Show HN: Hacker News AI link reading list (ai-reading-list.pages.dev)

arXiv announced that its Computer Science category will no longer accept standalone review articles and position papers unless they have already been accepted by a peer‑reviewed journal or conference. The policy change, explained on the arXiv blog, responds to a recent flood of low‑quality and AI‑generated submissions—many produced or polished with large language models—that overwhelmed volunteer moderation and degraded the signal‑to‑noise ratio in preprint listings. For the AI/ML community this is a consequential shift: review and position pieces often distill trends, synthesize literature, and shape research agendas, so gating them behind peer acceptance slows the rapid dissemination of community syntheses. Technically, it reduces the pool of potentially hallucinated or shallow LLM‑produced surveys, lowering moderation load and improving overall quality of hosted content. But it also raises concerns about increased gatekeeping, longer feedback cycles, and fewer publicly accessible syntheses for practitioners and educators. Expect authors to either wait for formal acceptance, migrate exploratory overviews to alternative platforms, or annotate AI‑assisted drafts more rigorously; meta‑researchers should monitor changes in citation patterns, reproducibility, and platform migration as the policy takes effect.