This article introduces the philosophical significance of large language models (LLMs) like GPT-4 in relation to classic debates in philosophy of mind, language, artificial intelligence, and linguistics. It argues that LLMs challenge long-held assumptions about artificial neural networks (ANNs) as cognitive models, but also highlights the need for further empirical investigation into their internal mechanisms. The article serves as a primer for philosophers, offering an opinionated survey of LLMs' significance in relation to classic debates. LLMs have achieved remarkable proficiency in a wide range of language-based tasks, often surpassing human performance in areas such as essay writing, dialogue responses, and standardized tests. They can also generate code, solve mathematical problems, and create creative works. However, their success may be attributed to a form of memory retrieval rather than true understanding, raising questions about whether they are merely "Blockheads" – systems that produce responses based on pre-programmed answers rather than cognitive processing. The article discusses the philosophical implications of LLMs, particularly in relation to the debate over whether they can be considered intelligent. It argues that while LLMs may not possess the same level of cognitive processing as humans, they can still exhibit flexible generalization and creativity, suggesting that their success may not be limited to mere memory retrieval. The article also explores the role of empirical evidence in evaluating the cognitive capabilities of LLMs, emphasizing the importance of experimental findings in resolving philosophical debates about their competence. The article also addresses the issue of compositionality, which is central to the debate over whether LLMs can be considered intelligent. It argues that while LLMs may not possess the same level of compositional structure as humans, they can still exhibit systematic generalization and compositional processing, suggesting that their success may not be limited to mere memory retrieval. The article also discusses the implications of LLMs for the debate over nativism in language acquisition, arguing that their ability to learn syntax without innate knowledge challenges traditional arguments for innate grammar. Overall, the article argues that LLMs are more than mere Blockheads, but that their philosophical significance lies in their ability to challenge long-held assumptions about artificial neural networks and their potential as models of cognition. The article concludes that further empirical investigation is needed to fully understand the cognitive capabilities of LLMs and their implications for philosophy.This article introduces the philosophical significance of large language models (LLMs) like GPT-4 in relation to classic debates in philosophy of mind, language, artificial intelligence, and linguistics. It argues that LLMs challenge long-held assumptions about artificial neural networks (ANNs) as cognitive models, but also highlights the need for further empirical investigation into their internal mechanisms. The article serves as a primer for philosophers, offering an opinionated survey of LLMs' significance in relation to classic debates. LLMs have achieved remarkable proficiency in a wide range of language-based tasks, often surpassing human performance in areas such as essay writing, dialogue responses, and standardized tests. They can also generate code, solve mathematical problems, and create creative works. However, their success may be attributed to a form of memory retrieval rather than true understanding, raising questions about whether they are merely "Blockheads" – systems that produce responses based on pre-programmed answers rather than cognitive processing. The article discusses the philosophical implications of LLMs, particularly in relation to the debate over whether they can be considered intelligent. It argues that while LLMs may not possess the same level of cognitive processing as humans, they can still exhibit flexible generalization and creativity, suggesting that their success may not be limited to mere memory retrieval. The article also explores the role of empirical evidence in evaluating the cognitive capabilities of LLMs, emphasizing the importance of experimental findings in resolving philosophical debates about their competence. The article also addresses the issue of compositionality, which is central to the debate over whether LLMs can be considered intelligent. It argues that while LLMs may not possess the same level of compositional structure as humans, they can still exhibit systematic generalization and compositional processing, suggesting that their success may not be limited to mere memory retrieval. The article also discusses the implications of LLMs for the debate over nativism in language acquisition, arguing that their ability to learn syntax without innate knowledge challenges traditional arguments for innate grammar. Overall, the article argues that LLMs are more than mere Blockheads, but that their philosophical significance lies in their ability to challenge long-held assumptions about artificial neural networks and their potential as models of cognition. The article concludes that further empirical investigation is needed to fully understand the cognitive capabilities of LLMs and their implications for philosophy.