9 ms·
Mastodon and federated systems like it will not succeed beyond a tiny niche, will not replace twitter, and don't even solve most of the problems they try to sol
by gonehome 4y ago
Mastodon and federated systems like it will not succeed beyond a tiny niche, will not replace twitter, and don't even solve most of the problems they try to solve.
They don't fix the upstream issues that lead to centralization. Incentives mean most users end up centralized on a handful of servers with a worse experience. It's the worst of both worlds. As designed the UX is bad and cannot be fixed because of these issues.
Twitter has been a terribly mismanaged company for a long time and there's general consensus on this. Jack knew this, but was powerless to actually make the changes necessary to fix it. Elon may not succeed, but I hope he does and what he's doing is a necessary prerequisite. People dismiss his obvious success elsewhere in other insanely hard domains, they can largely be ignored.
I want a world where people have actual ownership over their computing, but federated system designs like Mastodon will not get us that and we have thirty years of failure to prove it. The press pretending otherwise because they hate Elon and see tech companies as a threat are just embarrassing themselves.
To solve this problem for real is a lot harder, but I'm optimistic it's not impossible. https://zalberico.com/essay/2022/09/28/tlon-urbit-computing-freedom.html https://zalberico.com/essay/2022/09/28/tlon-urbit-computing-...
- eftychis 4y agoWhy do you think/know Jack was powerless to fix things?
- gonehome 4y agoHe had only a tiny amount of company ownership, didn't have total voting power, the board (which barely even used Twitter) made it impossible for him to make the changes he'd want to make and just generally seemed obstructionist/bad. Jack has talked a lot about how making Twitter a company (instead of a protocol) was a mistake, he's talked about Web5 which lands on a lot of the same ideas as Urbit, but trying to tack it on top of the existing stack. He supported Elon taking power and being able to actually make the necessary changes and recently tweeted support of the layoffs too. Pre IPO Twitter was doing interesting things with API access and being more protocol like, but they killed all of that stuffed ads on it and went public. Ben Thompson talks a lot about how they're a terrible ad product and I think he's generally right. Subscriptions are a more interesting model. If you read the texts between Jack and Elon that came out during the recent legal fights over the sale you can see these things for yourself. That's somewhat of a rambling answer, but a lot of this stuff is publicly verifiable.
- expensive_news 4y agoHe’s kind of said it himself on multiple occasions. “Twitter as a company has always been my sole issue and my biggest regret. It has been owned by Wall Street and the ad model. Taking it back from Wall Street is the correct first step.” https://mobile.twitter.com/jack/status/1518772754782187520 https://mobile.twitter.com/jack/status/1518772754782187520
- Reventlov 4y agoYou mean, federated systems like the web or email won't succeed ?
- gonehome 4y agoCorrect - email is a failure as a federated system, people are centralized on a handful of providers and you can't even run your own server these days without getting blacklisted. The promise of the decentralized web of the 90s is a failure. The majority of users are centralized on handful of megacorporation apps despite the ad business making this pretty awful. We're trapped in a local maximum because of the underlying system designs. Spam, linux server admin being too hard, terrible devex are some of the forces that lead to this [0]. It's a hard problem to get out of without thinking from first principles. [0]: http://moronlab.blogspot.com/2010/01/urbit-functional-programming-from.html http://moronlab.blogspot.com/2010/01/urbit-functional-progra...
- Gigachad 4y agoIt’s not quite that bad. I ran my own mailinabox server on a vps a few years ago and didn’t have much issues with spam lists. Only gave up because I couldn’t be bothered updating the thing.
- bombcar 4y agoThat's the death of all decentralized systems, it's too much work in the long run. They work great when they're the ONLY option in town, but eventually people centralize on one aspect of it, as they did with email. Email used to be so decentralized that it was literally run on your machine (not your terminal, which was dumb, but the machine the terminal connected to) and even a single company department might have 10+ machines that could receive email (email addresses would look like name@machine.department.company.com). Then it devolved to one per department, and then one per company, and now one per massive IT provider.
- Gigachad 4y ago
- nathias 4y ago> It's the worst of both worlds. Yea, I've been saying the same for some time. I really want for decentralized solutions to proliferate, but federation is a terrible model.
- greenie_beans 4y agomastodon's ux is terrible, but urbit's ux is like 100x worse
- gonehome 4y agoI don't disagree, but the design of urbit is such that there's a path to it being good - that's not true of any other system built on the current stack (imo) because they don't fix the underlying incentive failures that lead to centralization in the first place.
- miked85 4y agoI'm pretty sure no one (including myself) even understands what urbit is.
- xeeeeeeeeeeenu 4y agoIt's very simple, this summary explains everything: https://twitter.com/urbit/status/1184984759123922945 https://twitter.com/urbit/status/1184984759123922945 Kidding aside, I think it's unlikely that a deliberately obfuscated project will conquer the world.
- bembo 4y agoomg that has to be the worst explanation of a piece of software that I've seen in a long while
- colin_jack 4y ago> As designed the UX is bad and cannot be fixed because of these issues. For fun I just tried signing up and the process itself is confusing, sluggish, buggy, and ultimately didn’t work. It feels like it’ll need to improve very quiclkly if it’s going to capture any significant amount of Twitter users.
- PebblesRox 4y agoThe slowness and bugs might be due to the sudden influx of people from Twitter. I just joined a small server that is now having to scale up a bit to accommodate all the new traffic, and I assume the more popular servers are seeing an even bigger bump. I was actually expecting the sign up process to be a lot more difficult than it was for me, just based on the impression I'd gotten from various HN comments over the years. I encourage anyone curious to try it out for themselves.
- colin_jack 4y ago> The slowness and bugs might be due to the sudden influx of people from Twitter Yeah definitely, feels like they should have seen this coming though. They have an opportunity that won’t come around again, and poor performance and confusing UX seem like pretty serious problems.
- ceejayoz 4y ago“They” describes a non-profit, low-resourced set of volunteers. Seeing it coming and being able to do something about it aren’t the same thing.
- colin_jack 4y agoI dunno, on performance they could just have limited number of registrations to what they can handle. Better that than having potential users getting a bad impression because everything is too slow.
- 4y ago
- bragr 4y ago>don't even solve most of the problems they try to solve I mean it does somewhat solve the problem of Musk running twitter into the ground. For most people, I imagine mastodon is a lateral move, not a step up.
- code_runner 4y agoI’m always surprised when there are still people acting like Elon isn’t just BEGGING to be hated. I’m still of the opinion he is not doing the work and his success is mostly throwing money at eager (and smart) engineers. That’s a smart thing to do, good call on his part, but it’s almost like he doesn’t know that’s how his companies work. Maybe he is just great at running tech companies or something but acting like the media/everyday people hate him for no reason is just dishonest.
- torginus 4y agoIf that's all he is doing - paying smart people and expecting them to work hard - why isn't literally everyone else doing the same?
- p1necone 4y agoIt might have something to do with the paying bit.
- code_runner 4y agoMy theory is that getting smart people to make twitter vs landing reusable rocket boosters… will yield VERY different levels of passion/interest. Which again is FINE. Personally, I just wanna do cool stuff and have a great work/life balance. I don’t want to (and don’t have the expertise to) land a rocket booster. I guess… not everyone’s problem domains are interesting enough to lure a team like space-x etc. again, I don’t like him, but credit to Elon for at least trying to solve these tough problems. Just admit FSD isn’t a year away!
- sillysaurusx 4y agoI don’t agree with them, but this counter argument doesn’t quite work. In a system with x% chance of success, we’d expect most people to fail for small x. But some few will still succeed multiple times. Throw in that your odds get better the more you’re successful, and it’s classic survivorship bias.
- code_runner 4y ago
- pjc50 4y ago> People dismiss his obvious success elsewhere in other insanely hard domains People also dismiss his failures in hard domains, like The Boring Company.
- unsupp0rted 4y agoIf he didn’t fail sometimes in hard domains then they wouldn’t be hard. He’s intentionally choosing to do things that have a relatively high likelihood of failure.
- cplusplusfellow 4y agoHe’s in the arena. And that’s what people respect. But has he failed at Boring?
- systemvoltage 4y agoI am glad someone is trying. I don't see the likes of Sundar Pichai doing anything revolutionary.
- BurningFrog 4y agoThe Boring Company hasn't succeeded yet, but that's quite different from failure.
- josh2600 4y agoI’ve spent a lot of time at the edge of decentralized and useful tech. You’re right that most decentralized systems don’t accomplish the goals they set out to achieve. A few key things: decentralized means a network of at least 3 nodes where no operator controls more than 1. Anything beyond that is vanity or more censorship resistance depending on your vantage point. What we actually want is governance with checks and balances. I would argue that the problem with the US has been the centralization of power in the executive and judicial branches, particularly the executive. A functioning parliamentary system with three branches of governance is quite effective when there is a balance of power. Now to how emulate that in software is another quandary. I think ethereum for all of its faults is doing a decent job at decentralized governance. I think we’re still in the very early days of decentralized finance. A lot of lessons will be learned (and relearned from the past) as we move forward.
- Dracophoenix 4y ago> I would argue that the problem with the US has been the centralization of power in the executive and judicial branches, particularly the executive. A functioning parliamentary system with three branches of governance is quite effective when there is a balance of power. On the contrary, I'd argue that the power has towards the executive and legislative branches. The judicial branch exists to curb the excesses of the other two through the invocation of law that was already written. The other two either write whatever the hell they want and vote on it or invoke their powers by fiat (i.e. executive orders).
- sydney6 4y agoWhy are so many people calling Elon Musk by his first name, like he's their buddy or neighbor? This reminds me when Steve Jobs had passed away, many people were walking the streets with candles and cards in memoriam of "Steve". How come?
- adam_arthur 4y agoProbably because it's a unique enough name to not introduce ambiguity. Nobody can just say Steve or Bill without confusing people
- PebblesRox 4y agoPlus Musk is kind of a weird word.
- pjc50 4y agoCelebrities induce parasocial relationships. None of those come close to the response to the death of Diana.
- sydney6 4y agoThank's for your response.. what an immensely interesting [1] topic. [1] https://en.wikipedia.org/wiki/Parasocial_interaction https://en.wikipedia.org/wiki/Parasocial_interaction
- narag 4y agoThat's a measure of fame. When nobody knows you, they use the complete name, Elon Musk. When you're famous, they drop the given name, you're Musk. When you're ultrafamous, they only use the given name, now you're Elon.
- mmplxx 4y agoAnd when you're God, they cannot utter your name.
- PetitSasquatch 4y agoIt seems to me like there is a possible, albeit compromised, middle ground between highly centralised and the (in reality) partially centralised Mastodon model where users flock to popular monoliths: strictly per-user instances. Given the right tooling etc etc, expediting user's migration to their own instances, of something like a Mastodon node, could be an achieveable route to greater decentralisation without getting muddled down in things like blockchain etc.. And executed correctly the model could be more communicable than what it presently appears to be. From what I can see though, Mastodon isn't currently the technology stack to achieve, perhaps it could get there. No solution is going to be perfect, but surely an emphasis on data sovereignty would reduce the volume of arguments over what a given platform should be.
- unity1001 4y ago
- rvz 4y agoThis. They're right you know. > I want a world where people have actual ownership over their computing, but federated system designs like Mastodon will not get us that and we have thirty years of failure to prove it. The press pretending otherwise because they hate Elon and see tech companies as a threat are just embarrassing themselves. This is the media capitalising at the maximum over the emotions of the tech crowd to move to a platform that is worse than Twitter, especially in UX and is essentially a solution that is creating more problems than it is trying to solve. It really is not early days anymore with this, and the hype around Mastodon is manufactured by outrage due to 'Elon Musk' being the new villain of the month. If you have to explain what Mastodon is, which server you have to choose and how to find which people you followed on Twitter are on Mastodon due to its poor discoverability, then it tells us that it is another pointless hurdle for users to bear with, resulting in a worse experience. Mastodon clearly is a solution that is creating more problems for users and the media is clearly parroting over the hype and emotion to pretend that it is an alternative or a threat to Twitter.
- slater 4y agoHow is my favorite Mastodon-hater doing, you must've had a bad last few days? ;) (i kid, i kid...)
- rvz 4y agoI'm actually laughing at you and the entire Twitter chaos on this 'migration' made up of the same techies overhyping with the media taking advantage of the current outrage pretending that Mastodon is a viable 'alternative' when it remotely isn't. Given that it already takes much explaining how to use it, 'choosing and instance', struggling to solve it's poor discoverability since a minute amount of Twitter users out of 200M+ have tried to move themselves, it means that you are already losing. It is not 'early days' anymore. After 6 years of 'exoduses', Mastodon doesn't cut it as an 'alternative' and especially with the exact unchallenged reasons I've stated. It is not about getting manipulated by the current hype by the media, it is always about the retention afterwards that counts. Twitter is here to stay despite the 'current outrage' that the media is riding on. Sorry to burst your (techie) bubble as hundreds of millions of non-technical people are still using Twitter and won't delete their accounts over something worse.
- PhasmaFelis 4y ago> Elon may not succeed, but I hope he does and what he's doing is a necessary prerequisite. Everything he's done so far reads like he's deliberately trying to destroy the company. The only reason I don't really think that's his intention is that I don't see how he could hope to benefit from it. But it doesn't make sense.
- gonehome 4y agoThe Twitter status quo was one of stagnation (took them ten years to ship edit to a handful of people) - they're a good example of how you can basically do everything wrong, but if you have product market fit it's hard to fail. He's trying to pivot them into a focused company. I think his actions make sense in that context.
- PhasmaFelis 4y agoThere's no context in which many of his actions make sense. He ranked people based on lines of code written, immediately fired thousands of them based on that objectively stupid metric, than immediately tried to rehire a bunch of them because the company couldn't function without them. "Random flailing" is the most charitable way to characterize that.
- toofy 4y agoi have a suspicion that much of what he’s doing is just chest thumping for his new fan base, “look at how i do whatever i want. i behave how i want. they can’t stop me, i’m The Strongman!” just red meat for that particular crowd who worship strongmen. tho if he’s even a fraction of the irrationality that he’s portraying, i completely agree, twitters in trouble.
- A4ET8a8uTh0 4y agoOne could argue that the centralization process is already in place. Only servers that are deemed 'acceptable' are recommended ( I think the criteria for acceptability no hate speech, no transphobia, no homophobia ). It is interesting to the point that Guardian in its article trying to describe mastodon suggested it is left leaning. I chuckled. I mean, I suppose it is, but the whole point of an individual instance is just that.. I agree with your point. I think people are generally missing forest for the trees.
- throwaway0x7E6 4y ago>but federated system designs like Mastodon will not get us that and we have thirty years of failure to prove it sure, if you ignore the very only communication method that everyone has - the email. a myriad others came and went, and none that exist today will be relevant in 50 years, but email will
- gonehome 4y agoReplied here: https://news.ycombinator.com/item?id=33496458 https://news.ycombinator.com/item?id=33496458 Email is an example of a federated system failure.
- throwaway0x7E6 4y agoa failure that is infinitely more successful than any centralized solution. and it will outlive gmail
- sangnoir 4y ago> I want a world where people have actual ownership over their computing, but federated system designs like Mastodon will not get us that and we have thirty years of failure to prove it. Counterpoint: email has been working great for the past 30 years. The reason why email works over centralized schemes is because it has tiered decentralized controls - sending server, relays (when they were a thing), receiving servers and the client each have an opportunity to accept, reject or drop a message. Mastodon has a great chance of succeeding. All it needs is an easy-to-use client,decent discovery and giving users control of what they see. It does not have to be a constellation of mini-Twitters.
- gonehome 4y agoReplied here: https://news.ycombinator.com/item?id=33496458 https://news.ycombinator.com/item?id=33496458 Email is an example of a federated system failure.
- sangnoir 4y agoThat's very strange definition of failure: many non-homogenous systems have Pareto distributions. That in itself is not indicative of failure. 30 years ago, AOL may have been most popular email provider, it was displaced by Hotmail, then Yahoo, now GMail. Having different leaders over time is evidence of the resounding success of email federation, those domains still exist and can still send/receive emails universally. If email had hypothetically been non-federated and proprietary to AOL, it would be dead today (or as close to it as gopher)
- serverholic 4y agoCrypto is trying to solve this in a different way. Namely projects like lens protocol. The idea is that the blockchain is a shared, decentralized database and you store simple messages on the blockchain and larger files on something like ipfs. This forms a social graph that a frontend ui can then import and provide a view into (a lens if you will). It's nice because users just need one account (a wallet) and they have access to the whole social graph if they want, or they can configure their apps to only show certain things. The downside is that it's all public and once you lose your keys you're locked out of your account. Also, you can't really prevent anyone from posting things but you can choose which types of posts you allow on your feed. It's a different set of tradeoffs that's interesting to explore.
- Angostura 4y ago> Incentives mean most users end up centralized on a handful of servers with a worse experience. I don’t understand this point. I’m on a tiny server, but read the federated feed most of the time. What’s the problem?
- gonehome 4y agoMost users won't run their own server, they'll be on large centralized ones with all of the same problems that entails.
- INTPenis 4y ago>we have thirty years of failure to prove it There is a comparison to be made between the IP protocol and ActivityPub. When we all first went online we hopefully had more than one ISP to pick from, because they could federate data between each other using the IP protocol. I believe that the same type of evolution is going to happen in social media. Whether it will be with ActivityPub, ATP or whatever else I don't know but it must happen. Because having just one ISP, one social media platform, one online shopping platform is not a healthy system.
- dorfsmay 4y agoI would LOVE if there was a way to federate FB, instagram, twitter, etc... You post and read on your preferred platform but can see posts from your friends from other platforms, or even their own implementation on their own server!
- neetdeth 4y agoYou're describing a federation of platforms, and this can work, at least technically. XMPP worked wonderfully for a brief moment. Unfortunately, profit incentives are aligned against it. The Fediverse, at least as I see it advocated here, is not a federation of platforms but a federation of social cliques, where your ability to communicate with someone else is determined by connectivity in the clique graph. This is a great thing for the kinds of people who would have become deeply invested in a web forum or newsgroup back in the day, but it cannot (and arguably should not attempt to) address the mass market.
- zaik 4y agoXMPP still works wonderfully if you want to setup a chat server for your friends and family and be independent from all the walled garden messaging apps.
- vidarh 4y agoThe obvious counter to the notion that the issues with Mastodon can't be fixed is that nothing would prevent a carbon-copy of Twitter (or Twitter itself, for that matter) to federate. It's not about Mastodon, but about ActivityPub. And sure, most users will end up centralised on a handful of servers, and that is not a problem as long as they can move when a specific server does not meet their needs. The biggest importance is a safety valve. I don't know which parts of the UI you think can't be fixed, but I don't agree. Better search is needed, but can be externalised (nothing stops someone from federating with the main instances and signing up to a few relays and offering more comprehensive search). A better experience when looking up external users timelines is largely down to Mastodon's developers having very specific ideas about how to do things. The logic for following or boosting/liking toots when you're on a different instance's timeline is awkward but can be improved with an optional intermediary (falling back to the awkward request for your address for those who insist on opting out of an intermediary providing a mapping). The upside with Mastodon is that it's not a single thing. It's the server, and the web UI, and you can replace either or both. E.g. you can run Pleroma with the Mastodon web UI, or it's own, and there are additional web UI's. Or you can (and people do, and I'm doing so to) build your own, so we'll see a number of choices of platforms that will look very differently. Personally I hope Elon fails, because it will make things easier, but I think the long term belongs to ActivityPub over Twitter in any case for the simple reason that the presence of ActivityPub means that the threshold to stand up a viable social network has dropped massively. On one hand you have the software you can set up in less than a day (been there, done that), but far more importantly, sign up to a relay and within hours you have an instance that looks lively, even though most toots are via federation. Where there before was a massive chicken and egg scenario, there's now an increasingly vibrant starting point for people to experiment with their own visions for what social media should be, and each one of them will add to the network effects of ActivityPub. Mastodon may well fail, but social networks has a history of reaching tipping points that very abruptly changes things dramatically. In both directions. But the federated networks are far less vulnerable to people moving elsewhere. As for the failures of federation, e-mail and mailing lists is the obvious counter. As is the web. Social media is a newfangled baby in comparison to the successful federated services.
- djur 4y agoI don't know if he's going to succeed or fail (or even what "success" and "failure" are in this case), but I think it'd be better for the world if he failed. Twitter isn't necessary and can be replaced, but it would be good for Musk to be mired in legal problems for a while and hopefully lose a lot of money. Reducing the power and influence of both Musk and Twitter seems like a clear win-win in the long run. I feel bad for all of the people at Twitter being trampled over in the process, though.
- rchaud 4y agoThis is all-or-nothing thinking that's endemic to Silicon Valley's subsidies-as-a-service business model. Mastodon does not have to monopolize microblogging the way Twitter has, because it isn't competing with Twitter's cursed ad-based revenue model. Similarly, the tiny niches that do migrate are likely more valuable to have than the hordes of Twitter users who never post anything, and just juice the algorithm by liking and retweeting.