The emergence of Large Language Models (LLMs) has reignited longstanding philosophical debates concerning linguistic understanding, conceptual content, and the conditions under which a system may be regarded as a genuine language user. The article “LLMs and the Logical Space of Reasons” offers a sustained critique of claims that contemporary LLMs possess genuine understanding, arguing instead that they remain outside the normatively structured practices constitutive of discursive agency. This commentary examines and further develops that central thesis through the lens of an inferentialist account of discursive practice. Drawing on Wittgenstein’s remarks on rule-following and Brandom’s critique of regularism, it argues that current LLMs exhibit only conformity to rule-like regularities rather than genuine responsiveness to inferential norms. Although such systems generate linguistically sophisticated outputs and can often mimic inferentially appropriate responses, their performance remains grounded in statistical regularities rather than participation in practices of commitment, entitlement, and normative assessment. The commentary further explores the implications of this diagnosis for theories of assertion and speech acts. After examining competing accounts such as proxy assertion, proto-assertion, and make-believe approaches to chatbot discourse, it argues that LLM-generated utterances are best understood as simulations of assertion rather than genuine assertions. Building on this idea, the paper proposes the notion of “degenerate speech acts” as a useful conceptual category for understanding forms of communication mediated by artificial systems. Finally, the discussion situates the paper within broader debates concerning sapience, sentience, and artificial agency, suggesting that the challenge posed by LLMs is not merely technological but fundamentally philosophical. The commentary concludes that behavioral sophistication alone cannot establish genuine participation in the logical space of reasons, since discursive understanding depends on normative statuses that current LLMs do not possess.