The Signal-to-Noise Collapse in AI Science Coverage

The triage burden has shifted from institutions to individual researchers, and the shift happened without announcement. One Bluesky user captured the new workflow with exhausted precision: sorting AI output into research, collective delusion, individual delusion, and investor fraud ^[17] — a taxonomy that implies all four categories are present in meaningful quantities. That implication is the story. Peer review was supposed to make that sorting unnecessary. The fact that working scientists now treat it as a personal responsibility, not an institutional guarantee, marks a genuine break in how scientific credibility…