Every week, 4,600+ AI papers are published. AI Research Newsletter uses a 7-stage pipeline to score each on 4 innovation dimensions and deliver the 15 that matter most — every Sunday. **Key features**: Multi-source discovery (5 sources, deduplicated) Innovation scoring (novelty, impact, breadth, technical surprise) Hidden gems (high innovation + low citations) Practical use cases per paper Trend detection vs. historical baselines Full-text analysis, not just abstracts

Intercom — Startups get 90% off Intercom + 1 year of Fin AI Agent free

Startups get 90% off Intercom + 1 year of Fin AI Agent free

Hey Product Hunt! 👋 I built this because I was spending 3–4 hours every week scanning arXiv and Twitter for important AI papers — and still missing things. The core idea: instead of manually curating or using keyword filters, run every paper through an LLM-powered analysis pipeline that scores innovation across 4 dimensions. Then surface the top papers AND the "hidden gems" — papers that score high on innovation but haven't been noticed yet (low citations, no Twitter buzz). Each paper also gets practical use cases — not just "what this paper says" but "how you could apply this." The whole pipeline costs about $0.30 per run (~$0.004 per paper analyzed). Stack is Python + FastAPI + PostgreSQL + GitHub Actions. I'd love feedback on: - Is the innovation scoring actually useful? - What would make you switch from your current paper-reading workflow? - What topics/sources am I missing? Archive: https://ramitsharma94.github.io/... Subscribe: https://ramitsharma94.github.io/...

Intercom — Startups get 90% off Intercom + 1 year of Fin AI Agent free

Startups get 90% off Intercom + 1 year of Fin AI Agent free

PaperMine

The AI papers that matter — scored, summarized, delivered

The AI papers that matter — scored, summarized, delivered