[P] I tested Meta’s brain-response model on posts. It predicted the Elon one almost perfectly.

I built an experimental UI and visualization layer around Meta’s open brain-response model just to see whether this stuff actually works on real content.

It does.

And that’s exactly why it’s both exciting and a little scary.

The basic idea is that you can feed in content, estimate a predicted brain-response footprint, compare patterns across posts, and start optimizing against that signal.

This is not just sentiment analysis with better branding. It feels like a totally different class of feedback.

One of the first things I tried was an Elon Musk post.

The model flagged it almost perfectly as viral-like content.

Important part: it had zero information about actual popularity. No likes, no reposts, no metadata. Just the text.

Then I tested one of my own chess posts - absolutely demolished.

I also compared space-related content (science) framed in different ways — UFO vs astrophysics. Same broad subject, completely different predicted response patterns.

That’s when it stopped feeling like a gimmick.

I made a short video showing the interface, the visualizations, and a few of the experiments. I’ll drop the link in the comments.

Curious what people here think: useful research toy, dangerous optimization tool, or both?

Sources:
1. https://neural.jesion.pl
2. https://ai.meta.com/blog/tribe-v2-brain-predictive-foundation-model/