I built a system to read 2,846 newsletters for me
The messy reality behind my "perfect" productivity solution
My Gmail was a disaster. Hundreds of AI newsletters I should read, needed to stay current on, but never actually opened. Just sitting there creating this low-level anxiety hum.
I'm a newsletter junkie. Fear of missing out on the next big AI breakthrough drove me to subscribe to everything. Substack became my dealer. Every interesting writer got an immediate subscribe.
The problem wasn't the quality of the newsletters. These were brilliant people sharing genuine insights. The problem was quantity. My inbox became this constant reminder of how behind I was, how much I wasn't reading, how everyone else was staying current while I was drowning.
So in the spirit of my regular AI experiments I did what any rational person would do: I built a system to read them for me.
Claude + Zapier workflow that scans the last five days of labeled newsletters, extracts what matters based on my specific interests, and delivers a digest in my Google Drive. Summaries, insights, and patterns I wouldn't have caught. Contradictions between different sources.
Most of my FOMO was procrastination wearing a productivity costume.
I wasn't collecting information to get smarter. I was collecting information to feel smart. To feel prepared. To feel like I was doing something important by staying informed.
But consuming isn't creating. Having 2,846 unread newsletters doesn't make you informed. It makes you anxious.
The automation didn't cure my FOMO. It revealed that most of my FOMO was fake. Real insights come from building, not browsing.
The part they don't show you in the success stories
The original inspiration came from
at . His workflow was the spark. But getting the system prompt to actually work for me is a different story. Plenty of refinement and iteration.I wanted the system to
only surface infos relevant to my actual work
catch patterns I'd missed
flag contradictions between different sources
challenge my assumptions
and, obviously, also scan for good prompting techniques
Teaching an AI to think like you think, but better than you think, and putting that into clear instructions is harder than I thought. Try and define what your version of “good” actually means and explain that to an AI…
The first version got me 85% there which, to be honest, felt a little too good to be true. Then I hit my free Zapier credits and realized this miracle had a monthly price tag.
So now I'm building version 2.0 with Make, trying to crack that final 15% while keeping the costs low. The process continues.
The truth is, I'll probably always be someone who subscribes to too many things, builds overcomplicated solutions, and then rebuilds them because the first version wasn't quite right. That's just how my brain works.
But at least now my inbox anxiety has a purpose. It's a driver for something I'm actually making, not just evidence of everything I'm failing to consume.
Progress, not perfection. Always learning.
Admirative of your system for the effort it requires to build but honestly … i don’t believe it will be so useful (but I will be happy to learn I was wrong)
1/ I don’t think there is a magic prompt that will spot everything you need to read. Because what you need to read is at a very specific place hard to capture in a prompt: familiar enough so that it intersects your interests, but new enough not to be a repetition of what you already know.
2/ more fundamentally, I think there is another “right” way to treat unread newsletter (by “RIGHT” I mean effective, in the sense that you both learn something and you don’t create FOMO). The idea is to consider those unread like an infinite library of interesting content. Whenever you feel like reading you go there and something interesting pops up. That’s so great ! And with the number of newsletters, it creates diversity.
Sorry for this long rant :)
You wrote. I 100% related. “Most of my FOMO was procrastination wearing a productivity costume.” That’s the whole story for so many of us. Thanks for taking the time to choose just the right words.