I toyed around with using Language Embeddings as a way to categorize my RSS Feeds.
It works pretty well. But importantly, it's so cheap that I have never really seen it on my bill. An earlier prototype used OpenAI embeddings. I loaded 5$ API credits and after a year the credits expired.
There's a ton of selection bias going on here. Terence writes mostly about open web standards and topics that would be of significant interest to folks with an RSS feed reader.
Also, RSS readers are generally automated. I know I've had them around for years pulling in articles that I never read. Like a podcast "listen" is actually just an automated download, RSS traffic does not necessarily involve anyone actually reading the article, whereas search traffic is generally high intent and is at least resulting in eyeballs on the site, if not actual readers.
The data doesnt purport to cover any more than the 1 website. Its not like there are any generalisations about other websites derived from the data. Its just "These are where my hits come from"
> RSS readers are generally automated. I know I've had them around for years pulling in articles that I never read
They try to address that
> I added RSS and Newsletter tracking. These data are very lossy. If someone is subscribed to my RSS feed and opens a post and their client downloads a lazy-loaded image at the end of the post, I get a hit.
This is probably the case for most personal websites with considerable cachet in their niche.
+ With a strong enough social network you probably don't have to care about SEO as much
You can title your post about bad customer service practices in a unique way without a second thought [0] and your more traditionally titled posts can still make the first page of a Google search with a reasonable query [1].
+ Depending on your niche your target audience is likely to already be tapped in well enough to not have to rely on search engines for content catering to their interests.
I feel like search engine practices trend along the curve shown in that meme where it's the "fool" on one end and then the "normie" in the middle and then the "Jedi" on the other end who does the same thing as the idiot. Except in this case "Jedis" only search for what's not present in their feeds (which doesn't have to be only RSS feeds) and fools can eventually cultivate their own feeds for their interests and reserve search engine use for mundane purposes that essentially fulfill the responsibility of some kind of pop culture almanac, phonebook and portal to Wikipedia.
I've been toying with LLM based agents for staying on top of feeds lately. Staying on top of the fire hose was always the challenge with RSS readers, a lot of interesting low frequency stuff from people you follow gets buried under the non stop feed of stuff from high volume news that is actively trying to drown everyone out with many updates per day. Back in the day I solved it by simply not using RSS feeds for high volume stuff. But I was following a lot of interesting people. I miss that.
This is what lead to algorithm based filtering. Hacker News uses a simplistic algorithm but it is definitely using one and it works well enough. It's why I come here. We all collectively vote things up and what remains is nominally interesting enough to skim from the front page. With a bit of editorializing.
Social networks tried to game the algorithms for ad revenue. Which is why they are a lot less popular these days. Sites like Medium, Substack, Tumblr, etc. took over from simple blogs and immediately started raising walled gardens around them to become discovery platforms, have recommendations, etc.
But at least they support RSS. A lot of websites still do. If you run any kind of website publishing regular news or article content and you don't support a feed, you are being an idiot. It's easy, doesn't really cost anything, and you might actually get people using your feed once in a while. Your site might actually have one without you realizing. Most news papers have feeds. They are everywhere. The main issue isn't finding them but sifting through them. It always was.
With agent based approaches, you control the algorithm. That wasn't possible in the past. LLMs can summarize, aggregate, categorize, group, filter, etc.
I'm building Subweb.net (not ready yet, it's just a few test feeds without the LLM pipeline turned on yet) to LLM-tag RSS feed items with topic, relevance/interest, location, and translations, and present them as feeds. I'm thinking I could maybe let users specify their preferred custom prompts and ranking params or similar, though the standard prompt is already fine.
I think the open web needs to come back, but in a fair way for everyone, giving readers control over their feeds while also sending traffic and comments back to the original sources. Not quite sure how to do that yet.
Just as a suggestion, but I've been thinking in terms of clustering. If I can figure out that five different posts from five different feeds are about one story, I can collapse them into a single UI element.
Is it possible that the users who used to find us through Google are now satisfied with AI chat summaries and no longer feel the need to click through to the actual page?
Meanwhile, the long-time users who subscribed via RSS are still showing up like they always have. If this is the case, it’s a bit of a sad reality for content creators.
I toyed around with using Language Embeddings as a way to categorize my RSS Feeds.
It works pretty well. But importantly, it's so cheap that I have never really seen it on my bill. An earlier prototype used OpenAI embeddings. I loaded 5$ API credits and after a year the credits expired.
https://aws.amazon.com/blogs/machine-learning/use-language-e...
https://github.com/aws-samples/rss-aggregator-using-cohere-e...
There's a ton of selection bias going on here. Terence writes mostly about open web standards and topics that would be of significant interest to folks with an RSS feed reader.
Also, RSS readers are generally automated. I know I've had them around for years pulling in articles that I never read. Like a podcast "listen" is actually just an automated download, RSS traffic does not necessarily involve anyone actually reading the article, whereas search traffic is generally high intent and is at least resulting in eyeballs on the site, if not actual readers.
>There's a ton of selection bias going on here.
The data doesnt purport to cover any more than the 1 website. Its not like there are any generalisations about other websites derived from the data. Its just "These are where my hits come from"
> RSS readers are generally automated. I know I've had them around for years pulling in articles that I never read
They try to address that
> I added RSS and Newsletter tracking. These data are very lossy. If someone is subscribed to my RSS feed and opens a post and their client downloads a lazy-loaded image at the end of the post, I get a hit.
This is probably the case for most personal websites with considerable cachet in their niche.
+ With a strong enough social network you probably don't have to care about SEO as much
You can title your post about bad customer service practices in a unique way without a second thought [0] and your more traditionally titled posts can still make the first page of a Google search with a reasonable query [1].
+ Depending on your niche your target audience is likely to already be tapped in well enough to not have to rely on search engines for content catering to their interests.
I feel like search engine practices trend along the curve shown in that meme where it's the "fool" on one end and then the "normie" in the middle and then the "Jedi" on the other end who does the same thing as the idiot. Except in this case "Jedis" only search for what's not present in their feeds (which doesn't have to be only RSS feeds) and fools can eventually cultivate their own feeds for their interests and reserve search engine use for mundane purposes that essentially fulfill the responsibility of some kind of pop culture almanac, phonebook and portal to Wikipedia.
[0]: https://shkspr.mobi/blog/2026/03/bored-of-eating-your-own-do...
[1]: https://shkspr.mobi/blog/2026/04/does-mythos-mean-you-need-t... — I Googled "mythos and open source". Interestingly, a forum discussion about this post came before it: https://itsfoss.community/t/does-mythos-mean-you-need-to-shu...
I've been toying with LLM based agents for staying on top of feeds lately. Staying on top of the fire hose was always the challenge with RSS readers, a lot of interesting low frequency stuff from people you follow gets buried under the non stop feed of stuff from high volume news that is actively trying to drown everyone out with many updates per day. Back in the day I solved it by simply not using RSS feeds for high volume stuff. But I was following a lot of interesting people. I miss that.
This is what lead to algorithm based filtering. Hacker News uses a simplistic algorithm but it is definitely using one and it works well enough. It's why I come here. We all collectively vote things up and what remains is nominally interesting enough to skim from the front page. With a bit of editorializing.
Social networks tried to game the algorithms for ad revenue. Which is why they are a lot less popular these days. Sites like Medium, Substack, Tumblr, etc. took over from simple blogs and immediately started raising walled gardens around them to become discovery platforms, have recommendations, etc.
But at least they support RSS. A lot of websites still do. If you run any kind of website publishing regular news or article content and you don't support a feed, you are being an idiot. It's easy, doesn't really cost anything, and you might actually get people using your feed once in a while. Your site might actually have one without you realizing. Most news papers have feeds. They are everywhere. The main issue isn't finding them but sifting through them. It always was.
With agent based approaches, you control the algorithm. That wasn't possible in the past. LLMs can summarize, aggregate, categorize, group, filter, etc.
I'm building Subweb.net (not ready yet, it's just a few test feeds without the LLM pipeline turned on yet) to LLM-tag RSS feed items with topic, relevance/interest, location, and translations, and present them as feeds. I'm thinking I could maybe let users specify their preferred custom prompts and ranking params or similar, though the standard prompt is already fine.
I think the open web needs to come back, but in a fair way for everyone, giving readers control over their feeds while also sending traffic and comments back to the original sources. Not quite sure how to do that yet.
Just as a suggestion, but I've been thinking in terms of clustering. If I can figure out that five different posts from five different feeds are about one story, I can collapse them into a single UI element.
I did exactly this with Cohere language embeddings
https://aws.amazon.com/blogs/machine-learning/use-language-e...
Is it possible that the users who used to find us through Google are now satisfied with AI chat summaries and no longer feel the need to click through to the actual page?
Meanwhile, the long-time users who subscribed via RSS are still showing up like they always have. If this is the case, it’s a bit of a sad reality for content creators.
Make RSS Great Again. Besides the jokes, RSS is a great founding pillar to get out of the algorithmic internet that is today.
I'd much prefer correct sitemap.xml to an RSS. Please let me keep up to date with your site/blog/homepage without scraping it!
Ranked yesterday, although not much discussion (18 points, 2 comments) https://news.ycombinator.com/item?id=48022560
Interesting to see RSS is still relevant, thanks for sharing.