TL;DR With traditional mail, if you want to invite your friends to a party, you visit their homes and leave a copy of the invitation. This is roughly how ActivityPub operates. In contrast, with...
TL;DR
With traditional mail, if you want to invite your friends to a party, you visit their homes and leave a copy of the invitation. This is roughly how ActivityPub operates.
In contrast, with ATProto, a copy of every message is sent to everyone, and it's their responsibility to check which messages interest them. This is where relays come into play.
Anyone looking to host an ATProto relay will need to store as much of the network's data as possible to avoid missing messages. This leads to ever-increasing storage costs, which are prohibitive for the average person.
Even with self-hosted identity servers, BlueSky has control over the user keys, so it may never be entirely possible to escape its influence.
BlueSky is built by good people who understand and are working on this problem—federation and decentralization—but, for the time being, it is neither of those things.
Agreed. I went to Bluesky, created an account, and can now talk to everyone. Just like Twitter, Facebook, etc. On Mastodon, I had to do some research. What instance do I want to join? Where can I...
Agreed. I went to Bluesky, created an account, and can now talk to everyone. Just like Twitter, Facebook, etc.
On Mastodon, I had to do some research. What instance do I want to join? Where can I find lists of instances? What are their policies? What are their users like? Who do they federate with? Will I be able to find the content I want? Will I be able to talk with friends and peers from the instance I choose?
Then if you want to set up Mastodon on your phone, which app to use? Oh wait, I need to know the server address as well? Just remembering username and password is hard enough for some people.
There are too many decisions to be made, and each of these decision points is also an offramp from Mastodon altogether.
At least mastodon.social, the flagship instance, serves as "default" entry point. But again, that doesn't answer all the questions from above.
Just cherry-picking that one thing makes you miss the overall point by a pretty large margin though. Mainly because: It is one aspect of multiple things brought forward in the comment you replied...
Just cherry-picking that one thing makes you miss the overall point by a pretty large margin though.
Mainly because:
It is one aspect of multiple things brought forward in the comment you replied to. In isolation things are always easier/simpler, but they don't exist in isolation to begin with.
Even if this was the only point, if it is a little bit more complex than what people perceive as equivalent platforms many people will see it as barrier and go for the easier option.
Most of the time a login username is an email address, so people who struggle to remember a username/password combo are generally already struggling with "email plus one other thing". Adding more...
Most of the time a login username is an email address, so people who struggle to remember a username/password combo are generally already struggling with "email plus one other thing". Adding more things to remember is adding more friction regardless of how memorable those additional things are in a vacuum.
And the fact that it is a single weak link run by a corporation is also why it will succumb to the same enshittification cycle that comes to all corporate social media platforms. I have accounts...
And the fact that it is a single weak link run by a corporation is also why it will succumb to the same enshittification cycle that comes to all corporate social media platforms.
I have accounts on both. More of my social circle is on BlueSky, but overall I still feel safer on Mastodon, and if I were a betting person I would put money on either Mastodon or a compatible successor outliving BlueSky.
As opposed to the multiple weak links run by individuals AND corporations that caused other federated/decentralized services to succumb to the same enshittification cycle that came before? Let's...
And the fact that it is a single weak link run by a corporation is also why it will succumb to the same enshittification cycle that comes to all corporate social media platforms.
As opposed to the multiple weak links run by individuals AND corporations that caused other federated/decentralized services to succumb to the same enshittification cycle that came before?
Let's not pretend that something's protocol and the way it is structured as a service still isn't affected by the people behind it all running said service.
I'd argue decentralization poses MORE risks to data and how it is used as by its very nature, all of your data has to be duplicated multiple times over by each and every service that has access to it. It only takes one to not honor deletion requests. (Which has happened on multiple Lemmy instances where they refused to delete my account.)
There has also been multiple Mastodon and Lemmy services that all had to be Fediblocked due to bad actors or disappeared without a trace after people dug into who was behind the instances... a company like BlueSky is going to be a lot more careful about burning its reputation let it ends up becoming another X or Threads in a lot shorter amount of time before it can even pull off any further substantial funding rounds.
So far, it's been hard to complain about how Bluesky is handling things. Right now it's not much more than an extremely self-selecting and curated twitter clone, but that's all it has to be to be an improvement over Xitter itself. That can always change and users will bail when it does. That's the circle of social media life.
Honestly. We should just go back to self hosted blogs and webrings. Finding a related site always seemed to be more impactful when you knew you were just a few degrees away from personally knowing the person involved in it.
Here's some info about it: https://techcrunch.com/2024/10/24/bluesky-raises-15m-series-a-plans-to-launch-subscriptions/ I look at Bluesky as it another Twitter clone. It's not yet fkd up as...
I look at Bluesky as it another Twitter clone. It's not yet fkd up as Twitter, but as every other proprietary and corporate-owned IT product, especially corporate ocial media, it will be.
If one wants to use really human-centered and made by people for people social media, they should stick to something from Fediverse. Yes, it may be more difficult to choose server for the account, however I see it as a feature, I can ask my friends where are they and register there, or I can just look for thematic servers and register there. Also, since there are others willing to help, you can just ask and people will help.
Speaking as a non microblogging social media user, cause I personally dislike all of them, it was taking too much precious time from me.
I can't do this because none of my peers or friends or colleagues have heard of it more than in passing. I don't have real life people to ask about it either. And I'm ultimately not that...
I can ask my friends where are they and register there,
I can't do this because none of my peers or friends or colleagues have heard of it more than in passing. I don't have real life people to ask about it either. And I'm ultimately not that interested in that much work.
People had heard of Bluesky before the great migration, but weren't sure about making the pivot, especially early on when you could run out of Following easily.
The current debate I'm seeing FWIW is over verification and why Bluesky should or shouldn't do verification instead of the current self-verification method. (I land on shouldn't myself)
(I don't really care if they sell subs if they keep it to the benefits like longer videos and the like, people have wanted to support Bluesky for a while.)
They are registered as a public benefit corporation, so they have more legal obligations than just a fiduciary duty to their shareholders. This doesn't make them immune to enshittification of course.
They are registered as a public benefit corporation, so they have more legal obligations than just a fiduciary duty to their shareholders. This doesn't make them immune to enshittification of course.
I remember app.net in its heyday (https://en.wikipedia.org/wiki/App.net) and whilst novel and a refreshing alternative to Twitter, it eventually collapsed due to lack of revenue or recurring...
Bluesky is built by good people who care, and it is providing something that people desperately want and need. If you are looking for a Twitter replacement, you can find it in Bluesky today.
However, "credible exit" is a reasonable term to describe what Bluesky is aiming for. It is Bluesky's term, and I think Bluesky should embrace that term fully in all contexts and work that they can.
I remember app.net in its heyday (https://en.wikipedia.org/wiki/App.net) and whilst novel and a refreshing alternative to Twitter, it eventually collapsed due to lack of revenue or recurring subscriptions. See the Wikipedia article:
On May 6, 2014, the founders announced that subscription renewals had been so poor that there were no longer funds to retain development staff for App.net and future operations would be on a maintenance-only basis using contractors
I really wish there were a way to let software be “done” in the modern era. As in, it has implemented all the features people really want, all the major bugs are ironed out, and the only updates...
I really wish there were a way to let software be “done” in the modern era. As in, it has implemented all the features people really want, all the major bugs are ironed out, and the only updates are dependency version bumps (made easy via rigorous automated testing).
For whatever reason, it seems like there’s a pressure to continually add things until the product is crushed under its own weight or the money runs out. Iirc this happened to cohost too.
The problem I think is (1) software even when "done" requires ongoing maintenance due to changes in the environment (for example you at the very least often have to recompile it for new operating...
The problem I think is
(1) software even when "done" requires ongoing maintenance due to changes in the environment (for example you at the very least often have to recompile it for new operating system versions), if not regular bug fixes (most software is full of bugs).
(2) internet software requires that someone is around to host the service. Federation partially solves this problem, in that other people can go host an instance, but practically speaking you need a lot of money to do so.
I feel like the second issue is one of those things that is just invisible to most people and doesn't cross their mind. Even somewhat technical people get this wrong very often. Certainly when...
I feel like the second issue is one of those things that is just invisible to most people and doesn't cross their mind. Even somewhat technical people get this wrong very often. Certainly when media like photos and video are involved the costs of hosting something at scale go up quickly.
If the second issue of hosting was not as big of a factor I feel like we would already have open source alternatives to many platforms that are currently the monopoly of big tech companies.
Re. The second issue, agreed that media hosting costs (typically: egress and long tail, permanent, nearline storage) are what make self-hosted services expensive. Imo, but the existence of...
Re. The second issue, agreed that media hosting costs (typically: egress and long tail, permanent, nearline storage) are what make self-hosted services expensive.
Imo, but the existence of BitTorrent and scene groups feels like a solution that never became mainstream. Ie. users can be responsible for their own content, but that fundamentally alters their experience of the product in ways that would present steep adoption blockers to most folks.
Decentralized solutions aren't really solutions for platforms where you will need things like moderation. At the very least it they introduce extra complexity in that area. You can already see...
Decentralized solutions aren't really solutions for platforms where you will need things like moderation. At the very least it they introduce extra complexity in that area. You can already see this to some degree with lemmy, mastodon, etc.
Even ignoring all that, from a technical point of view I am not sure p2p solutions are actually viable.
Heck, even torrents themselves are a good example as many of them would not work without dedicated high throughput seed boxes. At which point they are effectively still hosted but with just a extra step.
I do have a response but it’s taking a while to write! In the meantime, without citations on my phone whilst procrastinating from work: sorry for the minor correction, but p2p networks are...
I do have a response but it’s taking a while to write! In the meantime, without citations on my phone whilst procrastinating from work:
sorry for the minor correction, but p2p networks are distributed, not decentralized. These are some formal networking terms that refer to how nodes connect to each other, so apologies for the pedantry; they really are genuinely different and have profound impacts on system architecture (eg NAT punching, super nodes, DHTs, etc.),
re. moderation I think we’re just discussing a media CDN, so you’d moderate a centralized platform the same way you always would, but resources would be hosted by users (BitTorrent ensures file integrity, for instance, so this isn’t a security risk),
I’ll have to dig up citations on this, and I don’t have much more than anecdote, but iirc seedboxed are mostly popular because symmetric fibre was hard to get and ISPs like to crack down on piracy every once in a while. For legitimate usages, and in 2024, I would put forward that seed boxes are unnecessary. Plus, we have stats from the 2010s which showed torrent traffic as being high competitive with Netflix traffic. Plus plus moving content to the edge has been a technical goal for the industry forever, and what’s edgier than your neighbour
Pedantry is fine, it is good to clarify that the technologies are different in various ways. I could also have been more clear in that as I do think that for the purpose of moderation they have...
Pedantry is fine, it is good to clarify that the technologies are different in various ways. I could also have been more clear in that as I do think that for the purpose of moderation they have similar issues where I even think p2p makes some things worse.
Specifically when it is about removing content, certainly when it is actively harmful is difficult to do that with both decentralized and decentralized platforms. It is not something that is entirely unsolvable I think, but it is a hurdle that needs to be taken into consideration. As you 100% will have to deal with bad actors otherwise who will abuse these mechanics for various nefarious purposes.
As far as it being 2024 and symmetric being more readily available, it will make it more viable to some degree but for many people around the globe fiber or any meaningful broadband connection isn't all that viable.
To be clear, I am not saying it is impossible but it isn't an easy solved problem either. If it was, we would see much more movement in this area as I know there have been several attempts to do distributed hosting of platforms in the past decades.
I’m really sorry for not putting together comprehensive replies >.< I have some other thoughts on your comment, but I wanted to pull this one out: I don’t think it’s trivially solved, but I’m...
To be clear, I am not saying it is impossible but it isn't an easy solved problem either. If it was, we would see much more movement in this area as I know there have been several attempts to do distributed hosting of platforms in the past decades.
I’m really sorry for not putting together comprehensive replies >.< I have some other thoughts on your comment, but I wanted to pull this one out: I don’t think it’s trivially solved, but I’m pretty sure the core technical problem is well understood and pretty easy these days. The near insurmountable hurdle is that social media platforms live and die based on network effects, and even the smallest hurdle (eg “remember to leave your computer on if you have a small user base, or subscribe to ‘pro’”) is enough to kick them off. Especially when the competition can burn VC money to unsustainably make storage and distribution AWS’ problem, until the cash runs out.
I think I get where you are coming from. But, I can't help but feel that if it truly was easy as you think it is we would have actual live examples. Mind you, examples that would have seen mass...
I think I get where you are coming from. But, I can't help but feel that if it truly was easy as you think it is we would have actual live examples. Mind you, examples that would have seen mass adoption but at the very least smallish communities of dedicated individuals.
On the federated side of things we have examples other such small initiatives that are having some moderate amount of success (mastodon, lemmy, etc) but I struggle to find anything that uses p2p.
I also know it isn't for a lack of trying, I have seen several attempts over the past few decades.
Of course, you can argue that early initiatives didn't take off because of bandwidth limitations as you mentioned earlier. But even then you would expect there to be something these days other than a list of dead past projects.
Simply put, if it was viable and fairly easy to solve I would expect people to be actively working on it these days in the same way people work on Fediverse projects. The fact that this doesn't seem to be the case combined with my understanding of the issue makes me fairly sure there is more to it than VCs being able to throw more money at it.
Oh, like, Peertube exists and you could view popcorn time/Kodi as other incarnations of this idea. That’s what I had in mind at least; may I ask if you feel they don’t? Eh … mastodon and...
But, I can't help but feel that if it truly was easy as you think it is we would have actual live examples. Mind you, examples that would have seen mass adoption but at the very least smallish communities of dedicated individuals.
Oh, like, Peertube exists and you could view popcorn time/Kodi as other incarnations of this idea. That’s what I had in mind at least; may I ask if you feel they don’t?
Simply put, if it was viable and fairly easy to solve I would expect people to be actively working on it these days in the same way people work on Fediverse projects. The fact that this doesn't seem to be the case combined with my understanding of the issue makes me fairly sure there is more to it than VCs being able to throw more money at it.
Eh … mastodon and activitypub were mostly dead until a couple years ago. It was a weird RSS/Atom successor that never went anywhere, until the pushback against Twitter and social media in general gave it some steam.
Overall I’d argue that CDN design is a more specialized skill set, so the folks who want to do this as a hobby and aren’t already burned out from doing it for a living are an empty set. … it’s not difficult, it’s just a less known thing, so the odds of finding someone to work on it for free is dramatically lower due to sampling sizes.
I don't consider popcorn time to be a social network, it effectively was just a fancy BitTorrent client with search and video playback build in. Peertube might be the closest thing to what we are...
Oh, like, Peertube exists and you could view popcorn time/Kodi as other incarnations of this idea. That’s what I had in mind at least; may I ask if you feel they don’t?
I don't consider popcorn time to be a social network, it effectively was just a fancy BitTorrent client with search and video playback build in.
Peertube might be the closest thing to what we are talking about as it integrates WebTorrent for videos where possible. This is somewhat limited to people watching the same video as far as I know so most hosting still comes down to the instance itself. Which in itself is another limitation to the idea I suppose. It is possible to have people keep a limited cache of things from the platform they visited but you can't expect them to store all of it for an indefinite amount of time. So the majority of content still will need to be hosted somewhere.
I am honestly curious now how much of the videos on Peertube is actually p2p provided and how much is directly coming from the instances.
While typing this out I also have come to a realization as to why Peertube didn't occur to me. It also is federated, which in my mind disqualifies it as I don't see federation as a solution in itself. It is also why I did focus on moderation earlier on. It is great in theory as you have all these different instance with their own focus, etc. In practice a lot of them are complete shit and there are no good tools to deal with bad actors coming in from other instances in a lot of federated protocols. You can de-federate but that comes with a lot of other issues, mainly that you then are dealing with a small community and might as well have something like Tildes.
Eh … mastodon and activitypub were mostly dead until a couple years ago.
I wouldn't say they were dead, just very small and used by a small group of users. But, they were in active development and got semi regular attention in the tech news, etc. For actual p2p distribution of content
Aah, I think that’s the disconnect — I’m mostly thinking of CDN/media distribution as an individual technical problem, but I think you’re considering how it would integrate into a platform? From...
Aah, I think that’s the disconnect — I’m mostly thinking of CDN/media distribution as an individual technical problem, but I think you’re considering how it would integrate into a platform? From my perspective, even YouTube could do this — ie, ask people to leave their computers on, since Google wouldn’t provide free distribution + video storage — so the question of success or failure of other usages of that CDN layer didn’t seem like the argument at hand. But perhaps it is!
While I love what torrenting enables, there's no way that competes with a service like YouTube. It's extraordinarily expensive to support the amount of data they have!
While I love what torrenting enables, there's no way that competes with a service like YouTube. It's extraordinarily expensive to support the amount of data they have!
2 problems: modern applications have dozens or hundreds of (transient) dependencies. Even if your app is "done", a security vulnerability can arise in any of those dependencies, and there is a...
2 problems:
modern applications have dozens or hundreds of (transient) dependencies. Even if your app is "done", a security vulnerability can arise in any of those dependencies, and there is a high probability that updating it would require non-trivial changes in your codebase.
people are attracted to shiny new things, and your competitors will add new features and evolve their design. Your app from ten years ago might still be totally safe and functional, but it will feel ancient and probably cumbersome to use when compared to an app that came out this year.
TL;DR
Ultimately, the fact that it’s not actually decentralized or federated is partly what allows blue sky to be successful now.
Agreed. I went to Bluesky, created an account, and can now talk to everyone. Just like Twitter, Facebook, etc.
On Mastodon, I had to do some research. What instance do I want to join? Where can I find lists of instances? What are their policies? What are their users like? Who do they federate with? Will I be able to find the content I want? Will I be able to talk with friends and peers from the instance I choose?
Then if you want to set up Mastodon on your phone, which app to use? Oh wait, I need to know the server address as well? Just remembering username and password is hard enough for some people.
There are too many decisions to be made, and each of these decision points is also an offramp from Mastodon altogether.
At least mastodon.social, the flagship instance, serves as "default" entry point. But again, that doesn't answer all the questions from above.
People still manage to remember their email addresses, a mastodon username isn't much more complicated.
Just cherry-picking that one thing makes you miss the overall point by a pretty large margin though.
Mainly because:
Most of the time a login username is an email address, so people who struggle to remember a username/password combo are generally already struggling with "email plus one other thing". Adding more things to remember is adding more friction regardless of how memorable those additional things are in a vacuum.
And the fact that it is a single weak link run by a corporation is also why it will succumb to the same enshittification cycle that comes to all corporate social media platforms.
I have accounts on both. More of my social circle is on BlueSky, but overall I still feel safer on Mastodon, and if I were a betting person I would put money on either Mastodon or a compatible successor outliving BlueSky.
As opposed to the multiple weak links run by individuals AND corporations that caused other federated/decentralized services to succumb to the same enshittification cycle that came before?
Let's not pretend that something's protocol and the way it is structured as a service still isn't affected by the people behind it all running said service.
I'd argue decentralization poses MORE risks to data and how it is used as by its very nature, all of your data has to be duplicated multiple times over by each and every service that has access to it. It only takes one to not honor deletion requests. (Which has happened on multiple Lemmy instances where they refused to delete my account.)
There has also been multiple Mastodon and Lemmy services that all had to be Fediblocked due to bad actors or disappeared without a trace after people dug into who was behind the instances... a company like BlueSky is going to be a lot more careful about burning its reputation let it ends up becoming another X or Threads in a lot shorter amount of time before it can even pull off any further substantial funding rounds.
So far, it's been hard to complain about how Bluesky is handling things. Right now it's not much more than an extremely self-selecting and curated twitter clone, but that's all it has to be to be an improvement over Xitter itself. That can always change and users will bail when it does. That's the circle of social media life.
Honestly. We should just go back to self hosted blogs and webrings. Finding a related site always seemed to be more impactful when you knew you were just a few degrees away from personally knowing the person involved in it.
I understand they recently got 15 million from a freaking blockchain company in 2024, so it's unavoidable
Here's some info about it: https://techcrunch.com/2024/10/24/bluesky-raises-15m-series-a-plans-to-launch-subscriptions/
I look at Bluesky as it another Twitter clone. It's not yet fkd up as Twitter, but as every other proprietary and corporate-owned IT product, especially corporate ocial media, it will be.
If one wants to use really human-centered and made by people for people social media, they should stick to something from Fediverse. Yes, it may be more difficult to choose server for the account, however I see it as a feature, I can ask my friends where are they and register there, or I can just look for thematic servers and register there. Also, since there are others willing to help, you can just ask and people will help.
Speaking as a non microblogging social media user, cause I personally dislike all of them, it was taking too much precious time from me.
I can't do this because none of my peers or friends or colleagues have heard of it more than in passing. I don't have real life people to ask about it either. And I'm ultimately not that interested in that much work.
People had heard of Bluesky before the great migration, but weren't sure about making the pivot, especially early on when you could run out of Following easily.
The current debate I'm seeing FWIW is over verification and why Bluesky should or shouldn't do verification instead of the current self-verification method. (I land on shouldn't myself)
(I don't really care if they sell subs if they keep it to the benefits like longer videos and the like, people have wanted to support Bluesky for a while.)
They are registered as a public benefit corporation, so they have more legal obligations than just a fiduciary duty to their shareholders. This doesn't make them immune to enshittification of course.
I remember app.net in its heyday (https://en.wikipedia.org/wiki/App.net) and whilst novel and a refreshing alternative to Twitter, it eventually collapsed due to lack of revenue or recurring subscriptions. See the Wikipedia article:
This is the age old problem.
I really wish there were a way to let software be “done” in the modern era. As in, it has implemented all the features people really want, all the major bugs are ironed out, and the only updates are dependency version bumps (made easy via rigorous automated testing).
For whatever reason, it seems like there’s a pressure to continually add things until the product is crushed under its own weight or the money runs out. Iirc this happened to cohost too.
The problem I think is
I feel like the second issue is one of those things that is just invisible to most people and doesn't cross their mind. Even somewhat technical people get this wrong very often. Certainly when media like photos and video are involved the costs of hosting something at scale go up quickly.
If the second issue of hosting was not as big of a factor I feel like we would already have open source alternatives to many platforms that are currently the monopoly of big tech companies.
Re. The second issue, agreed that media hosting costs (typically: egress and long tail, permanent, nearline storage) are what make self-hosted services expensive.
Imo, but the existence of BitTorrent and scene groups feels like a solution that never became mainstream. Ie. users can be responsible for their own content, but that fundamentally alters their experience of the product in ways that would present steep adoption blockers to most folks.
Decentralized solutions aren't really solutions for platforms where you will need things like moderation. At the very least it they introduce extra complexity in that area. You can already see this to some degree with lemmy, mastodon, etc.
Even ignoring all that, from a technical point of view I am not sure p2p solutions are actually viable.
Heck, even torrents themselves are a good example as many of them would not work without dedicated high throughput seed boxes. At which point they are effectively still hosted but with just a extra step.
I do have a response but it’s taking a while to write! In the meantime, without citations on my phone whilst procrastinating from work:
Pedantry is fine, it is good to clarify that the technologies are different in various ways. I could also have been more clear in that as I do think that for the purpose of moderation they have similar issues where I even think p2p makes some things worse.
Specifically when it is about removing content, certainly when it is actively harmful is difficult to do that with both decentralized and decentralized platforms. It is not something that is entirely unsolvable I think, but it is a hurdle that needs to be taken into consideration. As you 100% will have to deal with bad actors otherwise who will abuse these mechanics for various nefarious purposes.
As far as it being 2024 and symmetric being more readily available, it will make it more viable to some degree but for many people around the globe fiber or any meaningful broadband connection isn't all that viable.
To be clear, I am not saying it is impossible but it isn't an easy solved problem either. If it was, we would see much more movement in this area as I know there have been several attempts to do distributed hosting of platforms in the past decades.
I’m really sorry for not putting together comprehensive replies >.< I have some other thoughts on your comment, but I wanted to pull this one out: I don’t think it’s trivially solved, but I’m pretty sure the core technical problem is well understood and pretty easy these days. The near insurmountable hurdle is that social media platforms live and die based on network effects, and even the smallest hurdle (eg “remember to leave your computer on if you have a small user base, or subscribe to ‘pro’”) is enough to kick them off. Especially when the competition can burn VC money to unsustainably make storage and distribution AWS’ problem, until the cash runs out.
I think I get where you are coming from. But, I can't help but feel that if it truly was easy as you think it is we would have actual live examples. Mind you, examples that would have seen mass adoption but at the very least smallish communities of dedicated individuals.
On the federated side of things we have examples other such small initiatives that are having some moderate amount of success (mastodon, lemmy, etc) but I struggle to find anything that uses p2p.
I also know it isn't for a lack of trying, I have seen several attempts over the past few decades.
Of course, you can argue that early initiatives didn't take off because of bandwidth limitations as you mentioned earlier. But even then you would expect there to be something these days other than a list of dead past projects.
Simply put, if it was viable and fairly easy to solve I would expect people to be actively working on it these days in the same way people work on Fediverse projects. The fact that this doesn't seem to be the case combined with my understanding of the issue makes me fairly sure there is more to it than VCs being able to throw more money at it.
Oh, like, Peertube exists and you could view popcorn time/Kodi as other incarnations of this idea. That’s what I had in mind at least; may I ask if you feel they don’t?
Eh … mastodon and activitypub were mostly dead until a couple years ago. It was a weird RSS/Atom successor that never went anywhere, until the pushback against Twitter and social media in general gave it some steam.
Overall I’d argue that CDN design is a more specialized skill set, so the folks who want to do this as a hobby and aren’t already burned out from doing it for a living are an empty set. … it’s not difficult, it’s just a less known thing, so the odds of finding someone to work on it for free is dramatically lower due to sampling sizes.
I don't consider popcorn time to be a social network, it effectively was just a fancy BitTorrent client with search and video playback build in.
Peertube might be the closest thing to what we are talking about as it integrates WebTorrent for videos where possible. This is somewhat limited to people watching the same video as far as I know so most hosting still comes down to the instance itself. Which in itself is another limitation to the idea I suppose. It is possible to have people keep a limited cache of things from the platform they visited but you can't expect them to store all of it for an indefinite amount of time. So the majority of content still will need to be hosted somewhere.
I am honestly curious now how much of the videos on Peertube is actually p2p provided and how much is directly coming from the instances.
While typing this out I also have come to a realization as to why Peertube didn't occur to me. It also is federated, which in my mind disqualifies it as I don't see federation as a solution in itself. It is also why I did focus on moderation earlier on. It is great in theory as you have all these different instance with their own focus, etc. In practice a lot of them are complete shit and there are no good tools to deal with bad actors coming in from other instances in a lot of federated protocols. You can de-federate but that comes with a lot of other issues, mainly that you then are dealing with a small community and might as well have something like Tildes.
I wouldn't say they were dead, just very small and used by a small group of users. But, they were in active development and got semi regular attention in the tech news, etc. For actual p2p distribution of content
Aah, I think that’s the disconnect — I’m mostly thinking of CDN/media distribution as an individual technical problem, but I think you’re considering how it would integrate into a platform? From my perspective, even YouTube could do this — ie, ask people to leave their computers on, since Google wouldn’t provide free distribution + video storage — so the question of success or failure of other usages of that CDN layer didn’t seem like the argument at hand. But perhaps it is!
Steam does something a bit like this by allowing you to transfer downloaded games between systems on your local network.
While I love what torrenting enables, there's no way that competes with a service like YouTube. It's extraordinarily expensive to support the amount of data they have!
2 problems:
modern applications have dozens or hundreds of (transient) dependencies. Even if your app is "done", a security vulnerability can arise in any of those dependencies, and there is a high probability that updating it would require non-trivial changes in your codebase.
people are attracted to shiny new things, and your competitors will add new features and evolve their design. Your app from ten years ago might still be totally safe and functional, but it will feel ancient and probably cumbersome to use when compared to an app that came out this year.