Hello HN,

I have a confession: I’m addicted to X/Twitter. But I'm terrified that my brain is being force-fed digital sugar: a cocktail of controversy, clickbait, and cat videos. We ask what’s in our food. Why don’t we ever ask what's in our feed?

The deleterious effects of social media are quite well documented: more screen time = more depression[1]. But it's even more insidious. We no longer program the timeline; it programs us. I often feel anxious after scrolling through my feed because it is incentivized to feed me controversy. The usual advice is to quit the app and go cold turkey. But that doesn't work the same reason fad diets don't work. You can’t quit sugar. You can’t quit Twitter.

So I drew inspiration from the efficacy of nutrition labels on packaged foods[2][3][4][5][6] to add a 'digital nutrition label' to every tweet in my feed, analyzing its Education, Entertainment, and Emotion content. What started off as a quick and dirty bookmarklet turned into a full fledged (completely free and open-source) Chrome extension that I call 'Resist'[7].

For example, take these 3 posts, all news, and yet different: @OpenAIDevs comes across as just educational, @paulg's perhaps evokes anxiety, and @SamAntar's leans controversial. Resist understands the content, not just the app. For the first time, you see what you're truly consuming.

https://imgur.com/QFXwNd4

https://imgur.com/lXrtuNb

https://imgur.com/KFRiIed

With Resist, you can also set an attention budget — a digital diet for your mind. It tracks, it tallies, and when you go over, it autohides the junk. Willpower not required.

https://imgur.com/Ciu1KN6

That's the long bet: disengaging from the junk, starving them of clicks, and breaking the loop. Without the oxygen of engagement, I'm hoping the timeline adapts to what _we_ choose. Just like food labels reshaped supply chains[8], attention labels could reshape our feeds.

Resist is a proof of concept and has some serious drawbacks: it’s web-only, the categories are just a first draft, it currently works only on Twitter (but can be extended), and the classification models are sometimes way off. But it presents what could be: a world where you are in control of where your attention is directed.

Full blog post: https://bipinsuresh.info/blog/resist.html

Video demo: https://youtu.be/UmWCdgHQ_8o

Download: https://chromewebstore.google.com/detail/resist/paeflkahgaod...

Twitter thread (for irony): https://x.com/bipsandbytes/status/1963385093080404043

Code (MIT license): https://github.com/bipsandbytes/resist

[1] https://jonathanhaidt.com/social-media/

[2] https://pmc.ncbi.nlm.nih.gov/articles/PMC6340779/

[3] https://pubmed.ncbi.nlm.nih.gov/32493057/

[4] https://ijbnpa.biomedcentral.com/articles/10.1186/s12966-019...

[5] https://www.cdc.gov/pcd/issues/2014/13_0161.htm

[6] https://journals.plos.org/plosmedicine/article?id=10.1371/jo...

[7] https://chromewebstore.google.com/detail/resist/paeflkahgaod...