Much commotion about this new Anthropic paper, and it's legitimately interesting work. But they are playing coy with quasi-philosophical claims that are not justified by the results. By framing the results in the language of "access consciousness," Anthropic is, deliberately or otherwise, throwing more mud into already muddy waters, inviting an inference the data cannot support but that people will draw, and already are drawing. As Anthropic themselves concede, this does not speak to whether LLMs might have phenomenal consciousness, the "something it's like" of experience. Frankly, neither does the original GWT it's modeled on. This has been the source of enormous confusion for decades. Reportability and deliberate reasoning are behaviors that are co-extensive with phenomenal experience in humans, but they do not entail phenomenology. In fact, as I've argued elsewhere, the success of LLMs drives a hard wedge between these two things. Linguistic report turns out to be a self-contained generative process that can run in full, workspace and all, without any sensory grounding whatsoever. This suggests that consciousness-adjacent language is possible without conscious experience. If anything, this paper strengthens the case that reportability and phenomenal experience come apart. Good science. Really good marketing. Bad philosophy.