Discuss.
rabble's avatar rabble
We need a way to do algorithms which learn from user interactions. Ideally an open system, that users can choose where to give their interaction data and what their algorithms are optimized for. But we can't just put our heads in the sand and ignore their existence. All of the A/B studies where users were given chronological timelines show they'll switch their time to other apps / services that do provide an algorithmic timeline. The question is how do we do it in an open, nostrway... If we don't then we get a social network owned and run by Sam Altman... which honestly is terrifying. He'd make Elon and Mark seem like balanced stewards in comparison. https://www.theverge.com/openai/648130/openai-social-network-x-competitor
View quoted note →

Replies (23)

X, IG, and TikTok already exist. Whoever wants that on nostr should explore building it. I have switched to nostr exactly because those algorithms are not shoved in my face. I don’t want an algorithm to learn what it thinks I like and feed me more of that. I will be staying with my chronological feed and seeing only content from my follows πŸ˜‚
someone's avatar
someone 6 months ago
Yeah, some AI is good, for example dodging those nude pics.. I switched to LLM analysing the notes of annoying users, it is not perfect but it is a start.. Locally run LLMs will be a thing in the coming months or a year. Coupled with discovery LLMs that run on servers this may be a thing on Nostr.
A lot of people on Nostr have AIDS (algorithm implementation derangement syndrome). They like the chronological algorithm, and don't think people with a different opinion should have a choice. I hope someday there is a marketplace of algorithms that are selectable in client, or even a simple interface to create your own. Having only a chronological feed is like having email with only an inbox, or storing your files without a folder structure. When you have a large amount of data to deal with, it helps to have tools to sort, search and manipulate said data.
That's how I feel. If people want it, great. They should have that choice. I'm happy with much simpler filtering mechanisms than what I'm assuming is meant by algorithm in this context. My online social experience has vastly improved. It can wax & wain like normal life and that is a very nice feeling. πŸ’œ
nostr is already open, i don't understand. > "All of the A/B studies where users were given chronological timelines show they'll switch their time to other apps / services that do provide an algorithmic timeline." what studies? there are studies on nostr clients?
Pretty bearish on "AltmanBook" getting off the ground... At this point, incumbent social networks have pretty big network effects. That means that there's going to be a comparatively large need for a significantly better product for people to leave. Nostr is fundamentally different than what any centralized platform can offer. AI algorithm blah blah ... sure... AltmanBook will still need to have actual human beings responsible for centralized content policy decisions, and that's where it'll fail like all the rest.
plor's avatar
plor _@plor.dev 6 months ago
I'm looking forward to @nextblock for this type of experience. I don't know if this fits in their roadmap, but the neighborhood feature seems like a way to jump around different algorithms, be they manual curation or possibly TikTok style. I personally sometimes want discovery where I see people I don't follow, but might be interested in based on some non-creepy algorithm. Other times I just want the newspaper (what are the people I follow up to). I hope we can support both modes and easily switch between them.
What about, hear me out, forgetting all about algorithms and going back to discovering/seeking/actively interacting more with other humans? Are we going to miss lots of stuff? Absolutely yes, but Hakuna Matata, what’s this constant fomo? Let’s go back to less consumption and less discoveries, and the few we find will be remembered and valuable
richard's avatar
richard 6 months ago
100% agree, it seems like they can't grasp that possibility unfortunately. always screaming "let's get bigger and better, we'll win!" against big tech... damn near impossible, what a waste of time.
richard's avatar
richard 6 months ago
how about having no feed?!😨 mic drop. lets start feeding ourselves rather than getting fed. search and discovery tools are fine, but we are on a decentralized network, there will never ever be an algorithm as good or better than legacy social media algorithms.
richard's avatar
richard 6 months ago
i mainly meant no algorithm feed. but beyond that, maybe no feed period is the way to go. i don't know the how yet, but i think there's so much left to explore in UX/UI. feeds are restrictive and by design addictive.
richard's avatar
richard 6 months ago
im not against algorithms and exploring new ways to find content, i agree they could be necessary as there is a clear need from users to sort content. all i said is that we're not going to win that race against big tech algorithms and that our focus should probably be somewhere else. we want nostr to thrive as a decentralized network and algorithms are not compatible with that as much as you all want it to be.
richard's avatar
richard 6 months ago
i hear you, but rn when people don't like what they have, they already go out of their way to "rebuild" their algorithm by liking or viewing different types of content on purpose. (i think a lot feel like they can control it to an extent) and on the technical side, it seems that it would be so much harder to have algorithms on nostr anywhere close to how efficient the big tech ones are. so i do find it interesting that we're looking for better algorithms and more choice, but in the long run is it actually worth it or are we wasting our time? imo the outbox model is the future of nostr, and algorithms do not like that model. nostr does not thrive with algorithms.
I'm not looking for "better" or "more efficient" algorithms than big tech, however you define those terms. I want to control them, for myself. It seems like your argument is "I don't think it's possible so don't even make the attempt. It works for me as is so it should work for everyone." I reject that premise. Have a great day and enjoy Nostr as you wish. Other people may want to use it differently.
richard's avatar
richard 6 months ago
my argument is what i wrote, not whatever it seems to you. have a good day.
I get the commercial aspect that we should give users what they want (and most want algos that make them discover what they like), but like you said we cannot compete with big tech on this field, we should focus on other aspects unique to nostr (but then maybe not enough users will be interested in non-algo platforms with consequent insufficient revenues?)
↑