RSS users - how do you use, organize and maximize your enjoyment of RSS?

I subscribe to and parse RSS feeds using a module of a self-written & hosted Discord bot, separating feeds between channels depending on content type- but the majority of feeds I follow are web-comic or Youtube feeds. I refuse to own a Youtube account, anything Google, so that is how I keep up with creators I enjoy. Like your friend, I've tried following some of the "louder" feeds such as Ars, but I find that sort of content is more easily digested (or ignored) in e-mail subscription format. When it comes to web-comics I definitely find the time to read each new post, but this is easy when most comics update semiweekly at most (Mon-Fri, Mon-Wed-Fri, etc).

If I want to browse a large selection of semi-interesting content, I always have Reddit- a bad habit, I know. I treat RSS more like a "follow" button for content creators. It's less overwhelming that way.

Same as you I use the Web UI of FreshRSS. I also use Reeder on iOS.

Feeds are categorised like:

• Group 1: Where I want to read every post
• Group 2: Where I want to at least read every post title
• Everything else in some high level categories and I just scan through

I wrote a little custom rss reader in Python, I don't subscribe to that many feeds(yet?), so there's not that much to organize. But it has been a fun little project so far. One of the feeds i follow post daily and other one on average weekly, so I'm reading and filtrering the topics manually. I'm pretty new to the whole rss reader thing, but so far i get the feeling that I need to treat every feed differently and figuring out what best works for it.

Aside from using the Feeder android app for reading a bunch of news sites, I'm a big fan of Newsboat + mpv for subscribing and watching/queueing YouTube videos from my desktop.

I have this in my urls file and it's basically like a Subscriptions page:

"query:new videos:(unread = \"yes\") and (tags # \"yt\")"
"query:all videos:tags # \"yt\""
https://www.youtube.com/feeds/videos.xml?channel_id=UC..................JtAw "!some channel" yt
https://www.youtube.com/feeds/videos.xml?channel_id=UC..................M7dQ "!another channel" yt

The "!.*" special tag hides the individual feed from main view. Then, in the config file:

feed-sort-order lastupdated
browser "linkhandler %u"

And this is for opening/queueing videos in mpv:

$ cat linkhandler
#!/bin/sh

exec >/dev/null 2>&1

case "$1" in
        *youtube*)
                if [ -e /tmp/mpv.socket ] && pgrep mpv; then
                        echo loadfile "$1" append | socat - /tmp/mpv.socket
                else
                        setsid -f mpv --no-terminal --input-ipc-server=/tmp/mpv.socket "$1"
                fi
                ;;
        *)
                setsid -f $BROWSER "$1"
                ;;
esac

I wrote my own RSS reader, YOShInOn which uses machine learning to filter my feed.

I subscribe to about 110 feeds on Superfeedr which makes an https call to a web hook whenever a new feed items appears. It costs about 10 cents per feed per month which I find highly economical for high volume feeds but deters adding a few hundred or thousand independent blogs that get (say) one post a week or less. The web hook runs on AWS Lambda and adds the feed items to an SQS queue, YOShInOn fetches from this queue when it is convenient, putting the feed item into an

https://arangodb.com/

database. For each “cycle”, the system looks at the new articles and processes them into vector with

https://www.sbert.net/index.html

From there on it uses classical algorithms from scikit-learn, particularly it clusters documents with k-means clustering and classifies the documents using the probability calibrated support vector machine. It picks the top articles out of 20 clusters for diversity and then it blends in a few randomly chosen articles for more diversity and better performance evaluation.

It chooses 300 articles for me to look at which takes as long as it does (maybe 2-3 days) and then I run another cycle. It might start with 6000 or so articles in which case it is showing me just 5% of the articles. The main user interface looks like TikTok or the long forgotten StumbleUpon with a thumbs up and thumbs down button. The important thing is that it gets both positive and negative feedback over articles which is good for training the classifier, and, just as important, makes it possible to calibrate the classifier so it can predict that I have a 64% chance of liking an article which is useful when combining the rankings with some other information. (It has some other screens which will show me articles that it thinks would get a lot of comments on Hacker News, or articles that landed in a particular cluster, or came from a particular source.)

Favorited articles can be viewed in a list and I have buttons that will queue up a link to post to Mastodon or Hacker News. There is also a simple search facility. It shouldn’t be that hard to make it also function as a bookmark manager.

Nothing too fancy here, I just use NetNewsWire (one of the oldest Apple platform RSS readers, which is now FOSS) to subscribe mostly to tech newsfeeds and a few blogs. The reader is written extremely well and is lightweight and clean, making a point of being a good desktop citizen with no trace of “UI as branding” (it’s just another tool and has no reason to stand out). It looks like it could’ve shipped with the OS. It’s boring and functional and that’s why it’s perfect.

It can sync with a mobile app counterpart as well as with other readers via support for a wide range of RSS services (Feedly, etc) but I intentionally don’t use any of them so time spent on news is bounded to when I’m using my main computer.

I try to keep the signal to noise ratio as high as possible. I stay away from generic, high volume feeds, and filter feeds if necessary. I use services like Reddit as curators, as they allow me to filter feeds by vote counts.

I don't follow breaking news through RSS. I just visit a handful of news websites daily for the bigger developing news. I do subscribe to news digests however, to get summaries of what has been going on in the world.

With something like Ars Technica which you mentioned, I don't subscribe to their main feed, or necessarily even their main categories. Instead, I subscribe to their individual journalists who I know produce high quality content on a topic that I am interested in.

I divide feeds into thematic categories as it makes it easier for me to process articles quicker and more accurately when I have that context as a reader. For instance, if I come across a headline about AI in my culture category, I instantly know that the article is likely very different than a similarly sounding article would be in my tech category, or in my LLM category.

I sort articles within categories by publication date, oldest first. This again helps with context, as sometimes things build on top of earlier things.

Most of my categories are intended to be read through like news feeds, but there are some which function a little differently. My music recommendations category, for instance, gathers album recommendations from a handful of sources. I only touch that category when I want to listen to something new, basically just picking whatever the next unread article recommends. As I work from home, I can and do listen to music throughout the day, but typically that category is never fully "read" as it always has unread articles (unchecked recommendations).

At one point I spend too much time organizing, categorizing and filtering RSS feeds trying to optimize everything. But that was probably taking more time than simply being a bit more critical of what I subscribe to and the leave the FOMO behind and just mark everything as read if I scroll by. And removing feeds if I never read anything from it. For the heavy sites like Ars, I have signed up for their weekly newsletter instead as a way to just catch up on what was published of major things.

I really like https://kill-the-newsletter.com/

It’s a service to read newsletters in RSS. It’s included in some paid RSS clients (like Feedbin), but that service is free.

NewsBlur has the option to show content from all feeds together in one list, but I don't use it. Instead I scan the list of feeds with new posts and click on whichever ones catch my eye. Web comics get read first.

I don't read anything that's tremendously high traffic, but there are feeds that post multiple times a day, where I don't keep up and click "mark as read" every so often.

I use Readwise’s Reader app. It gives you an email that directs new content to the “Feed.” I can read it, skip it, save it to “Later.”

I guess I am quite similar to you and use it for content discovery. I have recently changed my methods, and have

• Blogs
• News
• "Notifications" (e.g a few github things I want to read every release notes for, severe weather events, data breaches)
• Parliamentary Bills (this one spams a little when in session), I might also add other stuff since nearly everything related to parliament has RSS feeds
• Tech news feeds

I USED to separate the news more into different regions, as well as separating my French and English news, but I really removed many of the feeds as it was just noise.

I'm still not 100% happy with my feeds, I think I'd want some way of rating what comes through my feed. Conceptually, Nunti achieves this but it doesn't yet have a way to have separate databases for different languages and also doesn't seem to like my OPML file.

I'd also like to curate a feed of investigative journalism from a plethora of sources, low noise, long form, and high quality. I've yet to really find enough dedicated feeds to this to achieve it sadly.

Nothing too complex. I have one international news sub to Wikipedia Current Events, so I only read them once, and one local news one that is more frequent. I'm subbed to a few forums that I've been part of for over a decade, some gaming sites. Some comics (SMBC, xkcd, etc), github app updates and some blogs.

I used to be subbed to my YouTube channels as well, but I discovered it works better to use AntennaPod for that, since they sit in my inbox and clicking on them doesn't make them read. I have it set up so that they open into NewPipe, so it's pretty efficient.

I subscribe to Feedbin as a service to keep my feeds synced across devices.

I then consume feeds through the Reeder iOS and macOS apps.

For my RSS usage I just subscribe to several feeds (mostly news and blogs) in QuiteRSS. It has a built in browser with ad blocking which works pretty well but it is easy to open items in a full browser for some sites that don't play well or I just want to view in my main browser.

My favorite reader was Google Reader (RIP). The thing that comes closest to that feeling of minimalism, but nice, is CommaFeed. It's open source and available as self-hosted or hosted.

https://www.commafeed.com
https://github.com/Athou/commafeed

The default screenshot looks a little bleak-- it's customizable and has non-dark modes as well. Here's an example of mine, just for comparison's sake:

https://ibb.co/YRdrJyr

As far as content, it originally was oddball blogs that I had no other way of reading about, but many of those over time became extinct. I found a random post on Hacker News that listed like 500 top blogs that were somehow categorized as worth reading, so I pared through some of those as well. That is available here:

https://github.com/outcoldman/hackernews-personal-blogs

I also like Uncrate because it shows me ridiculous things like $34 million dollar estates up for auction, as well as titanium sporks, which I could (probably) afford.

https://uncrate.com/