Some people’s reaction to visual illusions is, “Here you have it. The proof that our brains are just a collection of hacks cobbled together!” Adaptive systems might use different processes to perform one task; when the results don’t fit neatly together, (for example, when our visual system oscillates between several versions of reality), we feel like we have been tricked. Yet these failures and discrepancies are a powerful tool to understand not only how complex adaptive systems work, but also to design models of these systems, and even design new adaptive systems that use “collection of hacks cobbled together” to efficiently deal with tasks in their environments. I argue that we should push biological systems to their limits, and that replicating these limits in artificial systems could be an efficient design process.