How can machines and minds be related? The analogy between machines and minds has profoundly influenced cognitive science and AI research. Before delving into how machine models can be used to simulate mind in the field of AI research, it is important to clarify typical category mistakes in understanding minds. Clarifying category mistakes about what the mind is helps to orientate the practical goals of AI research and development. After that, this chapter will explore whether and how these category mistakes could be dissolved by the main approaches in AI research. Analogies between minds and machines will be discussed here, in particular the analogy between neural and circuit systems in cybernetics, the early functional analogies between minds and Turing machines in symbolic approaches, and the connectionistic analogies between brain plasticity and the learning capacity of artificial neural networks in sub-symbolic approaches. An examination of these approaches suggests that understanding minds in a naturalistic framework alone is insufficient for developing thinking machines. A normative framework is needed to articulate the conditions under which intelligence can be recognized and explained as taking place.