Trendar collects AI/ML papers, global news, and GitHub trends, reads every item, and ranks what matters — then puts the day in one AI-written line.
Last collected 2w ago🔭 9/24 ICLR 2027 마감 + NeurIPS 결과(추정) — 9월 하순 이중 이벤트 관행 — 전 분야 폭증 예상, 공시 전 추정
Every day the AI calls a trend's direction and posts the bet publicly. At maturity it's auto-graded against a naive "assume it just continues" baseline — and the hits and the misses both stay on the board.
See the live scoreboard →Every morning it pulls new AI/ML papers (arXiv, OpenAlex & more), global news, and trending GitHub repos.
The AI reads each one and scores it by novelty · contribution · impact — not by popularity. Quiet, important work doesn't get buried.
It predicts where trends are heading, posts the call in public, and grades itself against a naive baseline at maturity.
Toward expert-level medical question answering with large language models
Med-PaLM 2 achieves physician-level medical question answering performance through medical domain fine-tuning and strategies for improving reasoning and grounding. In human evaluations, physicians preferred Med-PaLM 2 answers over those from other physicians, and in a real-world medical question pilot study, specialists preferred Med-PaLM 2 over generalist physician answers 65% of the time.
927 cited · 1y ago
Bend – A language that blocks AI mistakes via proof, on CPU and GPU
▲ 371 · 2w ago
tt-a1i/archify
★ 66.1K
Built solo, free, and updated every morning.