Social Epistemology Review and Reply Collective 15 (8):10-18. 2026In Hauswald (2026b), Hauswald responds to Mizrahi (2026), which is a reply to Hauswald (2026a), which is itself a reply to Mizrahi (2025), which is a commentary on Hauswald (2025). I am very grateful to James Collier for inviting us to continue this fruitful dialogue. I also thank Rico Hauswald for his thoughtful engagement with my arguments. In what follows, I respond to his most recent response, i.e., Hauswald (2026b). It seems to me that the arguments outlined in Mizrahi (2025) still stand. In fact, Hauswald (2026b, 1) now grants that the argument from bullshit machines is sound. As for the argument from responsibility and accountability, I provide additional reasons to think that human novices should not regard epistemic sources as epistemic authorities unless those epistemic sources can take responsibility for their assertions and be held accountable when those assertions turn out to be false. I also argue that Hauswald’s attempt to draw parallels between human novices deferring to human experts and human users prompting LLMs breaks down when the indicators novices use to assess the trustworthiness of epistemic sources are taken into consideration. ( direct link )