as a tool for critically addressing behavioural engineering within the human-dog-artificial intelligence paradigm. The starting questions of this practice-based research are: What can artificial intelligence learn from dogs? What would happen to the artificial neural-network plasticity if AI were made of the metacommunicative skills of a dog? The idea of dogs training machines and machines behaving like dogs predates machines. The ‘paradox’ of dog computation aims to point out that ‘the more obsessively we attempt to compute the world, the more unknowably complex it appears’ (Bridle 2018). We are comparing a trajectory of two relationships: Human Computer Interface (HCI) which has evolved over the last 40 years, and the human-dog relationship which has evolved over the last 40 000 years. The comparison of these two orbits opens up a new area of research: speculative Dog Computer Interface (DCI). As an open ended meta-dialogue between a dog and a machine, DCI may offer navigational tools to help us better understand the traps and highlights of the HCI paradigm. '