Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Open Social

overreacted.io

Post reply on HN
Print view View original

61–70 of 409 posts

Re: Open Social

#61

Post by verdverm »

b_e_n_t_o_n wrote: post #12

I'm a simple man, I see a Dan post and I click. I'm a bit concerned that the open web only won because of first mover advantage. What gives me hope is OSS winning. I'd love to see something like atproto win though. It's clear that a major issue with social media is network effects preventing better apps from becoming popular.

> a major issue with social media is network effects preventing better apps from becoming popular.

One thing ATProto does is enable real competition in social apps, assuming they all run on the atproto fabric. One of the core hopes is that we can get everyone over to something like atproto once, to get them out of the silos, such that this is there last time they have to "move" their social network

Re: Open Social

#62

Post by jrowen »

tetrisgm wrote: post #18

I feel very conflicted about this work. On one side I find these ideas extremely compelling. This is aligned with the Indie web body of work, that pictures anyone having a personal website of their own content and ownership over that. And this page an article are beautifully put together. On the other hand, we haven’t really seen a lot of developers adopting these standards for their own projects (like using this for…

I am deeply concerned about the apathy people have towards the idea of ownership, openness and interoperability. It gives the idea that people just want to be fed TikTok and Instagram reels.

Can you expand on this feeling? Why is it deeply concerning? Why should people care about the abstract concept of data ownership? People were totally fine when they had zero ownership or agency over media and they were fed TV, books, movies, radio, etc. Most people do just want that, their primary motivation to engage with media is just to be entertained in that moment.

Now that they have places where they can publish stuff and their friends and family and maybe even some other people might see it, why should they care that they don't "own" their Instagram post, whatever that means?

Re: Open Social

#63

Post by emseetech »

numpad0 wrote: post #47

I'm starting to feel many of "next big Twitter to fill its power vacuum" projects are tackling the problem slightly wrong - they all perfect the Twitter feature set, then hit the wall with user growth and content deprivation chicken and egg problem. People gather where there are others and that's still around the rotting whale. That OpenAI timeline thing that just launched is more better approach, it solves content p…

Social networks rarely come up by being "the same but..."

They come up by doing something unique that can't be done on older platforms.

Re: Open Social

#64

Post by verdverm »

steveklabnik wrote: post #19
tshaddox wrote: post #13

The bit about aggregation is interesting, but it's not clear to me what the performance characteristics will be for very popular accounts. Presumably Justin Beiber's repo cannot be expected to handle 100 million WebSocket connections, all of which push out a message the instant he posts something. Is it vital to have more centralized hosts which can implement the sort of hybrid push vs. pull models that Twitter famou…

In atproto, those websocket connections aren't between users's repos, they're between an application and user's repos. Bieber has one connection per application doing aggregation, not per follower.

Relays also provide an important scaling building block, such that every app can listen to the relay, which listens to all the repos, instead of many apppds.

Re: Open Social

#65

Post by mozzius »

RobotToaster wrote: post #58
LightChaser wrote: post #10

Sadly, it's hard to imagine a world where something like this will ever catch on. The target audience for "traditional" social media is very different from the niche of people who want decentralized social media. Most people just use social media as a means to an end and don't really care about the systems behind it. If the answer is that most people should just make a bluesky account, that defeats the whole purpose…

AFAIK bluesky isn't even properly federated yet, everything relies on a single "BGS" router server.

Not true, there are many independent relays (one went online today, in fact: https://bsky.app/profile/upcloud.com/post/3lzqkrrqap22n).

This also completely misunderstands the architecture. Things don't hinge on the relays at all and they don't act as routers.

Re: Open Social

#66

Post by danabramov »

RobotToaster wrote: post #58
LightChaser wrote: post #10

Sadly, it's hard to imagine a world where something like this will ever catch on. The target audience for "traditional" social media is very different from the niche of people who want decentralized social media. Most people just use social media as a means to an end and don't really care about the systems behind it. If the answer is that most people should just make a bluesky account, that defeats the whole purpose…

AFAIK bluesky isn't even properly federated yet, everything relies on a single "BGS" router server.

This is misleading. I’m not sure if you’ve read the article so it’s difficult to elaborate — it aims to explain precisely that.

There isn’t such a thing as “Bluesky getting federated” — that doesn’t on its own mean anything. In Mastodon world, “getting federated” means many copies of the same webapp emailing each other. In atproto, you don’t create many copies of the same app. Instead, it’s shaped like the web — individual users can host their data in different places, and apps aggregate over that data. There’s no point in having many copies of the same app.

The BGS server you’re referring to is the “relay” mentioned in the article. Running your own relay is possible (Blacksky does it, as mentioned in the article). It costs about $30/mo with the current traffic. However, note that a relay is very dumb (it’s just a retransmitter of signed JSON over websocket). It’s cool that anyone can run one but by itself this isn’t a vanity metric to chase. We’ll probably see more independent relays but usually someone would run one for a reason — to insulate a company or a community from upstream failures, or maybe to censor things (in repressive governments).

Re: Open Social

#67

Post by verdverm »

brap wrote: post #42

I think we tend to do a lot of idealization. The vast majority of non-techies don’t care whatsoever about decentralization.

One can explain features that are possible on ATProto but not Big Social without talking about "decentralization". My pitch to the non-technical typically follows or covers these points

- Social today is not healthy

- Single account instead of N

- All apps keep your data in your database

- User level choice over apps, algos, moderation. Esp algos, my social media usage patterns have changed for the better since I started using custom feeds

- Real competition in social media

- Take back our shared digital experience from a handful of billionaires deciding everything and keeping us locked into their attention economy

Re: Open Social

#68

Post by dgaffney »

Thank you Dan for the post! I think two other things to point out:

1. Because open social has to actually compete for a user's business, any sufficiently mature platform build in the ecosystem will necessarily trend towards being more responsive to those users needs, which will trend towards a better product than the legacy crop,

2. Precisely at a moment where governments lean on large, visible corporate entities to enact desired policies, splintering that ownership helps ensure a resilient communications network

Re: Open Social

#69

Post by dgaffney »

jrowen wrote: post #62
tetrisgm wrote: post #18

I feel very conflicted about this work. On one side I find these ideas extremely compelling. This is aligned with the Indie web body of work, that pictures anyone having a personal website of their own content and ownership over that. And this page an article are beautifully put together. On the other hand, we haven’t really seen a lot of developers adopting these standards for their own projects (like using this for…

I am deeply concerned about the apathy people have towards the idea of ownership, openness and interoperability. It gives the idea that people just want to be fed TikTok and Instagram reels. Can you expand on this feeling? Why is it deeply concerning? Why should people care about the abstract concept of data ownership? People were totally fine when they had zero ownership or agency over media and they were fed TV, bo…

idk if the normal user should necessarily care about data ownership, but I think the incentive structure it creates would be immediately legible to most people

Re: Open Social

#70

Post by a022311 »

mcny wrote: post #4

Now here is a controversial question... Can we have a free of cost top level domain? What are the actual costs associated with registering a domain? If let's encrypt can provide secure certificates free of cost, why can't a different no profit provide domains free of cost as well? It doesn't have to be pretty. It could be a UUID v7 stacked on top of another UUID v7 for all I care but it would be globally unique and a…

Umm ever heard of .tk domains?
Post reply on HN