Language Models Break Researchers' Mental Filters for arXiv Papers
——0——
Doesn't seem worth it to hone it again. Earlier it happened due to the multiplicity of reading groups and similar during graduate school, wide interests, and picking up certain promising "thin slice" gestalts over time.
Over the years, I had developed a pretty strong filter (according to my tastes) for scanning a few arxiv lists rapidly despite the volume. Now due to LMs, the filter seems broken, can't just eyeball stuff and mentally categorize it into different "interesting" buckets.
5:33 AM · May 24, 2026 · 468 Views
5:40 AM · May 24, 2026 · 153 Views