The Question Concerning Large-Language Models

In 1954, Heidegger's "The Question Concerning Technology" warned that technology's true danger emerges when we mistakenly view it as something "neutral," urging rigorous questioning of our relation to emerging technologies. Today, nearly 75 years later, technological development has spiraled to heights likely inconceivable in Heidegger's time—most recently with the arrival of artificial intelligence. This paper revisits Heidegger's seminal work to undertake a phenomenological analysis of large-language models (LLMs). I begin in Section II by expounding Heidegger's phenomenological method and unpacking his key terminology, clarifying what it means to question technology's essence. In Section III, I explore Heidegger's claim that technology's essence is a "realm of revealing," adopting this as my guiding framework for investigating LLMs. Section IV argues that LLMs, as a realm of revealing, disclose our relationship to language and understanding—what Heidegger explores through "discourse." In Section V, I unpack Heidegger's theses on authentic and inauthentic discourse, explaining why the degree to which discourse is inauthentic correlates to how thematic it becomes. Finally, in Section VI, my investigation culminates in the conclusion that the essence of LLMs is idle talk. My central aim is not merely to argue that LLMs exemplify inauthentic discourse, but to suggest that their proliferation reveals—and is enabled by—a "flattened" understanding of language, one that reflects the pervasive influence of what Heidegger calls the they.