Newsblur has a training feature [1], with a thumbs up/down on article tags, author, and keywords in a title that can bubble up more interesting articles or completely hide them.

Unfortunately training is per feed, there's a newer Premium Pro tier with global training coming out with it, but it is much more expensive [2].

1. https://www.newsblur.com/faq (under Intelligence section)

2. https://forum.newsblur.com/t/global-keyword-training/5419/6

Wow, $299/year. you weren't kidding.

I assume this is targeted at professionals that are trying to stay abreast of developments around certain topics/subjects. If you think about it through that lens it's reasonably priced, I suppose.

lol all these RSS readers really go in on solving problems as broadly as possible - not just ocean boiling, it leaves each user with a subpar experience even if it’s novel (having to laboriously refine results yourself). it would be easy for them to for instance scrape many popular news sources or APIs for ranking/top story info to at least cover these extremely common and extremely noisy feeds

Yes. Feedly seems to be marketing itself at investors and speculators lately (or trading bots I suppose), because they're using NLP to highlight and categorize articles based on whether they mention acquisitions, mergers, leadership changes, etc.

Feedly seems to be marketing itself as whatever interest group they think you are, because they're using NLP to highlight and categorize articles but at least for me not regarding whether they mention acquisitions, mergers, leadership changes, etc (though some of the articles my feed do) but rather whether they mention new security vulnerabilities or attacks - and I'd guess the same for many other specific niche domains.

I also saw the "new exploit" notifications. It may just be a question of feed curation, too.

Hey mate, you can give a go to lenns.io (an opinionated RSS reader). You can assign different priorities to both sources and categories to have the important (for you) articles show first.

I added rss filter/grep support to my homebrew rss->email script. So for example, I follow https://news.ycombinator.com/rss but only for articles with titles matching certain patterns:

    mike@klaus:~$ rss --list|grep ycomb
    155. https://news.ycombinator.com/rss
    mike@klaus:~$ rss --list-grep 155
    74. title = (?i)\b((e-?|web)?mail|hardenize|irc|internet relay chat|grpc|hashicorp|rust|debian|c\+\+|perl|(bit|name)coin|tor|pgp|gpg|gnupg|openpgp|digitalocean|ovh|linode|grepular|email\s*privacy\s*tester|parsemail|ssl|https|backdoor|apache|exim|distribut|peer (to|2) peer|vpn|secur|anonym|webrtc|torrent|webtorrent|nextcloud|owncloud|graphql)(ity|ous|e?s|ing?|ed?)?\b

I also randomly hit the front page of course, otherwise I wouldn't have seen this. Maybe I should add "rss" to my regex

I had a similar problem and created nooshub.com some years ago, it groups similar articles and uses the group size to detect news trends, but everything can also be configured per feed collection if you prefer chronological feeds. I use this to quickly check what is going on with high rate feeds, for other feeds like podcasts I keep it classic... You can see it in action on some pages that do not require login like https://www.nooshub.com/pages/7-news-international

That's one of the reasons I build lenns.io. An alternative RSS reader. I gives you some additional vectors of control. For example, you can limit the number of items per source to show on your feed, as well as assign priorities to both sources and categories. Combined - it gives you a calm feed following all the sources your are interested in.

I'd be more than happy if you give it a go and share how it seems.

If you just want highlights, https://brutalist.report is my go to

With FlipRSS.com we let subscribers choose only the content they want to receive. Using RSS feeds, matched to interest groups, content creators can keep in regular contact but deliver more personalised subscriber experiences. We love RSS!

I had this problem too, until I installed and configured sfeed[1] on my server --- now it's all on one page, and much less "push-y" :P

[1]: https://codemadness.org/sfeed-simple-feed-parser.html

(You can see my setup at https://acdw.casa/planet)

I ran into the same problem and ended up building https://legiblenews.com/upgrades/the-weekly-edition

We need to get back to a world where we’d have daily or weekly issues of news that we all read so that we can’t point at the same thing and say, “that’s right” or “that’s wrong”.

Not a complete solution to your problem, but the New York Times at least offers different RSS feeds for different sections.

So you can have a different feed for Science and Technology, and one for Australia, and one for New York. By not using the front page firehose, you can keep the number of inbound articles to a more manageable level.

The BBC as well: https://www.bbc.com/news/10628494 ("top stories", "world", "politics", "health", "technology", ...)

Though like many sites, I have no idea how you would reach the "feeds" page from the front page (I gave up and searched Google for it)

If you have Reeder, you just paste the regular web page URL into it and it tries to figure out the RSS address on its own. It's possible that other RSS readers do the same, since the RSS information is usually in the page's metadata.

Yeah there is usually a <link> but none to be found on bbc.co.uk...

newsboat supports regex filters and macros/outside scripts which can be powerful

Newsboat also has a sqlite backend, so you can use scripts directly on the database which is handy.

I think feedly has features to help filter less popular articles