Show HN: Paper Radar – daily ArXiv triage, now runs on Claude Haiku too (github.com)

Paper Radar, a novel tool for daily arXiv triage, promises to revolutionize how researchers keep up with the overwhelming influx of scientific papers. By leveraging TypeSafe's System One model, Jev, Paper Radar assesses every new arXiv submission, providing a probability-based yes/no judgment on relevance tailored to user-defined interests, at an astonishingly low cost of about $0.06 per day. This stands in stark contrast to conventional methods like keyword alerts which can miss relevant papers due to varied phrasing or embedding algorithms that only recommend based on past reading habits. The significance of Paper Radar lies in its ability to democratize access to cutting-edge research without the need for extensive filtering or manual triage. It guarantees that users can efficiently navigate the avalanche of research—over 1,500 new papers daily—by providing quick, reliable assessments on what deserves attention. Notably, it introduces an effective benchmarking system, achieving a 96.9% consistency with human reviewers in preliminary trials, demonstrating its potential as a pivotal resource in the AI/ML community. As researchers increasingly rely on AI-driven tools for research synthesis, Paper Radar's unique approach could set a new standard in research management and discovery.