Social media
Some quick thoughts on Threads
This is very obviously v0.9
Some fairly fundamental things are missing. Support for ActivityPub, which would allow you to follow Threads users from Mastodon, isn’t there yet. There’s no web interface, which means the only way to post from the desktop is to use the Android app running on the Windows Android Subsystem (I presume it works with ChromeOS’ equivalent). You can’t run the iOS app on Apple Silicon Macs, either.
Most egregious for an app which is spun out of an image platform, there’s support for alt text. This should have been there from the start, no excuses.
It feels like it’s really 2-3 weeks away from launch, but Meta obviously saw a window of opportunity thanks to the last few days chaos on Twitter and decided to push the big red button.
I’m not seeing much to worry Mastodon users
Having seen it and read what Eugen has to say I don’t think the sky is falling in for Mastodon. It will be cool to be able to follow people who exist in a Threads world from my Mastodon account. I’ll have the advantage too of a straightforward chronological feed. Which leads me to…
The feed was a mess at the start, but it got better fast
The feed is clearly algorithmic (and I would say that it really needs a chronological option). There were a lot of complaints about the number of posts from people you didn’t follow in the feed from very early adopters.
Break it down, though, and that’s entirely to be expected. You might follow lots of people but early on they may not even have actually been registered – it effectively flagged anyone you followed on Instagram so that when they registered for Threads, you followed them. And of the people who created accounts, many didn’t post much to start.
So Meta did what Meta does: it gave you something fresh to look at, every time you refreshed your feed or opened the app. Hence, lots of posts from people you don’t follow that the algorithm thought you might be interested in, based on almost no data.
At first my feed was horrible. It’s now fine, and unless I do some doom refreshing all I see is people I follow. If you judged it from the first ten seconds you saw your feed, you really should calm down a little.
Twitter is dead, dead, dead
Within 24 hours on Threads you can find most politicians, media brands, celebrities and sports teams. A lot of what made Twitter good for the kind of ordinary person was being able to connect to and talk about that things – and Twitter was the only platform you could do that, other than Instagram. Now you can connect with all the same people, with text posts.
Less than 24 hours in, Threads has 30m users – it could eclipse Twitter’s scale in a month or too. There is just no reason for Twitter to exist now. It’s just Gab with a couple of hundred million dormant accounts.
There are some serious tests of moderation capabilities ahead
Which leads us nicely on to moderation, and there are many tests of the skills of Meta’s moderation teams and policies to come. Already, some of the most horrible abusive accounts like LibsOfTikTok are on the platform. Is Meta going to ban them when they start spewing hatred and encouraging violence? Zuckerberg has spoken about Threads being a “kind” platform. If he really means that, they are going to have to come down on accounts like LibsOfTikTok hard and fast. We shall see.
The future for Bluesky is not that bright
Unfortunately we can probably consign the likes of T2 and Post to the “where are they now?” file of 2024. Which leaves Bluesky, which recently opened up invites and got quite a few more people in. It had also been pretty vibrant last weekend, when Twitter was melting down – I think that was the last straw for quite a few people.
I think Bluesky is dead. It has some nice technical ideas – shared blocklists, feeds you can programme yourself – but the whole idea of building this on a new, alternative protocol rather than using ActivityPub has meant it’s been trying to tap-dance and unicycle at the same time. It’s been too much, it’s slowed down its ability to scale up, and it probably means it’s missed its window. Those two technical sweeties are outweighed by the fact that most people I know are already on Threads, and I don’t have to search for them.
I quite like it
I wasn’t expecting to like Threads, but I do – or rather I like being connected to friends, old colleagues and family on a text-based social network. I like being able to follow the odd team, celeb, brand or politician.
I don’t hate the algorithm, although I would prefer to be able to default to reverse chronological. When ActivityPub support arrives, I’ll probably just follow a bunch of people on Threads from my Mastodon account, so I get to use a reverse chronological feed and use the powerful filters you have there.
There’s a lot for Meta to do to make it really work, but it’s a good solid start.
Meta and Mastodon – What's really on people's minds?
There has been a lot of chatter about the decision of some instances on Mastodon to pre-emptively block Meta's purported new ActivityPub-compatible service:
It’s a weird own goal for various Mastodon admins who are running a decentralized social network based on interoperable protocols to pledge that they won’t interoperate with services from existing social networks if built on the same open protocols.
It’s not even the hypocrisy, it’s just dumb and undermines the entire point of interoperable protocols.
The Mastodon server admins who are preemptively promising to defederate Threads are effectively guaranteeing that users on their instances will be isolated from a potential majority of users on the overall platform.
I’ll close this post with the same question I did yesterday: Is the goal of the Fediverse to be anti-corporate/anti-commercial, or to be pro-openness? I think openness is the answer. Others clearly disagree.
I cited email as an example yesterday of big companies running big instances and not extinguishing the openness that made the platform great. Podcasting is maybe an even better example. I don’t take podcasting’s future for granted, but 20 years in, it’s still thriving as an open medium, despite the presence of titanic players like Apple, Amazon, YouTube, and Spotify.
The anti-Meta #Fedipact can only achieve one thing: make sure that #ActivityPub loses to the Bluesky protocol. Is that what people here want?
I think this criticism misses the point, perhaps because some of the authors simply don't understand what Mastodon is and what its priorities are.
Federation, not an open protocol, is what defines Mastodon
Dare and Tristan's focus is on ActivityPub and its openness. But ActivityPub ≠ Mastodon, and services can adopt an open protocol without committing to use it to connect to anyone and everyone on the planet using the same protocol. Gab uses ActivityPub (and no one connects to Gab), and a quick view of the average block list for any instance should be enough to show that interoperability is limited by social norms as well as technical ones. Tristan's point is doubly nonsensical seeing as BlueSky's AT protocol isn't even out yet, so Meta couldn't use it even if it wanted to.
What defines Mastodon is not the use of a protocol. The protocol is just an enabler. Instead, Mastodon is defined by allowing communities (instances) freedom of association. It is the ability of communities to choose not to federate with anyone else which gives Mastodon its strength.
Mastodon is not a social network, which is where I think John and Dare start from. It's a set of communities which may, or may not, choose to connect to each other. Those relationships are based on shared values and trust: my instance connects to yours because I trust you to moderate effectively, not allow spam, or whatever other ground rules we can agree on. Some communities choose to apply this loosely, and some more strictly (some communities, for example, won't federate with others who don't have the same expectations around moderation for everyone they federate with).
For some marginal communities the freedom of association which Mastodon is based on is the difference between having a space that's safe for them to express themselves and not having a space at all. It's not about having the biggest possible audience – it's about being able to share in a semi-public way with other trusted communities.
Freedom of association is the foundational human right and Mastodon is the only service I know of which prioritises it over everything else. That's what makes it important – and arguments like Dare's, which claim that if you use an open protocol you lose that right, are just wildly off-base.
Federating with an instance isn’t just about letting individuals connect to individuals
If you want to understand why some communities are wary of Meta, it's worth remembering how ActivityPub works. Let's say I run an instance called EvilMole.social. If I don't defederate (block) Meta's server, any user on their server and any user on EvilMole can follow each other. So suppose that @molesworth@evilmole.social follows @billy@meta.social. What happens?
First, every single one of Molesworth's posts are now available to Meta, including all the replies they make to everyone including people who don't follow anyone on Meta's server. In other words, Meta gets access to posts from people who may very well have an issue with that.
It’s not (just) about anti-capitalism
John's question is whether "the goal of the Fediverse to be anti-corporate/anti-commercial, or to be pro-openness" is simply a false dichotomy, because resistance to Meta is based on its past behaviour, not simply anti-capitalism.
Just for the record, I definitely am an anti-capitalist, and I look forward to the day when something more equitable takes over. But you don't have to be an anti-capitalist to believe that Meta, in particular, is not worthy of even a shred of trust
Remember that federating with a Meta service means handing over post data even for posts which are not directly towards a user on their servers. For marginalised and vulnerable communities in particular, keeping data out of Meta's hands can be a priority. It's not that long since the company was tracking people seeking information about abortions, allowing advertisers to target based on sexual orientation, and, of course, making millions of dollars off the back of a transphobic "documentary". The idea that Meta can be trusted with yet more data to harvest is pretty naive at best. For some communities, pre-emptively blocking Meta is entirely appropriate given its track record.
Claiming that a commitment to openness requires giving Meta access to your posts, that you somehow must cooperate with them or allow them into your community is another of those occasions when it's obvious that someone doesn't understand that freedom of association also includes the ability to choose not to associate – and that freedom of association is, if anything, an even more important right than freedom of speech. It's the freedom not to have to listen.
(As an aside, it's interesting to me that American commentators ofter seem to prioritise freedom of speech over freedom of association. Perhaps this is because, while freedom of speech is an explicit part of the 1st Amendment, freedom of association is not – it was only with NAACP vs Alabama in 1958 that freedom of association became a recognised right on the basis of its necessity for both free speech and due process. Meanwhile, in Europe, freedom of association is a separate protected right.)
The email analogy is terrible, and we should all stop using it
John – and many others – often cite email as an example of "big companies running big instances and not extinguishing the openness that made the platform great". There's a bunch of problems with this, but the biggest is that in fact email is a perfect example of how "open" platforms can become dominated by large corporations and effectively lock others out.
Email runs on open protocols, so it's technically possible for me to run a server on my Mac mini upstairs and start sending and receiving mail. The reality, though, is that virtually none of the mail I send would get through – because Gmail, Microsoft, et al would mark it as suspicious and end up automatically blocked as spam.
Blocklists for email exist and are shared across services – and blocking is often pre-emptive, not based on suspicious behaviour of that server. Sure, email is an open set of protocols, but it's also highly restricted by large companies and not at all open to either smaller providers or individuals.
This isn't just an abstract issue: I know of friends who have had to abandon email servers they ran themselves, sometime literally on a box in the corner of a home office, because the big corporations that dominate email simply wouldn't deliver anything they send.
John – and everyone commenting on this – must also know the history of how companies use open protocols as a method of "embrace, extend and extinguish" – but since the Microsoft era they have learned that they don't need to extinguish, just extend in ways which make it incredibly hard for anyone to compete. The canonical example of this for me is how Google has used Chrome to influence the web and its protocols, in such a way as "Chrome-only sites" now being a thing. It's incredibly naive to think that openness automatically makes things better when you have companies totally willing to exploit that openness to dominate a market.
Mastodon, BlueSky, and Highlander Syndrome
I am total agreement with Jamie Zawinski here. There is no way I will trust anything that Jack Dorsey has anything to do with. He’s either incredibly naive – in which case he should be nowhere near a service which requires trust and safety to be at the front of everyone’s minds – or he knew what Musk was like and didn’t care (in which case… you guessed it).
I don’t, though, really care if people jump from Twitter to BlueSky. There is room for more than one successful microblogging platform and different people will have priorities which aren’t the same as mine. I don’t particularly want to spent a massive amount of effort on a centralised service, but you might feel differently.
There is a lot of Highlander syndrome here: “there can be only one” social network which wins, there can be only one platform which everyone must be on. Some of this comes from the narratives which tech journalists love to write about. Conflict sells, conflict drives clicks, and if you can personalise the conflict to two “heroic founder” figures duking it out, all the better.
This isn’t, by the way, some kind of failing solely in journalism. Our oldest and most fundamental narratives frame things as battles between giants. The myths of gods and heroes are full of them, and seeing founders in the same vein is just part of the same old story. We do it with kings, religious leaders, scientists, you name it. Even our efforts to celebrate the collective often devolve back into hero worship. We’re just not very good at celebrating the collective, unless we have a clear enemy to stand against.
Weeknotes, Sunday 7th June
Some notes on anger
I’ve found myself getting astonishingly angry over the course of the week. There’s a lot to be angry about, but anger never sits well on me for long. The anger is, of course, well placed. Whether you’re angry about the government’s utter incompetence over COVID-19, the structural and personal racism which oppresses black people the world over, or a famous author’s transphobia (and yes, please, let’s not call it anything else), there is much to be angry about.
I’ve come to see social media in general and Twitter in particular as forces for ill in society, not good. That puts in me a small minority: there’s still plenty of people in tech who see the effects of social media as, on balance, a social good. When you see the impact that the awful death of George Floyd had, amplified to billions of people via social media, then you can see their point. Perhaps, now, we will get real change.
But anyone who has worked in social media management will tell you that the way to maximise your reach isn’t to make well-honed arguments but to to provoke emotion, and there is no better emotion to provoke in politics than anger. Twitter is a hate machine. We mock Trump’s Twitter use, but he’s a master at it, because he understands the fundamental rule: when you have people angry, if you want to reach more people, get them more angry still. Pile anger on to anger, until the world is burning.
Social media spread the news of George Floyd’s death further and with more impact than any other medium in history could have, and that is undoubtedly a good thing. But social media doesn’t give us any way out of the anger. It doesn’t give us any “and now what”. All it can do is keep making us more angry, because it rewards making you feel, not making you think.
But anger alone isn’t enough to solve social problems and, worse still, it’s addictive. It feels good, and it overrides the moral brakes in your brain. Anger drives hate-filled cops as well as justified protests. It’s a way to feel powerful in the moment, to feel in control of things that they have no control over.
That makes it doubly dangerous. Not only does it mean you lose control and do things beyond your own moral framework, but it’s gives you the delusion that you’ve already achieved something. Protestors throwing rocks today will have no more power tomorrow than they had yesterday, but get a sense of accomplishment. They feel like they’ve already made a difference.
Getting off social media
Related to all this: Inspired by a conversation with Phil Gyford I’ve set myself the task of writing something on how to remove yourself from Facebook while preserving the benefits of Facebook.
Technically, of course, it’s easy. There’s plenty of platforms which deliver the functionality of Facebook in a more open and ethical way. The challenge is actually discoverability.
The one unalloyed good that Facebook has brought to my life is that it’s genuinely brought me closer to my family. When your parents are alive, they’re often the glue that binds together you and your relatives. They tell you what’s going on, they keep track of who is where, and who has done what to whom. Then they die, and that bond with the extend family vanishes.
Facebook lets you preserve those bonds, but also makes it easier to rediscover them. Without Facebook, I wouldn’t be in touch with my Aunt Shiela, my dad’s last remaining sibling, who lives in Cypress. I wouldn’t be in touch with so many of my cousins, who prior to everyone being on Facebook I wouldn’t have known how to contact. And none of them would have been able to find me, either.
So the real issue with replacing Facebook isn’t “how do you remake the experience” but “how do you make yourself as easily discoverable”? That’s a much harder one to crack.
Things I’ve been reading
These examples of early computing design are almost heart breaking for me. Machines like the PET had the promise of science fiction about them.
The early days of home computing – in pictures | Technology | The Guardian
Nearly Half Of The Twitter Accounts Discussing ‘Reopening America’ May Be Bots - there isn’t much doubt in my mind that social media is, on balance, bad for democracy.
I'm fascinated by Ian Schrager - from Studio 54 to basically inventing boutique hotels, via a spell in prison.
Ian Schrager Is Still Creating Buzz - The New York Times
I've noticed these meeting notes generated by AI creeping into Outlook at work. One of Microsoft and Google’s key words when talking about AI is “useful” - think of Google calling the Pixel 4 “the most useful phone”.
Meeting Insights: Contextual assistance for everyone - Microsoft Research
Unsurprisingly, this had the MAGA crowd foaming at the gills, and even drew a tweet from El Presidents himself.
Donald Trump, the Most Unmanly President - The Atlantic
“What I was hired to do was to create a 21st-century media company,” Lynch told me in his glass-enclosed office on Condé’s new executive floor, once the company’s dedicated gallery space. “Part of that is defining what that means, because they don’t really exist yet.”
Condé Nast’s Future Under Anna Wintour and Roger Lynch
Oh Vice.