The number of monthly submissions to the popular arXiv preprint server has reached 30.000 in March 2026. There is a need to filter through the noise of irrelevant research articles.
My primary source for discovering new papers (as a machine learning researcher) is an RSS reader that aggregates multiple research feeds. Each day, I review the main feed and save promising papers for later reading in a reference management software.
Software
As an RSS reader, I use Reeder Classic. It has a minimal user interface and a one-time fee. Because most research alert services rely on email, I use the free, open-source Kill the Newsletter! service to convert email newsletters into RSS feeds I can then add to my RSS reader.
For reference management, I use Zotero. It is free, open source, and supported by an extensive plugin ecosystem.
Sources
I primarily use Google Scholar Alerts as the aggregation backbone. It likely has the most comprehensive coverage and allows both keyword and author-specific searches. I create one separate RSS feed for each alert (using Kill the Newsletter!). This allows sorting alerts (and therefore, specific keywords or authors) into separate categories and folders within the RSS reader.
For a more tailored service, Scholar Inbox provides very personalizable paper recommendations. General trends can be added using popular research newsletters (e.g., Hugging Face Daily Papers, The Batch, AI Newsletter, Causal Python).