With a decent client (I've bounced around a couple of them so far), even in a beta state, the experience of writing and interacting with people on "Mastodon" is better than it was on Twitter. I'm bummed out when I have to talk on Twitter now.
With a decent client (I've bounced around a couple of them so far), even in a beta state, the experience of writing and interacting with people on "Mastodon" is better than it was on Twitter. I'm bummed out when I have to talk on Twitter now.
Wasn't expecting Twitter to start switching off systems like a ship in Star Trek trying to conserve energy for life support..
Looking forward to Tapbots new client. Don't really care what's powering the backend as it turns out, it just needs to stay working.
Nice work Mastodon and the rest of the federation, I'll be looking into server options over the weekend. Gotta have the vanity domain.
> As with LOCAL_DOMAIN, WEB_DOMAIN cannot be safely changed once set, as this will confuse remote servers that knew of your previous settings and may break communication with them or make it unreliable.
https://docs.joinmastodon.org/admin/config/#local_domain
They also recommend a "301 Moved Permanently" implementation.
You can always host a lightweight ActivityPub server like honk (https://humungus.tedunangst.com/r/honk).
For those new to Mastodon, I wrote a huge 8K word post about it recently, from my POV as someone who worked at Twitter for 9 years: https://paulstamatiou.com/mastodon/
I go over some of the constraints that the federated model brings that might be particularly interesting.
A humble review: https://notes.ghed.in/posts/2023/ivory-mastodon-app-review/
Anyway, Mastodon has no reason to stand against third party apps. Its official app was released only last year, and by no means to threat others nor become the only app in town. Experimentation is good, and Mastodon's third apps are shining right now. There are dozens of them, each one bringing fresh ideas and new concepts.
edit: typos.
I don't think that's right. You don't need to run your own server in order to add verified links to your profile - but you do need to have pages you can link to on trusted domains which can rel=me back to your Mastodon page.
There's "this random person on a random instance is who they say they are" (after clicking on their name and checking the verified part).
There is also "this person, by value of the name of the instance is who they say they are."
@mfowler@toot.thoughtworks.com - I don't even need to go click through to their name to see if they are an employee of thought works (and thus very likely Martin Fowler).
(One of the challenges with this is also finding the company sites - I'd love an old school yahoo directory of them)
The Mastodon that stands a very good chance of killing Twitter is Mastodon-the-software; "ActivityPub", if we have to call it that. This Mastodon isn't a coherent social network at all; it's a successor to RSS. People run their own servers; "scaling" them means the same thing as scaling a blog would. Of course, most people don't want to run their own servers, but that isn't going to matter by the end of the year, when 10 different providers will boot up a "Mastodon instance" for you with a single button push. Nobody is going to be thinking about instances at all; they'll just have an address, the same way they do for email.
In this world, "Mastodon" is sort of a combination of most of the good features of Twitter, Blogger, and Google Reader, all at once. Users are as "discoverable" as blogs were --- meaning: very discoverable.
The problem with social media right now is the lack of a non-adversarial, sustainable business model. All these changes stem from the fact that advertising-based business models are on their last legs and are fundamentally flawed because they are adversarial to the users - the Twitter API shutdown is at least partly because they want to drive everyone to use the official client where it's easier to impose user-hostile functionality.
Decentralisation merely side-steps this problem which works on a very small scale but not only will break down at a larger scale (operating a social media platform costs money) but also brings a lot of its own issues. Part of the appeal of a social media platform is its popularity, network effects and a sense of community where most people are happy with or at least tolerate the rules and moderation policy.
A Mastodon-powered future will have 2 outcomes:
1) every instance federates with everyone and the entire thing becomes flooded with spam and other unsavoury (or outright illegal, at least in some jurisdictions) content because there is no common moderation policy. Users eventually get fed up and leave to a centralised competitor.
2) instances federate on a case-by-case basis which fundamentally breaks network effects and makes global conversation and community building impossible. Good luck explaining to non-technical users why they can't see/interact with the same posts as their friends because they happen to be on different instances that don't federate with each other, or because the content they both want to see is on a separate instance that doesn't federate with theirs. Users get fed up & leave or can't get started to begin with and sign up on a centralised competitor instead.
In both cases I haven't even addressed the issue of funding the network itself - there is still no business model (and any business model where users pay would require the service to have enough value for them to begin doing so - chicken & egg problem when the value of a social network is in its network effects), and even if there was, it will be more expensive because decentralisation requires a lot more system resources.
That's how it works now, but I don't think it's going to stay that way for long. I don't think large shared Mastodon instances, at least larger than you'd see for a group blog, make much sense. It doesn't make sense for all the reasons you're spelling out: the amount of effort to keep those big instances going is expensive, and there's no real funding model for it, and beyond that there's not enough win to justify the work.
But a "network" of hundreds of thousands or millions of tiny instances, maybe paradoxically, doesn't have this same spam/trolling problem, any more than RSS did when we were all using Google Reader. You see posts from people you follow. Because most people aren't on big shared instances, you're not pulling lots of random traffic in to begin with.
Again: think of this like Blogger and Google Reader mashed together, with Twitter's surface-level user experience. I think it'll work fine, and I think it does most of what Twitter did (and some additional things). People forget that the "blogosphere" worked pretty great before Twitter and Facebook wrecked it. I think there's a good chance we're heading back towards something that works like blogs did.
RSS/Google Reader was mostly a "pull" medium. Creators create content while consumers passively consume it but can't write back. This works but is fundamentally different from Twitter and the modern concept of social media. Furthermore, RSS offloads all the work to the clients, so RSS servers don't have to store nor process any state regardless of how many consumers there are (you can serve the same cached response to all of them).
If you want it to be like social media, you need people to also be able to "push" things (that others who don't directly follow you see) so people can reply and comment and do social-media-y things. This then not only opens the floodgates for spam, but also makes everything harder if every instance has to federate with tens of thousands of small ones instead of a few hundred like we currently have (the current system works fine with a "hosts"-like file manually curated by instance admins, but if instances start proliferating you'd need a DHT of sorts so they can all discover each other, and again need some way of vetting the quality of them so you don't get spam into it). This turns instances into hosts not just for the original content of the instance admin but also for the replies of people, again raising both server resource issues as well as issues relating to spam, illegal content and moderation in general.
> think of this like Blogger and Google Reader mashed together, with Twitter's surface-level user experience
Blogger could potentially pull it off because it's a closed system that has one consistent moderation policy. It doesn't have to broker or host unvetted third-party content. But blogging/publishing is fundamentally still a different medium to something like Twitter.
> People forget that the "blogosphere" worked pretty great before Twitter and Facebook wrecked it.
I think that the blogosphere and social media are fundamentally different mediums. Blogging is write-once, read-many, with minimal, mostly consumer-to-publisher interactions (comments on the blog itself). For a lot of people, social media also involves direct consumer-to-consumer communication, merely adjacent to the publisher's content but completely outside of their control. Your proposal does not address this use-case.
RSS doesn’t have this problem because it doesn’t require the client to have a static hostname, thus mobile RSS clients are totally possible.
People are sleeping on how well the blogosphere worked, and how much Twitter and Facebook were able to divert it purely through a catchy UX. The blogosphere now has that UX: it's called Mastodon. Meanwhile, Twitter and Facebook are collapsing, because the centralized social network is inherently unstable and susceptible to gaming.
RETVRN! As the kids say.
I was on twitter when it was 140 characters, and appreciated the increase to 280. I've always agreed it shouldn't get too long, or you lose the character of the platform. But even with 280, I'm always wordsmithing my posts just to squeeze in enough context to avoid conflict and misinterpretation.
On Mastodon, I haven't once had to reword what I'm trying to say, and I have never felt that others' posts are verbose. The Mastodon post length limit seems to hit the sweet spot for this kind of platform.
For what it is worth, in my own iOS usage, I used Mammoth for a few months (and still follow its developer) and then eventually settled on Toot! Toot!'s developer was on hiatus just before Mastodon 4 (and the Twitter meltdown) so it got mentioned as "no longer maintained" or ignored/overlooked entirely by a lot of suggestions lists for people coming fresh to Mastodon. But it was rock solid and feature complete with Mastodon 3.x and the developer had earned a hiatus for a job well done. The developer also came back and worked to quickly catch up on Mastodon 4 changes (none of which had been truly a show stopper, mind you), though I think a slight bit late for so many newcomers from Twitter seem to have overlooked Toot!. Anyway, Toot! is great if you need a recommendation from a random HN user who has been Mastodon since a previous era. (I started getting serious about Mastodon round about 2016, myself.)
I'm in the class of iOS users without any personal macOS devices, so that's not a consideration I was aware of, but that does make sense.
Good news: It looks like that's since been resolved shortly after Gruber's post. https://mastodon.social/@icecubesapp@mastodon.cloud/10971756...
Now obviously a lot of these features are also toxic engagement-maximizers so you don't want to necessarily force them onto users the way Twitter does, but they're also positives in their own ways since they provide content discoverability and legibility.
Twitter 15 years ago was a home of weirdos and journalists mostly. There's a reason it needed changes to take off.
Since Mastodon is decentralized what the leadership wants doesn't always matter. There are several instances which have patched their code to add full search in.
This is one of the big advantages I see to the fediverse as a whole- different instances can experiment around with changes and even entirely different software stacks, and if someone doesn't like the way the mainline software is being run they can fork it (and there are several successful forks already).
Journalists ignored Twitter until the mainstream was on it. It was definitely full of a community of weirdos. I was there.
I was an early adopter of Twitter and the "no algorithms, no engagement metrics" Twitter was something I missed and was part of why I left Twitter around about 2016. It is something that I like that Mastodon provides a somewhat clean slate on.
Twitter was about ephemeral day to day life. I still remember Twitter was never more useful to me than those early days during conferences/conventions when you'd turn on SMS notifications (!) of a friend or two to help coordinate meetups and meals and use the rest of the feed flying by for a general zeitgeist of exciting things around the next corner to maybe give you a direction to head. If you missed a Tweet when it flew by it was probably too late to see the thing it was talking about and content "discoverability" didn't matter, no one cared.
I know that's not what Twitter has been in a while. I feel it fair to say that when a lot of the algorithm stuff came into play, especially with its toxic engagement-maximizing, but also with its toxic drive-by miscontextualizations, that stopped feeling like a useful Twitter to me.
I'm not going to stop instances from exploring some of that stuff on Mastodon, but I also am going to use rights to silence and block them and otherwise defederate with them at my discretion for how useful I think they are to the overall community. I think the community is just fine without the algorithms and the everything ever tooted is always searchable and scrapable for ever and ever and always relevant for miscontextualization later. I preferred Twitter without those things, those things were stuff that started me rethinking my relationship with Twitter and eventually lost me to Mastodon because I wanted to escape that.
This is what I like about Mastodon now. People(including me) posting about their mundane life.
There's no "growth hacker" looking to game the algorithm, there's no listicle about JS and the algorithm didn't try to push for the latest "viral dumb takes" to you.