My AI Signal is live!
By ericblue Categories: Open Source AI LLM | 0 comments
Check out https://ai-signal.ericblue.com/
Earlier this week I shared a small local first pipeline I built to help me step back from the constant flood of AI news and find actual signal.
I decided to take it one step further.
I used to blog heavily from 2004 to 2012, and for a while experimented with LifeStreaming by
curating links and articles I found interesting as I went. Over time that faded, mostly because the
volume of information exploded.
AI Signal is an experiment in bringing that idea back using automation and AI assisted filtering.
Behind the scenes, it builds on the same framework I shared earlier:
– pulls discussions from communities I care about, starting with Reddit
– scores and ranks posts using real signals like engagement, recency, and depth
– runs a second pass analysis with explicit agent instructions to identify novelty, themes, and
notable quotes
– publishes a short, readable digest instead of an endless feed
The goal is not to cover everything. It is to surface a small set of genuinely interesting stories
based on the communities and topics I care about, and to make it fast to scan.
This is intentionally early, opinionated, and evolving. For now, it is an experiment in automated
curation and AI assisted microblogging, and a way for me to blog again without turning it into
another job.
If you are feeling the same AI news fatigue, you might find it useful too. This will also publish
automated weekly summaries of the most interesting AI news every Tuesday.
If you want to build something similar, the core building blocks are open source and available in
the original repo here:
https://github.com/ericblue/curated-post-digest
Feedback welcome!