The discussion revolves around the claim that large language models are inevitable and how that framing shapes the debate. Practitioners repeatedly treat the persistence of LLMs as a given, assuming the technology cannot be rolled back once embedded in workflows. They also take for granted that investment will continue despite the lack of a clear business model, framing compute and methodology as the only remaining bottlenecks. A common thread is the view that adopting LLMs is essential to stay competitive, and those who question the premise are labelled as deniers.
I notice that the participants treat inevitability as a comfortable premise, allowing them to sidestep deeper questions about economics or ethics. It is striking how quickly the conversation moves from accepting the premise to debating implementation details, as if the former were settled fact. The prevalence of analogies to past technologies feels both predictable and a way to legitimize the current hype. What surprises me is the extent to which the term “inevitability” is used as a shield against criticism, rather than a hypothesis to be tested.