I started writing this late on Tuesday, 13 December, 2022 in Portland, Oregon. It was the kind of day where you could see patches of blue in the sky, blue that you hadn’t seen for days and might not have realized you missed.
I have been away. I have not been well. It has been the kind of not well that is in part mitigated by lots of stupid mental health walks. I am starting to do better now.
Now it’s Friday, 16 December, 2022 and I’m on a plane the precise location of which I am not able to disclose on Twitter.
I am sorry. I am still thinking about Mastodon, the thing that is not Twitter that might be the thing after Twitter, but if it isn’t the thing after Twitter, it’s certainly a thing after Twitter, and certainly at this point, one of the stepping stones.
First, some background: it feels like Robin Sloan beat everyone to the punch is accurately describing what it feels like now for a subset of people who’re paying attention to “internet stuff” with his post/thing A Year of New Avenues1, in which he flings open the doors and recognizes the cautious energy of experimentation ushered forth, in part, through the destruction of a cherished/hated/abusive relationship with a place of social media (that is: THIS IS NOT A PLACE OF HONOR, IT IS A PLACE OF SOCIAL MEDIA) and the corresponding protracted birth of the mewling of open source homesteading libertarians excited at the prospect of letting a thousand homesteads bloom, but also accidentally on purpose doing so with openness, allowing a whole host of people to play with said social network in ways previously impossible or at the very least gated, at the place that was cherished/hated.
(I too have been reduced to writing Python programs to mess with my personal Mastodon instance and honestly, the fact that I can just go and click on something and get an API key is… super weirdly exciting?)
Actually, this is less about Mastodon, and more about social media gestures expansively.
Here’s the big idea.
Some of the content [sic] we publish online is on the more-like-oral-speech end of the spectrum as opposed to the writing end of the spectrum.
Thinking more about this – thanks to the 2003ness of what’s happening – I’m starting to become much more open to the idea that conversational social media should by default be ephemeral.
Why? Well, because it’s speech. I caveat all of this, as usual, with the note that I have not gone and read and researched papers on social media as phatic expression and small talk.
This isn’t to say that all speech on Twitter and other short-form social networks is small talk and inconsequential. It is obviously not. But I want to make the case that much of it is, and in the same way that much conversation can be inconsequential relationship maintenance and fall into patterns of cultural convention of course important conversations can happen.
But a concept I keep coming back to is that it is easy, as a technologist who is excited about the prospects enabled by the digitization of things, is that digitization is in a very real way massively destructive. I am also not saying that there aren’t significant benefits to digitization! Clearly there are! Don’t make me list them!
I gave a talk a long, long, long time ago about how we’re embodied agents in a physical world, and a lot of how we act and work and interact is based on a whole bunch of environment that we get for free, and that it’s easy to forget in the way that you’d say a fish forgets (or doesn’t even know) that it’s in water.
Speech as spoken is by default ephemeral unless we expend various amounts of energy and time and effort to record it. That energy, time and effort has been reduced over time thanks to the fact that while we might in some respects be dumb monkeys, we’re also dumb monkeys who’ve figured out the whole concepts of tools, tool-making and tool-using, which includes such groundbreaking advances like “the written word” so we can build, hopefully, on the backs of those who came before us.
So. Speech as spoken. It… vanishes. This is because speech is vibrations in the stuff we’re surrounded in and, you know, things like acoustic attenuation. When we say something, it goes away.
It’s very, very easy to keep things for longer than the time it takes for speech as spoken to dissipate. And it’s getting easier.
If I were to make a casual, careless argument based on handwaves pop-psychology airport books that have later turned out to be at the very least problematic and at the very least-worst, outright based on false and non-replicable premises, then I’d say from a cognitive evolutionary psychology perspective, we’re a species that’s the product of ephemeral speech as spoken, and (massive caveat and handwave of not doing any research at all, and of speaking to generic white-presenting British/American culture) culture of the same.
But ever since Twitter started and for me, personally, ever since I started using it something like 15 years ago, it has been very speech-like. And the default has been to record it, effectively, in perpetuity.
In this way, I’d call conversational social media like Twitter and its contemporaries and followers, just like Mastodon, just like group text threads careless social media, where I mean something like conversation without concern.
If you think about online speech that way, speech that is most like speech-as-spoken (which I would even include say high-traffic messageboards, forums and mailing lists), then the EU’s right to be forgotten makes a lot more sense in the specific context of a speaker revoking the storage of their conversation. It is not so much then a right to be forgotten (ie in the sense that you’d very much like newspaper and media reports about you removed, thank you), but a right to have your conversation respected as ephemeral.
I’ve got a couple reasons why I’ve been thinking about this more. First is that I just blew away about twenty three-odd years worth of blog posts thanks to how shitty the internet has turned in the last decade or so. Losing that was painful to me - it’s a reflection of myself and my self, it’s a record of my growth, and it is in effect the diary I kept because of my somewhat atypical approach to writing in the open.
At the same time, I’ve been considering whether it’s time to just blow away those 15 to 16-odd years worth of Tweets. I have an archive, of course, so I have a personal record. But why the reticence to remove all this content, to not make it public-world-readable anymore? How come I’m not comfortable with that?
One reason is that the record of my writing is how I’ve built my reputation over the years. The records illustrate my thinking, and my thinking and approach is what’s led to me making friends and developing a career. It is a piece of me.
The counterargument though is that I’ve made my reputation now. And in the last 6/7 years, my reputation and writing has not been on a blog, but on this newsletter. And the, well, consequential speech that I’ve uttered on Twitter, the good stuff, I’ve memorialized that elsewhere.
No, I think I’m now comfortable with letting my speech fade out. It is nice that it could be permanent. But it doesn’t have to be. And yet because it’s been easy to store, making it permanent has been the default – until the advent of differently opinionated platforms like Mastodon, that offer the built-in ability to auto-delete your utterances. I could go on about reasons why permanency makes sense for commercial reasons: more surface area for ads, more opportunity for outrage-based engagement, etc. And yes, outside of the commercial aspect, the personal ability to go sift through your lifelog and, for some people, effectively scrub through and search your conversation can be good! Of course: it’s a tool, and you know what I think about tools, their biases and reasons for existing. But again: I have my own archive. I can scrub and search through it offline, away from the public view, or also prevent anyone else from scrubbing through and searching my conversation.
(There is a side note here in that this corpus of conversation across all these platforms is an unprecedented trove of information of value to researchers. I agree!)
One thing that came up as I talked about this in public and with friends was the reluctance to be responsible for link rot, to break what makes the web the web. And I get that. I really, really do. The web is made of links and all of this is built on the web, but the thing is, not breaking links isn’t a law. It’s an opinion. And right now it feels like an opinion born of an earlier time with less experience and different experience.
I mean, if we really didn’t want link rot then we wouldn’t have the HTTP status code 410 for gone: there was a fragment of a conversation here, but it is no longer available and will not be available again. It has drifted off, dissipated, lived its natural life.
Here, I have an example. Metafilter is an old site. It carries with it the opinions and ideals of a much earlier internet, and one of those was don’t break the link. But over twenty-odd years, people will have real, reasonable needs to delete their posts. Is that okay? Should it be? I think yes. Respecting people is, I think, more important than respecting the integrity of the archive. It would be nice if the archive were preserved, but I do think we should be able to withdraw consent. People need agency, and I’m not even getting into vulnerable peoples’ needs. It shouldn’t matter whether you’re vulnerable or not.
So. Let conversation be conversation.
Perhaps experiment with a default of ephemerality.
Just because we can keep everything for an effectively trivial cost doesn’t mean we have to, whether or not we think we should.
(Yes, I know about disappearing messages. I know about Snapchat. I would like much, much more of things like this)
There is this thing called context collapse which I encountered during the early awkward years (as if these years aren’t awkward either) of social media, and again, mainly as a result of stuff that happened on Twitter. A lot of the main characters of Twitter are a result of context collapse, and some of this is covered very well in Rebecca Jennings’ piece for Vox, Every “chronically online” conversation is the same2.
I like to build on the concept of context collapse by linking it to the destruction of context that I talked about above when we digitize… stuff. My examples used to be turning the mixtape into a CD into a playlist: you instantly lose the physicality, the art, the personalization. These may and are in many cases outweighed by utility like ease of use, but they’re also actual things in a physical universe that exist and just are.
If I build on what I wrote above, speech acts online are intended for certain spaces. I think we can’t avoid that as a function of how we’ve evolved and that our lives are, well, embodied.
A way I’d explain this is that applications and social spaces are exactly that: spaces. They are different rooms with different contexts. If you’re giving a presentation at a conference, you’re in a BIG ROOM and you can see LOTS OF PEOPLE (or even a smaller one!). If you’re in the kitchen at a house party, that’s a unique context. Digitization and social media and all that jazz has meant that in theory at least it’s trivial to transplant content from one place to another because, duh, it’s just bits and we have things called standards and interoperability and protocols and so on.
I have a slight feeling that this is not, you know, great. Or at the very least it might have some downsides! (Of course it would) And I hate writing this because I worry it’s making me sound like some sort of shitty reactionary luddite when that couldn’t be further from the truth, I just really, really don’t want the technological tools we build to fuck things up when there’s the chance they might not have.
Context collapse in this case means when we take something from one context – one room, say – and we transplant it to another, it means we’ve designed expectations where a New York Times post on Facebook looks like any other post on Facebook. Any photo on Facebook looks like a Facebook photo. There are hardly any contextual cues as to provenance, and those contextual cues as to provenance are key to providing context.
The shape of a thing matters.
A related concept to this that I’ve been thinking of is that of digital genotypes and phenotypes. The raw bits of content make up the genotype - the information, and, say, metadata with hints or specifications as to how that information expresses itself. And the expression would, of course, be the phenotype. The photo is the photo, and a photo in Instagram has its own genotype and phenotype (the specific white border, the specific typeface for the title, the iconography and so on). What makes an instagram photos is more than just the image data. But when you cross-post an Instagram photo to Facebook (which is practically intra-platform), the phenotype and context of an Instagram post is stripped and then suddenly you’re left with the genotype of the photograph, and the information takes on the phenotype presentation of its new container: Facebook, and Facebook’s feed.
It doesn’t have to be this way. The container could be different. Your Facebook feed could look very different. It could look like there were literal copy/pasted Instagram posts in there.
And we get this, right now, when Tweets are screenshotted or TikToks are cross-posted but have the bug burned-in, when Reddit posts are screenshotted and shared elsewhere. That’s useful context, I think, and often when I see that context stripped, I think the reason is that it’s, well, stolen content (such as can be or is) and is being used by someone else to ultimately make money. (We are not going to have a discussion about intellectual property here).
Your Facebook feed could look more like someone’s screenshotted or embedded an actual Vox story in it next to a TikTok next to whatever. All that would stay the same would be the parent container. Sure, you can still add platform/container-specific data around that imported content. You could have Facebook comments attached to your embedded Instagram post. (I’m not sure why you’d want to, as a user, sounds like that’d be confusing).
That’s one kind of context collapse.
An aside: the contextual cues we lose when “content” is imported into a platform include things like “this is a respectable news source” which are things like “they use a professional looking font”. These design cues and tropes were appropriated (and have been, forever) by those seeking to trick and others, from political mailers designed to look like official voter registration communication to those not actual news sites pushing misinformation and propaganda that do their best to look like local or national news media. Now, it would be interesting if one had a list of such sites publishing misinformation that are dressed up in sheeps’ clothing and applied at the browser level anything from the janky custom stylesheet to a slightly better implemented browser extension all the way through to built into the browser in the first place or, I don’t know, a fucking injection at your ISP level, something that stripped the disguise or costume of misinformation sites. Rendered them in Comic Sans (yes, I know, this probably wouldn’t work). Would it reduce trust in those sites? Maybe? I would like to experiment and see if it did! Presentation matters, right?
Anyway. Another context collapse example. Different rooms, different conversations, different contexts. One of the horrible things about how Mastodon is designed right now is how it distinguishes between three types of communication:
These are very different contexts and the problem is they look the same. They’re hardly clearly visually distinguished despite being radically different contexts, and in this case, no wonder people are confused about this whole “federated” business when they look exactly the same. One of the earliest hacks I saw on Mastodon was custom CSS to call out DMs in your personal timeline because they looked just like everything else and also, seriously, why are what I thought were private DMs turning up in my timeline?
I feel like the view that “these are messages” and “how do we design these messages” has been designed away into “just make them look all the same”, and look if you’re a designer who wants to insist that they’re in different places or different tabs, sure, fine, but seriously: they could look very different. One of the reasons different conversations and different kinds of social media exist in different apps is precisely because they are different apps and present, mostly, radically differently. (Until, weirdly, and of course, design starts being imitated everywhere like circular avatar profiles and the virus that is “stories”).
Try making it messy. Try making the rooms different, the contexts immediately, clearly, visually distinctive. I think it’s worth trying. I would love to see it and sure, I expect that people will be confused and you might lose engagement for a while, but here’s the thing: there’s something in having common architecture (rooms in the real world after all all have doors and walls and ceilings and probably lighting and so on) and affordances that are better than others (you know, the shitty doors and the not shitty doors, the shitty chairs and the not shitty chairs) but ideally, rooms for different purposes look different because their design reflects their purpose. They do not have to look so similar.
That’s it for today.
How’ve you been? Like I said, I’ve not been doing great. But I think I’m doing better. And I send you hope that you are, or that you will be, doo.
Every “chronically online” conversation is the same, Rebecca Jennings, Vox, December 7, 2022 ↩