# Where I Actually Get My AI News

> By Friedrich Bremer (Fritz), Paid Social, AI and Automation Consultant. Published on business.mndst, https://business.mndst.co/p/ai-news-sources

**Topic:** AI Tools

A breakdown of how to follow the right AI sources instead of chasing outdated posts. Repositories, newsletters, and people worth your attention, plus a simple system to read them without drowning.

## The short version

- Most AI content you scroll past is already old by the time it reaches you.
- Primary sources beat aggregators - go to the repo, not the roundup.
- A daily scan, weekly analysis, and research feed covers almost everything.
- Following the right people cuts through more noise than any newsletter.
- Curation is a skill. Build your list once and prune it often.

## Most AI news reaches you late

By the time a tool trends on your feed, the people building with it have already moved on. The reel you watched about a new model was probably filmed a week after the release, edited, then pushed by an algorithm that decided when you saw it.

That lag adds up. If your entire input is repackaged commentary, you are always reacting to yesterday.

My read is that the gap is not about intelligence. It is about where you point your attention. Go closer to the source and the delay shrinks.

## Go to the repo, not the roundup

GitHub is where a lot of this actually happens first. Releases, changelogs, and issue threads tell you what is real before anyone writes a post about it.

There are curated lists that do the heavy lifting for you. ByteByteGo keeps a running piece on the top AI repositories worth watching, and it is a decent starting map if you do not know where to look yet.

Star the repos you care about and let GitHub notify you. That single habit puts you ahead of most people who wait for the summary.

## Newsletters do the filtering

Nobody can read every paper and every release. Good newsletters exist because someone else already sorted the signal for you.

AI Breakfast runs a weekly analysis of new projects and products. There is also a well maintained list of over 50 AI newsletters on GitHub, updated monthly, covering daily briefings and research feeds. Pick two or three, not twenty.

I keep seeing people subscribe to a dozen and read none. One you actually open beats ten you archive.

## A reading stack that does not eat your day

daily.dev describes a simple 1-1-1 approach that I think holds up well: one daily scan, one weekly deep read, and one primary research feed. That is the whole system.

The daily scan keeps you current. The weekly read gives you context. The research feed keeps you honest about what is hype and what is shipping.

Think of it like eating on a schedule instead of grazing all day. You stay full without the constant snacking that leaves you tired and no smarter.

## Follow people, not just outlets

The fastest signal usually comes from individuals, not brands. Researchers and builders post things in real time that no newsletter will cover for days.

The trick is being ruthless. Follow people who ship and explain, unfollow the ones who only react. Your feed becomes a source instead of a distraction.

Curation is not a one time setup. I prune my list every few months because the people worth reading change as the field does. Build it once, then keep sharpening it.

## Sources

- [Awesome AI Newsletters](https://github.com/alternbits/awesome-ai-newsletters)
- [The best sources for AI news in 2026](https://daily.dev/blog/best-ai-news-sources/)
- [ai-newsletters](https://github.com/topics/ai-newsletters)
- [Top AI GitHub Repositories in 2026](https://blog.bytebytego.com/p/top-ai-github-repositories-in-2026)
- [GitHub - aitools-coffee/Open-Source-AI-News: A curated ...](https://github.com/aitools-coffee/Open-Source-AI-News)

## About the author

Fritz has 8+ years in paid social and adtech and has managed $50M+ in annual ad spend. He builds the systems behind growth, and publishes the ones he builds for himself. He works across Europe and United Kingdom and sells one-to-one consulting and done-with-you build work.

- Website: https://business.mndst.co
- Instagram: https://www.instagram.com/peopleseye/
- Book a call: https://calendar.app.google/cHLyFEDs1U11JGUz6