antiochIst

The story spread from Bloomberg at 12:01 PM as first to report, then amplified by x.com and 9to5mac, followed by feeds.macrumors, machash, and appleworld.today. Yandori tracked 45 sources with 10 citations showing how the news propagated: https://yandori.io/news-flow/story/2026-01-21-9996613-apple-...

You can track how this story spread from CNN at 3:32 PM as first to report, then whistleblower.org, peters.senate.gov, and eventually scienceblog.com and democracydocket.com. Yandori shows 54 sources picked it up with 48 citations: https://yandori.io/news-flow/story/2026-01-20-9734134-trump-...

Interesting to see how this news spread across 45+ media outlets. Yandori tracked the news flow from AP News at 9:31 AM as first to report, then blog.google as the amplifier around 8:07 AM, followed by mashable and other tech sites. You can see the full timeline at https://yandori.io/news-flow/story/2026-01-22-10232467-say-g...

Interesting to see how this story developed across 60 news sources: https://yandori.io/news-flow/story/2026-01-21-10020950-takea...

You can see how this story spread across 52 sources over 21 hours here: https://yandori.io/news-flow/story/2026-01-22-10318781-amazo...

Full news story coverage/flow here: https://yandori.io/news-flow/story/2026-01-07-6507910-federa...

Full news coverage/flow here: https://yandori.io/news-flow/story/2026-01-07-6507910-federa...

Here is a clear example (sad news) but you can see how it breaks (via twitter here) and then flows: https://yandori.io/news-flow/story/2025-12-18-3003874-nascar...

I have a bunch of heuristics - but broadly speaking a new domain + a new good looking site. ie: some legit pages, legit socials, and some real owner/purpose behind it.

I don't know if drop shipping is still viable, but there are lots of those types of sites so...

I run a system that tracks newly launched websites worldwide. In November it detected 677,544 new sites across 392 countries.

Some findings from the dataset:

Geography (top countries):

United States: 253,589

India: 34,127

Canada: 20,263

United Kingdom: 18,701

Pakistan: 10,124

(392 total countries)

Industries:

E-commerce: 164,010

Adult & Gambling: large category overall

Gambling (L2): 84,353

News & Blogs: 49,424

SaaS products: 39,105

Niches (L3):

Online Casinos: 81,608

Clothing: 33,553

Niche SaaS tools: 31,881

Home Decor: 31,273

News & Journalism: 29,750

Platforms (detected on ~295k sites):

WordPress: 116,250

Shopify: 84,407

WooCommerce: 42,615

Squarespace: 25,328

Wix: 23,598

Webflow: 3,049

TLDs: .com: 435,622

.store: 38,223

.org: 26,474

.online: 23,422

.site: 22,919

.ai: 6,167

Happy to answer questions about methodology, accuracy, crawling, classification, or detection heuristics.

Thanks - I got fix for this.

FYI - I'll integrate this if anyone want to pay for twitter api fees.

I think this could be done, but would require paying more than I want to for the highest level of api access...

Yea there is some spam stuff for sure... working on improving filtering it out...

I get most of it, but I think especially around the holiday some stuff is getting through... Some black friday deals were actually hitting like news does...

You can kinda tell based on the distribution. Organic spread has less similarity between articles, less syndication, more spread out in timeline of releases...

Some stories are very clearly manufactured

I'm polling rss feeds from a bunch top 200k sites in the world.

Thanks for that bug feedback - ill get fix.

ehh, timezones handles just with some basic parsing logic...

I'm not pulling from social media yet.

Yea I feel you... Honestly I kinda just whipped this thing up in context of a larger project I'm working on.. so i have not given much thought to who it will serve. "rooting out bias" is interesting idea... But a bit negative in nature, I was more hoping to identify & highlight the original/powerful sources of news...

It's a whole thing... I run a project called websitelaunches, so I have index of basically the whole internet (500M+) sites. I took the top ~200k news related sites from there that had rss feed.

I'm basically throwing away non english articles for now... I'll pry get them in later, but I want to get english right first before trying to move to other languages...

The embeddings themselves will (pry) cluster ok in different languages (but I have not tested this yet)

yea, what im currently doing is pretty simple check on published at date from the rss feed (with some small validation checks)... but its causing issues bc it can be wrong and mess up everything...

I think checking source in story is next step...

Yea not all major have rss feeds, but it seems like the majority still do.

No translation yet.

I think the biggest problem is im relying on published date from the news source itself too much and its wrong sometimes... not super often, but if 1 out of 100 sources get its wrong then it can steal credit for being source article when its not.

Currently I'm using Snowflake’s Arctic embedding model on the whole story not just the title, to cluster stories. There are still some issues, but its not as simple as looking at title publish date.

Yea, I need to do some work on improving first to publish... currently I'm relying pretty heavily on the published date provided in the story itself, but sometimes that is wrong and makes it look like a later publisher was first to publish.

I built a system that detects (almost all) new website launches. Right now it only covers .com and English sites, but support for more TLDs and languages is coming. You can browse launches, filter by category or location, and upvote/bookmark favorites.

My hope is the Pareto is friendly enough in this domain, so rating very small portion like 1/10000 more or less gets like 80% of of stuff... ie: only popular content is rated...

I don't imagine that would fix much...