Eliminativism, or eliminative materialism, is a radical form of physicalism that rejects the validity of folk psychology, claiming that familiar mental states like beliefs and desires do not exist. Instead, it advocates replacing them with concepts from cognitive neuroscience. Originating in mid-twentieth-century critiques of introspection and language analysis, eliminativism was later formalized by the Churchlands into the neurocomputational perspective, linking mind theory to brain-based computation. In the 1980s, Paul and Patricia Churchland proposed a paradigm shift—from representing mental states in a linguistic format governed by morphosyntactic rules to representing them in a neurocomputational format (i.e., vector coding)—which allows such states to be processed by artificial neural networks. The latter are capable of learning and performing generalizations, thus paving the way for neurophilosophy and for new philosophical research fields such as neuroethics and neuroesthetics, as well as for emerging sciences like neuroeconomics. Despite persistent criticisms, including charges of absurdity and self-refutation, eliminative materialism remains a viable framework. Recent advances in cognitive neuroscience and artificial intelligence (AI), especially large language models (LLMs), reinforce its promise as a computationally grounded theory of mind.