A few reason not to have algorithms in nostr:

1. Higher signal from people you follow.

Without algos, you have to be selective about who you follow, and you’re more likely to curate that group of people often. Since content is not surfaced via some relational query, you just see what your circle posts.

Nostr without algorithms also resembles life more closely than other social networks. You have to literally be online to see what someone said. Sleeping? You may miss their note unless it is not buried enough that you can go back to it, or if they mention you directly. I think anything that more closely resembles real life interactions is ultimately healthier for us.

Of course that doesn’t mean you’ll miss everything. If the note is great, enough people will like or zap it that sorting can still surface it for you when you are back.

2. Better relationships from smaller number of direct connections. If the Dunbar’s number is to be trusted, we are limited at 150 people with whom we can maintain social connections. Without algos and curation of people you follow, you are more likely to form stronger connections. This in turn makes the network feel more intimate and friendly. Even as we speak I follow 178 people and it feels just about right.

3. Ability to step away without being punished. At least on Twitter, anecdotal evidence suggests that if you step away, your reach diminishes and you have to “earn” it back. Without algos on Nostr you don’t need to worry about this. Just step away as long as you need to - better overall for mental health. Come back when you like and you’ll still reach everyone who follows you. This is exactly how real life works - you enter a room of people and they will all hear you instead of a tiny fraction.

While other social networks optimize for engagement and time spent on website, I think nostr should optimize for maximum mental health and least stress.

Curious to hear ya’ll thoughts…

Reply to this note

Please Login to reply.

Discussion

The main reason I will consider algos is to discover content. I see Nostr as far more than just a decentralized twitter, possibilities are endless.

So, I agree with you, in my Twitter like client I am fine without algos. I just want my lil tribe to face the future with. But, if I am browsing for content (blogs, books, music, videos, research papers…) I foresee that an algo I can control could help me finding more stuff that is of interest to me.

Better than standard sorting? (Most liked, most zapped; liked by friends, etc…)?

Possibly, it’s all possibilities for now. I am curious to see what algos could be capable of doing if they are coded by people with good intentions.

Decentralizing algos development would surely provide some interesting outcomes I believe.

I think it would only centralize over time and go back to exactly what some seek to escape

Do you have the same fears for #nostr as a whole?

Not as much, no, because we’ll have so many clients to choose from.

So why wouldn’t that be the case with algos? We could have many algos to pick from, or even build our own combining predefined blocks from algos service providers.

Well, there’s no way to stop anyone from adding algos to clients. I’m just making a case for why you should probably think about it first…

Fair point. Let’s hope to have clients and relays who will leave that choice to the user. If that is possible.

I am not a coder, so when I browse the NIPs I might get confused. I thought algos would be something a user could control.

Sort by zaps , sort by reactions within, sort by comments and new. And also set time period (default 24hours)

I think these four filters/sort should suffice discovery

I feel that is a waste of all the tags and metadata that will be available. But, obviously if I cannot chose/control the algo, I do not want it.

Even in its simplest form, I could design an algo that gives parameters for all the things you have listed and combine zaps, retweet, likes, comments and more in one feed you configure to your linking.

Similar to Reddit *Hot* section

Agreed what you’ve mentioned works so why change anything? If it’s not broken why fix it? Real is rare and it’s working here 💜

But I don’t think this is something you can avoid. I mean.. Nostr is a protocol.. so you won’t have an Algorithm In Nostr but you may have algorithms in some clients

Then people will chose .. if there are clients with algorithms that improve the experience to the user then people will use it.. otherwise people will go back to their raw untouched feed

Oh yeah for sure, can’t stop it. I’m just hoping someone hears me out and doesn’t build into their client, so I can use it

An algorithm that I could consciously press the button on and be shown a number of posts I might find interesting would be complementary to what you are saying I think.

Real human curation and connection at the core, certainty that this isn’t being fucked with, and then add-on features for discovery.

That’s how Apple Music works and I like it. I curate a library but enjoy letting the AI DJ from time to time.

“Surprise me” button

Bingo. “Create a feed based on…”

Also don’t forget this public posting style allows for direct interaction with prominent people, which is a very compelling and interesting thing that real life doesn’t make possible

I agree with everything you said.

#[2]​ , really curious to see what you said but keeps coming up blank.. What relays are you on?

NM, I got it..

Totally agree, human interaction is the best algorithm 🫡

I was using Twitter for years without algo, it says I follow 0 people and my For You tab is empty but I follow many people through curated lists. The only useful algorithm for me would be "What did I miss?" and it should only be from my lists. I'm working on bringing this to Nostur.

Ah yes lists. I’ve heard of people doing this but was never able to get used to it. Wish it was the default behavior on Twitter.

Nostrgram.co has lists already. Also has Following+ which is your followers and their followers which is an expanded view without the cruff of Global

Continual algo development is the same central planning slippery slope that made all legacy social suck.

💯

My feedback: YES.

>> Of course that doesn’t mean you’ll miss everything. If the note is great, enough people will like or zap it that sorting can still surface it for you when you are back.

Isn’t that ranking just an algo?

I don’t think algorithms are inherently bad. What’s bad is if they are forced upon you and you have no choice whatsoever. Those ones are built entirely to profit maximize.

I guess it depends on your definition of an algorithm. I don’t think of sorting as anything advanced 🤷‍♂️

The nice thing about nostr is that it doesn’t matter what I say, people will build it and everyone will have a choice. I suppose it would be less harm if you can choose which algo to use at a personal level.

💯

But has to be time based ideally which is another parameter, and then you take out spam/blocked which is another, and then so on and so forth until you’re working towards “relevance” and ✨

I see your point but it’s a slippery slope.

They aren’t but with years they probably would be exactly in the same place Twitter algos are right now…

Are we not corruptible? Single people may be stronger, but corporations that will run many clients will make worst algos possible. Maybe the clients that stay in individual hands will still manage to have decent algos, but… 😬

Ainda não entendo pq temos o “like” se hj temos o ZAP

Some people use likes as an acknowledgment of having read a note. It doesn’t mean they agree with it, just sort of nodding to let you know they heard you.

Very true

[782887]

Com a opção de editar o Zap para valores simbólicos como o tradicional 21 Sats não faz diferença e evitaria robôs dando like para manipular resultados de engajamento

Zaps are easy to manipulate too. But besides that fact, most majority of users don’t have wallets connected last I checked, and have no way of expressing liking or acknowledging having read a thing. They might just leave if it’s not there where they’d expect it.

Agora entendi… faz sentido!

How cool is it that we’re having a conversation in different languages!

Isso muda tudo, tbm acho incrível!

This is the best. Tower of Babel

So interesting, was that in your dream from sleeping?

👀

Percebi, mesmo com tradução, a forma de se expressar oriental usa muitas metáforas difíceis de ser compreendidas pelo ocidente!

东方人彼此之间也看不懂,比如我大概只能读懂一半翻译成中文后的日语。

True but it’s sometimes enough to get the gist of what is being communicated.

O Chinês em alguns momentos parece poesia… com o avanço das AIs espero um tradutor de metáforas ou verbetes de cultura local 😅

すごいな~

Yes!!!

#[4]​ achamos o seu irmão perdido!

Mesmo assim um zap e um like pela resposta.. 🙂👍

I want lots of different tools and ways of sorting and sifting. For me Lists (in Nostrgram) is a good low noise approach, but I also want to be exposed to something new. I would be okay with an Explore tab that tells me the criteria it is using and allow me to tweak it, that is very powerful. We are using Algos every day, they are not all bad. I would like a chat box where I could say "Summarize the current sentiment of the US Dollar in Turkey" and get a selection of notes. I don't think we should think of Nostr in a twitter framing.

TLDR: there should be lots of clients doing lots of different things

I agree, tough it would be nice to have the ability to selectively prioritize a few people, making sure you get notified when they post and being informed the next time you log in.

In theory it could work with push notifications on mobile as well, but i would assume that would take up too much power to be useful, unless someone were to run a server that did all the requests on nostr.

Agreed to everything. If you must have algo, I’ll suggest that to be an additional feature, and pick your own algo. Let user select what they want in the mix.

If algos don't push content it also means you need to work harder to promote other people's notes to see quality stuff. This is a healthy motivator to enlarge the overall participation and make passive spectators active participators.

Good point