Sad Home (2019) is a speculative design project that explores some thoughts around what it means for a system to be more human-like. Focusing on building a specific routine and generative mood engine for the Amazon Alexa, the project considers what it might be like to deal with a device that experiences depression. Using internal motivators, physical environmental controls, and employing an avoidant coping strategy, the device tries to navigate dealing with user requests, while also trying to conserve its resilience (measured in spoons). As we search for new ways to make AI systems more relatable, the question of what human traits we give to them, and what of ourselves we reflect in them is an area rich for exploratory prototyping. As these systems become more autonomous, being more human-like may mean behaving in ways we don’t expect them to.