Hacker News new | past | comments | ask | show | jobs | submit login

Bluesky also doesn't show me videos of people dying every 10 minutes.



Bluesky doesn't have a working only followed feed.


What's wrong with the "following" feed? Works for me.


Same. And Bluesky also allows you to bring your own feed algorithm, so if you don't like the discover feed you could make a better one! It's not a hypothetical either, I follow and use several custom feeds today (for example, one that shows post by people who follow _me_, so I can discover people to follow back).


It shows couple days worth of posts by people you follow and then turns into discover feed.

While OnlyPosts feed shows posts only from 1 to 4 days.


Infra team recently switched the timeline backend to a new cluster and didn't backfill. It will regrow over time.

Sorry about the confusion that caused. It was one of those things that happened in the scaling rush.


Literally never had this issue on X.


This is far from the first time I've seen it mentioned. Seems quite possible that if you're a heavy user it has enough source material for your recommendations that they'll be good, but for less regular users it'll draw from whatever is 'trending'.


I don't either, but that's because I stopped using Twitter ages ago. It's a very reliable solution to all of Twitter's problems.


Because there are no algorithms yet.

I try to follow the situation in Ukraine and I see lots of videos on X with russians murdering ukrainians as well as their own.


> Because there are no algorithms yet.

One of the major selling points of Bluesky is that you get to chose your own algorithms (or rather, chose what algorithms you want to use, that others built).

So there are way more algorithms than any other social media, today. Only difference is that with Bluesky you get to chose which one you use.

https://bsky.social/about/blog/3-30-2023-algorithmic-choice


Bluesky has a number of algorithmic feeds. For instance "What's Hot Classic".

And while that feed doesn't have violence, it does have a surprisingly high number of "dudes holding their cock" and furry-related images, which seems like something that should probably be filtered out or at least tagged.


Well, no, violent content is against Bluesky's terms of service so it would be removed if flagged. Twitter has no such restriction so violent content spreads widely.

I understand the free speech principle at work but from a purely pragmatic point of view I simply don't want to see that stuff in my day to day browsing of social media.


X doesn't do that for me either... If you see posts you don't like, block/hide them. X learns.

That's the cost of a free platform though. Rather have that than being banned for thinking wrong.


Except that it is exactly the same, as Twitter will use those signals to derank globally the users whose posts are blocked too often.

Meanwhile, Musk himself is banning people for thinking what he decides is wrong.



(apologies for AI summary of the above articles; was just trying to be helpful!)

Elon Musk's tenure at X (formerly Twitter) has been marked by a series of actions that starkly contrast his self-proclaimed stance as a "free speech absolutist." Despite his vocal advocacy for unrestricted expression, Musk has overseen a significant increase in account suspensions and content restrictions, often targeting critics and journalists.

Notably, X has suspended numerous journalists who have reported critically on Musk or his ventures. For instance, in December 2022, the platform suspended accounts of reporters from The New York Times, The Washington Post, and CNN, among others, who covered Musk's activities. These actions have raised concerns about the suppression of dissenting voices and the erosion of journalistic freedom on the platform. Associated Press

Furthermore, under Musk's leadership, X has shown a heightened willingness to comply with government censorship requests. During the 2023 Turkish presidential election, the platform restricted access to certain content within Turkey, a move that critics argue was influenced by Musk's business interests in the region, including SpaceX's satellite launches in collaboration with the Turkish government.

These actions suggest a departure from Musk's professed commitment to free speech, revealing a pattern of behavior that prioritizes personal and business interests over the principle of open expression. The increasing suppression of critical voices and compliance with authoritarian regimes' censorship demands highlight a concerning inconsistency between Musk's stated ideals and his actions as the steward of a major social media platform.


Blue sky supports federation, and X is selectively censored by the top level of the US executive branch.

How do you define “free”?


In practice, Bluesky is a fully centralized system. You can run your own server to feed posts into the single centralized system.


You can also choose your own moderation, algorithmic feeds, and frontend

The API is open and free too, making it much easier to build on the network. The system is designed so that we can design new record types and create all sorts of applications on ATProto. It's going to go well beyond just a Twitter clone and each app shares the same social fabric, instead of having to attract their own, which is going to accelerate the network effect

Is it perfectly federated, no, there are weak points that will be made less weak with time. It's not a reason to put it down


The moderation is mostly federated. You can choose which lists to subscribe to to mute/block/label as you see fit.


I created a parody Donald Trump account as an experiment. It was deleted for "impersonation". Are you saying if I had run it on my own PDS, other people would get to choose whether it was deleted for them or not?


I obviously can't speak to your specific situation. As I understand it, both X and bluesky require parodies to be labeled.

From what I can tell, most content moderation on bluesky is done via labelers, which you need a pds to run, but not to use.

Both X and bluesky have semi-automated moderation processes with an inherent false positive rate. The two high profile cases I'm aware of (Jessie Singal and Laura Loomer) that were suspended automatically eventually got turned back on once a human was involved.

https://bsky.app/profile/jessesingal.com

https://bsky.app/profile/loomered.bsky.social/post/3lemtdvls...


No, they sent an email to the account's email address saying it had been deleted, and they deleted it.


Extreme content aside (well sometimes included):

Personally I found the "job" of managing who you follow / feed and pruning on twitter tiresome and at times impossible.

Bluesky has been better, maybe only due to fewer people, but if so then so be it.


In the end, the only thing Musk has done is cripple the spam filter of twitter.


Maybe, but I'm pretty sure that thing is not a lot of user's motivation ... at least as far as the discussion goes / their motivation to go elsewhere.

Side note: It is amusing how the current ownership's concern about bots vanished as soon as they took over.


Bluesky has a nuclear block which really helps.


Bluesky, or more accurately ATProto, has stackable moderation which is more sustainable in the long run

https://github.blog/news-insights/the-library/browser-monito...

tl;dr - anyone can run a labelling service, users have choice which one(s) they use (Bluesky &| 3rd party), competition and network effect will surface the good ones and weed out the bad ones.

This concept is going to be extended to community notes, sometime this year I suspect


Yeah. Think that got the broad strokes of the atproto architecture right for social media. Individual data stores -> aggregation/crawler layer -> services/app views. Feels like how search engines work with html web sites.

I think it will be more successful than Mastodon has been.


It's already 3x+ the size and growing faster, I would say it is already more successful in one regard. Next up is to build out the app ecosystem that APub has


Note: the reasons Bluesky is more popular than Mastodon have nothing to do with their technical architectures (except insofar as some people have learned helplessness about x@y), and everything to do with one being heavily marketed by a billionaire.

Though people do seem to prefer a consistent moderation policy, even if it's harsher. People don't like to have free speech about the things they want to speak about, and have another group of people choose to ignore it. They want to either speak to everyone at once or be blocked from everyone at once (at which time they make a new account).


I'm not sure what / which you are talking about in that second paragraph, but it does not describe moderation and preferences on Bluesky. ATProto has the best moderation scheme I have seen. It enables user choice and provider competition

The reason Mastodon | ActivityPub is not on boarding users like Bluesky | ATProto has a lot to do with the friction and discovery differences. ATProto was designed in part by understanding where ActivityPub had limitations and rough edges


I dunno about that. Bluesky deleted my entire alternate account parodying Donald Trump. Could I host it on my own PDS instead? Maybe, but they still own *the* relay.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: