Blog
Cutting The Noise: Building A Personal AI Driven News Desk
Fed up of the constant noise and other people's algorithms, I decided to build my own local news desk app to create a curated Daily Brief of AI industry and research updates from a diverse range of writers, bloggers and researchers.
The goal is to cut through the clutter of AI hyperbole, regurgitated news and opinions to find interesting, useful information that I would personally like to read and learn from.
This idea had been brewing in my head for many years but recently the launch of new Codex models and an X post by Andrej Karpathy talking about RSS feeds prompted me to finally investigate building out a news desk app for myself.
What The News Desk Does - At A Glance
Reads everything (email newsletters and RSS feeds) and ranks it based on my preferences to deliver on one page:
- A Daily Brief
- Top 10 Headlines
- 1 Top AI Research Paper
- Trends Tracker (tracks topics mentioned over time)
Too Much Content, Not Enough Time
I subscribe to a number of email newsletters and I follow a few publications too, what I wanted to do was have something that could read through the emails and pull in from RSS feeds and then rank those based on my preferences.
While just having everything in one place is useful, I didn't want just a dump of info that I would still need to scroll through.
Other Useful Features
- I can vote for what I like/dislike
- I can save bookmarks to read later
- I can add in new sources
It's still learning my preferences over time which will be a work in progress but I already love it as there's no endless scrolling or checking of emails.
Future - a Roving Reporter
I want to build out a get to a place where the there is a roving reporter agent who scouts everywhere for the latest AI news and updates and feeds in back into the news desk for curation.
So that eventually I'm in a place where I'm not having to add in the RSS feeds manually as I still want to quality check each source and not just mass ingest websites and blogs.