←back to thread

I ditched the algorithm for RSS

(joeyehand.com)
744 points DearNarwhal | 3 comments | 16 Jan 25 12:18 UTC | HN request time: 0.413s | source
Show context
bluGill ◴[16 Jan 25 19:11 UTC] No.42729542[source]▶
A good algorithm is a good thing. However what a good algorithm is for me is often different from what it is for those who maintain them. Outrage gets attention and sometimes it is needed, but there is a level of too much, and also a lot of outrage unfairly represents the issues and so it makes me mad even though if I understood the details I wouldn't be mad just concerned.

I want an algorithm that surfaces things of interest to me, then says "you have seen it all, go outside" (with an option of if I'm confined to a hospital bed to go on). Algorithm maintainers want me to keep scrolling for more ad dollars.

replies(11): >>42729694 #>>42730030 #>>42730343 #>>42731465 #>>42732223 #>>42732697 #>>42733474 #>>42733961 #>>42735114 #>>42764093 #>>42764671 #
1. EvanAnderson ◴[16 Jan 25 23:22 UTC] No.42732223[source]▶
I'd like to see an RSS item-level recommendation / discovery algorithm driven by recommendations made by a cohort of people who have similar likes to me. I wish there was a standards-based to publish a stream of my "likes" from my feed reader, and to "consume" the "likes" of others. When I add somebody's blog to my feed reader I'd expect it to pick up their "likes" (much like how people used to have a "blog roll" on their own site) and begin to consume the "feeds" from the sites they "like". It reminds me a bit of PHP's "web of trust".

Hacker News, arguably, functions in this capacity for me now. The cohort is the entire population (since we all see the same item rankings), though.

replies(2): >>42733680 #>>42739780 #
2. lonk11 ◴[17 Jan 25 03:07 UTC] No.42733680[source]▶
What you are describing is similar to how https://LinkLonk.com works (my side project) - when you "like" a link you get connected to the RSS feeds that posted that link and other users that also liked it. Then you get content from feeds and users that you are connected to. The more links in common you have with a feed or a user the more weight their other links have.
3. mjmsmith ◴[17 Jan 25 16:30 UTC] No.42739780[source]▶
Do any of the web-based readers offer discovery based on matching your RSS list to other users' lists? Mine (Feedbin) doesn't, but I'd like it as an opt-in option.