Newsletters, podcasts and channels on AI, data and tech

The site follows a growing list of them, with podcasts and YouTube channels beside them: independent writers such as Import AI, Interconnects, Latent Space, One Useful Thing, The Pragmatic Engineer, Stratechery, Data Engineering Weekly and SemiAnalysis, and company newsletters such as The Batch from DeepLearning.AI, TLDR and dbt Labs’ Analytics Engineering Roundup. The table ranks them by how often they really send an issue, measured from their own archives.

Each newsletter shows Free, Free and paid, or Paid. For Substack newsletters it is measured every week from the public archive: the share of the latest 12 issues that are for paying subscribers only. For the others it comes from the subscribe page. A newsletter that goes paid or free shows up on the Changes page.

Cadence is measured, not taken from the sign-up page: issues a month over the last six months, or over the days the archive covers when it holds less. Active means an issue in the last 45 days, paused 45 to 120 days, stopped longer than that.

Yes. The newsletters page has a Download OPML link (newsletters.fru.dev/newsletters.opml) that Feedly, Inoreader, NetNewsWire, Reeder and most readers import in one step, and newsletters.fru.dev/rss.xml carries every new issue from every newsletter in one feed.

No. The site keeps only each issue’s title, date and link, from the newsletter’s own public feed or archive. Every row links to the original, where the author decides what is free to read.

Every newsletter page has Suggest a correction, and the newsletters page has Suggest a missing newsletter. Each suggestion is checked against the newsletter before anything changes.