This just won’t work. If RSS becomes popular, there will be discovery platforms with “algorithm”s. It will be the same thing, just the discovery and content separated.
RSS appears good now only because it’s not popular enough for LLMs to meddle with. I don’t use RSS, so I don’t really mind, but those who use RSS are making disservice to its _purity_ by trying to popularize it.
RSS is just one element of the ecosystem - the input.
I envision that the filtering mechanism CAN use any rules - hand-written, heuristics, old-school machine learning, LLMs. Just with a key difference - you are the one controlling it. No hidden tricks to make you "engaged" (read: addicted) or "sold".
If you feel it is too much politics, you reduce it. If too little - add. If you want less clickbaits and intellectual fast food, you filter it. Etc, etc.
> it’s not popular enough for LLMs to meddle with
About that, I was sad to see that TDMRep [1] doesn't provide a way to signal reservation for RSS feed, so it has to be done at the HTTP level, otherwise the same content delivered in RSS feed can be legitimately scrapped and mined even if the author opted-out using an HTML meta tag on the website.
[1] https://www.w3.org/community/tdmrep/
> If RSS becomes popular, there will be discovery platforms with “algorithm”s.
So? If plain RSS exists, then you can still consume it the way you want.
I'd like to remind that when RSS was really popular we had "planet" aggregators everywhere, where someone interested in particular topic bundled posts from multiple people.
RSS exists but those authors who don't publish through it probably wouldn't care about it either. Like, if by magic, RSS became popular as a technology, they would publish through it, but then there would be demand for discoverability and algo feeds would win the engagement race and then RSS is in the background and th platform would naturally decide to just focus on the algo and drop RSS and the regular users wouldn't care and authors would only care what regular users care about. Except for the tiny techie bubble.
It's not a technical problem. Less effort will always be more popular and drown out more effort in the mainstream.
Imagine if you could order completely free McDonald's food to your doorstep anytime and could also choose to cook your meals at home. Guess what portion of people would choose which option.
You don't need "that technology to become popular" to make it even more popular. It already was popular enough and it already worked.
Your whole comment makes no sense to me. Completely confusing.
Who are you arguing with? Why RSS has to compete with anything? Why do you even refer to it as "technology" - it's a text file people used to edit by hand in notepad. And maybe automate that with a script in their html editor.
It was popular, it's a fact. It was and is included in multiple blogging platforms. It was used by techies. It was used by non-techies. Learning curve was non existent and it was trivial to use on both ends.
What created friction was: killing the biggest RSS reader service that was free for all and killing very good support in browsers.
It used to be trivial - every browser was showing an orange button if site had rss. You could click it. You could add the feed to browser bookmark bar. It would display feed as nice bookmarks, downloading it live. This is what we lost - and we lost it because big companies wanted us to be entrenched in their socials. The rest was literally trivial.
Blogs kinda dwindled in importance as a whole. Substack brought it back to a degree, through email distribution, which is a more familiar technology to regular people compared to RSS. But even Substack is becoming more of an algo feed based social site nowadays.
You are talking about bookmarks and stuff but that's not how regular people use the internet. They open a handful of social media apps and scroll whatever is shown to them.
> there will be discovery platforms with “algorithm”s.
We already have platforms like feedly that has optional AI curated feeds.
I am more optimistic, good blogs will continue being there, and crap ones are no new invention or menace, be it LLM slop or Markov chain SEO babble content of 10 years ago.