it’s one thing to convince you to buy something and it’s an entirely different thing to condition to you accept that genocides are necessary sometimes; vote against your own interests or else the other team will win; propagandize you against foreign political enemies; distract you from global rings of oligarchical pedofiles who torture, rape, kill and eat children for funsies; and many other things.

like, i see what you mean but i’m afraid that modern society doesn’t see it that way. “moral relativism” (a.k.a. post-structuralism) states that there is no absolute truth, neither is there an absolute set of ethics; and as a consequence, it must be possible to negotiate these ethics on the “free market of ideas”, a.k.a the internet where big influencer institutions pay to sway your opinion. It’s all just a market game: Buy and sell opinions, and see which ones perform best as a consequence.

astroturfing

i mean that’s what it’s there for? always has been

like, how else do you think companies on the internet make money? by manipulating public opinion

Just to continue on this: Ok, i see now that the concept of “identity” has at least two aspects.

One is to prove that one has a certain private key, the other is to own a globally unique username.

A globally unique username can be derived from the DNS system (i.e. @gandalf_der_12te@discuss.tchncs.de is derived from the discuss.tchncs.de hostname). That is something that you cannot generate from your personal computer alone.

Thanks, i wasn’t aware of that.

I’d rather not do complicated algorithms to try to figure out interest / community adjacency based on user activity, as this could get really complicated

I agree that per-user recommendations could easily get out-of-hand in terms of computational expense and probably also it would be a bit intransparent if the algorithm is really complicated, which would undermine user trust.

I do think that maybe having a few explicitely-defined feeds/lists would be better than trying to dynamically and implicitely generate a list for each user.

So instead of having a “suggested for you (personally)” feed for each user, there could be a few large feeds like /feed/technology, /feed/science, /feed/events_you_can_attend_in_germany, … that people can view.

I’m not sure how computationally expensive such a feed would be. I imagine each feed is defined by a small text file that might look like this: (example feed_science.txt hosted on mander.xyz)

/c/science@mander.xyz
/c/publichealth@mander.xyz
/c/mycology@mander.xyz
...

And then the server can re-generate a cached version of the feed every 10 minutes. That is then served to each user who accesses /feed/science@mander.xyz . In this way, the feed only needs to be generated once every 10 minutes and is then served out to all users that want to view it. That means that feed-generation time is independent of number of viewers.

what if we filter ourselves into a bubble by choosing to use a site that does it?

hmm, i think you’re overthinking this. what if the recommendation algorithm simply gives you stuff from communities and you’ve subscribed to and “similar” communities (these would have to be linked from the original communities / link to the original communities)?

that should be reasonably easy and not involve any neural networks. i think basically it constructs a “feed” (post list) which is basically a remix of other lists (which are the individual communities that stuff is taken from), maybe weighted with a certain scalar factor.

https://quiblr.com/understanding_your_private_personalized_feed

Unfortunately it’s very spare with technical details. What are the technical details? How much network traffic does it cause for the servers?

that’s great :D

“Discuss interesting topics and join communities on the Fediverse.”

IMO it’s better than “link aggregator” because we don’t only share links here, but also images and text.

Also it would be neat if instance admins could automatically set a certain set of “subscribed instances/communities” and “blocked instances/communities” for each new user that signs up on their instance. That would make it easier for instances to give themselves a certain “flavor” so users have an easier time finding interesting stuff.

How about the instance selection wizard (click “join a server” on the homepage), which lets you select topics and languages to select instances. Do the current options make sense?

I do think the “topic”/“language” selection screens make sense overall.

Also what i noticed is that on http://join-lemmy.org/ when you select certain topics (like “Art/literature” and “german speaking”), then the resulting instance list is empty. It might be wise to show at least one instance for every search query, even if it’s not 100% fit. For example, at least display one generic-topic instance for your language or one english-language instance for your topic.

One user told me they got to https://join-lemmy.org/docs/users/01-getting-started.html when they tried to sign up instead of to https://join-lemmy.org/ so maybe it would be good to have an easily-visible link there that brings you to the main landing page?

maybe the instance admin can “force-flag” communities as political if the community mods refuse to do so themselves …

there already is a politicalmemes community, it’s the only politics community i haven’t banned because it’s the only one that actually manages to make me laugh :P

ATP synthase also happens during sleep.

frankly, i disagree with the “everything is politics” proverb. if i talk about ATP synthase, i don’t consider it politics.

that is indeed a good question. maybe if the community self-identifies as political? or maybe a majority vote? Or maybe it depends on the instance that you’re using? (something that the server admin can configure)

I think UUID might solve this, as it’s supposed to generate worldwide-unique identifiers:

The DOI system might also be an interesting approach to give each piece of content on the internet an instance agnostic link.

oh so that was that “test” comment for xD

i was wondering …