What everyone's getting wrong about Threads

Plus generative AI invades newspapers, which welcome it with open arms

Prompt: “multiple threads of different colours coming together from the left to form a colourful weave which fills the right half of the picture”. We need to talk about what “half” means, Diffusion Bee.

Somewhere this week I saw a comment that the US tech media only has three stories:

1: this company is booming!

2: this company is failing!

3: look how much money this company got! (Often along with either of 1 or 2.)

Sure, you can point out that there are slight variations on these: Any of those three plus “and it’s bad/good for us” probably covers most of the waterfront.

Anyway, possibly related: earlier this week Mark Zuckerberg gave the troops at Meta an update on the progress of Threads, which got 100 million signups in its first few days—making it the fastest-growing product online ever, ahead even of ChatGPT’s introduction. Welllllll, Zuckerberg said, turns out that amazingly enough, not every single one of those signups had stayed around. In fact, fewer than half of them had.

The narrative thus was that Threads had gone from being Story No.1 to Story No.2. It being Meta, there was no obvious visibility on No.3, though the Washington Post had a story a few days ago about how Threads was built: a lean team of about 60, who started work last November and essentially played at being a startup to get the job done.

And yet it wasn’t really a surprise to anyone who has watched a new software launch. These always follow the same pattern: huge interest at the start (with a ton of signups), followed by a rapid dropoff which has to be warded off by adding new and more interesting elements.

The key criticism of Threads, at the beginning, was that it only had an algorithmic timeline. This is a criticism that doesn’t make a lot of sense. Most people absolutely need an algorithmic timeline that throws all sorts of things at them, because they don’t have a large established number of followers or friends when they join a social network. Social networks function on Zipf’s Law, the 80-20 principle: a tiny number of people post most of the content, and a gigantic number post barely anything. So if you’re trying to keep people engaged with your brand spanking new network, you’re going to give them an algorithmic timeline, and only that, for a while.

Bearing that in mind, of course the thing that all the power users clamoured for immediately was: a non-algorithmic timeline. (Or, more accurately, a timeline whose algorithm was reverse chronological.) Which did appear, early last week. Cannily, though, Meta made sure that it was well-hidden, and that the app would often revert to the (I’m going to call it the) algorithmic timeline.

Even so, we were left with Story No.2: this company (well, app) is failing! Which is an amazing sort of story to try to fit onto this situation. Threads is well-funded and, as I said a couple of weeks ago, what it’s really aiming to do (confirmed in the Washington Post story) is disrupt Twitter, since rebranded to X.

For that reason, Threads is rolling out new features, carefully, with a view to producing the best possible experience. You can see the thinking by what is being offered: the aim is to keep the power users on board, and keep them Adding Content.

From the start there was quote-posting (what we used to call quote-tweeting, but Twi...ah, X now calls them “posts” so hey ho). That’s notably missing on Mastodon, though available on the still invite-only Bluesky. There was muting, there was blocking. There wasn’t a desktop app—only the phone one (even the iPad version is, like Instagram, just a phone view rather than an iPad-specific product), and there wasn’t a web browser version. There weren’t lists—which are in effect alternate reverse chronological timelines. Those latter three are power user features, and while they’re nice to have—the desktop app and web app both increase the convenience for power users, meaning they’ll post more—they’re also a lot of work. Given that the vast majority of users are inevitably going to be on their phones, the focus has to be on the phone app.

One thing that particularly interests me is this:

This is absolutely fascinating, and what it says to me is that the Threads team, led by Adam Mosseri, is taking the problem of moderation very seriously and at the same time, putting it into the hands of the users. “Hide for everyone” really is smart. Sure, Twitt—X has it, and has had it for a long time. That’s because Twi—X has been around for ages. But what we’re looking at here is the priorities of the network, as exposed by the choices made by the design team.

I’m still finding Threads interesting: the level of discussion is far less, uh, confrontational than Tw—dammit. The algorithm throws up some unexpected people (still not enough from my favourite sports; I muted football and rugby, and it seems to take that as a lack of interest in sports. Bad algorithm). Generally, the tenor is of a more relaxed neighbourhood, even while many of the same people are posting there as on T…witter.

On the specific topic of social warming, besides the efforts that the Threads team are putting into tamping down the potential for rows, I think there’s a general effect from having so many networks all of a sudden: it diffuses the anger. An article in Slate recently pointed out that it’s become harder for far-right grifters to find people to annoy on Twitter, because folks can just drift away to those other, less confrontational networks; or, indeed, ignore them. There’s far less of the feeling that “It’s not me locked in with you, you’re locked in with me”.¹

We’re almost spoilt for choice, though I find the number of very active users who post on multiple networks really intriguing. I wish there were some way to measure what proportion of users are common across both Twitter and Threads, and how active they are relatively on each.

In any event, Threads is an ongoing experiment, and if you thought that losing half its initial signups was bad, consider that that still means it’s probably pushing north of 40 million users who are coming back regularly. For something cooked up in eight months, that’s more than respectable. The people prophesying its end are, I think, wildly premature. But that’s how hot takes tend to arrive: far too early, and often wrong.

---

Glimpses of the AI tsunami

(Of the what? Read here.)

• The Swedish newspaper Aftonbladet is getting AI to write summaries. Though it will get humans to check them.

• The Nottingham Post is getting AI to write summaries of “longer” stories which will then be added as bullet points to the top, or possibly middle, of the story. (Though in a bit of looking around the site I couldn’t find many examples. More common was the popular headline-writing technique “Half of population to develop this medical condition as study sends warning”, which doesn’t say what the medical condition is until you’ve clicked through. “Mental health disorder”, since you ask.)

• News Corp using AI to produce 3,000 Australian local news stories a week. Can you see a pattern emerging? A team of four staff generate the stories, which are about “weather, fuel prices and traffic conditions”. A News Corp spokesbod said it would be more accurate to say the articles provide “service information”.

• Meta released an open-source system which will generate audio (ambient sounds, music, whatever you want) from text prompts. Listen to the samples. We’ve come a long way.

• A rather good post on why generative AI won’t disrupt books. Phew. Speaking of books…

---

• You can buy Social Warming in paperback, hardback or ebook via One World Publications, or order it through your friendly local bookstore. Or listen to me read it on Audible.

You could also sign up for The Overspill, a daily list of links with short extracts and brief commentary on things I find interesting in tech, science, medicine, politics and any other topic that takes my fancy.

• I’m the proposed Class Representative for a lawsuit against Google in the UK on behalf of publishers. If you sold open display ads in the UK after 2014, you might be a member of the class. Read more at Googleadclaim.co.uk. (Or see the press release.)

• Back next week! Or leave a comment here, or in the Substack chat, or Substack Notes, or write it in a letter and put it in a bottle so that The Police write a song about it after it falls through a wormhole and goes back in time.

1. Love that film. Rewatch it every so often. Never sure if they’re meant to somehow have superpowers—apart obviously from Doctor Manhattan—or just happen to be superhumanly strong. If you know the answer, please share.