We turn again now to the Laboratory Of Bad Ideas, formerly known as Twitter, where the billionaire notionally only in charge of decisions about “product and engineering”, ie all the relevant bits, said earlier this week that he was considering charging everyone, not just those vain enough to want to demonstrate it, for access to (post on?) the platform:
“We’re moving to having a small monthly payment for use of the system,” Musk said.
Saying that bots cost “a fraction of a penny” to set up, Musk added that raising the cost of an account to “a few dollars or something” could put off operators of the software. He added: “Plus, every time a bot creator wanted to make another bot, they would need another new payment method.”
Ah yes, those damnable bots. They’re so troublesome! They keep on.. well actually what do they do? OK: they flood some people’s replies, and flood hashtags with spam. In my case they Like tweets—that’s literally the extent of the annoyance. (Not having a decent app that allows me to discount Likes as a notifiable event, among other omissions, is far more annoying.) But I’ve been on the site two years longer than Musk (ahem) and can’t say that bots have ever been enough of a problem as to make me want to leave it. Other things, such as not having a third-party app, have. But not the bots.
Anyway, hearing Musk make this apparently off-the-cuff announcement, I tried to remember how many announcements he’s made about what he’ll do with Twitter in the past year or so, and whether they’ve come to pass:
• buy it ✅
• not buy it, because management lied about extent of bots ❌
• let people banned under previous admin back on ✅🙈
• allow offensive speech to remain on the network ✅✅🙈
• remove verification ticks from people verified under previous admin ✅
• let anyone who pays get a “verification” tick without having done the previous step, so that one could confuse “verified real” accounts and “verified paid” accounts ✅
• add an “Edit” button ♒︎ (only for “verified paid”)
• only allow “verified paid” accounts to vote on polls ❌
• stop being CEO (post-headhunt) if people voted him out ✅
• make Twitter’s algorithm public ✅ (Link)
• wipe out the “bot armies” ❌❌
• get rid of the “block” function and replace it with “something better” ❌(? so far?)
• rename the world-known brand “Twitter” to the unknown “X” ✅
• fund legal bills for people “treated unfairly” by employers for things they’ve written or done on the platform 🤷♂️ (plenty of responses, zero evidence of any bill being paid)
• introduce video and voice call features 🤷♂️ (only said this at the end of August 2023)
• let users send each other money via the platform and process payments 🤷♂️ (announced November 2022)
• offer a revenue share with “verified paid” users based on views their tweets get an advertisements that run with them ✅❌ (the requirements are stringent, and rapidly drew complaints about tardy payments)
• turn Twitter/X into an “everything app” like WeChat in China 🤷♂️🤷♂️ (there’s not the faintest chance of getting the integration that WeChat has)
The surprising conclusion one draws from this list is that Musk doesn’t just throw out daft ideas; he then follows through on them, even when they’re daft and people tell him they’re daft (or, in the case of “verified paid”, he knows they’re daft), especially where they involve money. On that basis, you should operate on the assumption that yes, indeed, he is going to levy a charge for using Twitter. Whether that will be a charge for posting on Twitter, or for reading it, or both, is hard to figure out; if he puts the entire visibility of the site behind a paywall, it will effectively vanish from sight for many.
Let’s briefly rehearse the reasons why social networks don’t normally put up paywalls. Wealth is very definitely not evenly spread around the world: what seems like a pittance to pay for a daily coffee in the US could look like a week’s wages in India. Thus you see differential pricing for services such as Spotify across different countries where per-capita income varies widely. You’re also competing against other social networks and websites which don’t put up any price barrier, so you need to give people a very good reason to spend their money with you. An advertising-supported offering, on the other hand, plays on the fact that in the attention economy, everyone—from the richest to the poorest—is allotted exactly 24 hours each day to attend to whatever they want, and if your product can capture more of that attention, you can charge someone (either the person, or an advertiser) more in response. Hence Facebook, Instagram, Threads, Snapchat, TikTok, and all the various Twitter diaspora being free.
On the normal calculation, which is that putting up a paywall converts between 5% and 10% of your existing users to paying ones, you’d expect that Musk raising the drawbridge around Twitter (both posting and reading, I assume; given that most people just read rather than posting, allowing people to read doesn’t collect much revenue) will cut the number of users pretty dramatically. How many users are there now, though? From The Guardian story linked above:
Musk also said X had 550 million monthly users generating up to 200m posts a day. Previously, the platform had measured its user base by a different method, monetisable daily active users (MDAUs), which stood at 238 million before Musk bought the business in October 2022.
So we might expect that user numbers would head towards 24 million users—I think Musk’s 550 million number is the total of humans plus automated accounts (which includes those which just push information from other computers; and also includes “bots”, ie the bad automated accounts used for swarming and spamming). The automated group isn’t “monetisable”, because automated accounts don’t read ads. Musk’s figure does though provide an interesting perspective on the balance between humans and automated accounts. If we assume that MDAUs have remained the same, rather than seeing humans flee to other offerings such as Threads, Bluesky and Mastodon (highly unlikely, by the way), then there are roughly twice as many automated accounts as humans.
One can’t draw any conclusions from the “up to 200m posts a day” stat against the 550 million monthly users; it’s an apples/pears comparison, plus the carefully phrased “up to” which tells you nothing about variability. But it’s known that spambots are more active than most humans. A team from the University of Washington and Xi’an Jiaotong University in China used a machine learning model fed with tweets to estimate that spambots were about 8.5% of the MDAUs in December 2022. Higher, as they say, than Twitter’s 5% estimate, but a long way off Musk’s 20% claim at the time.
One can note in passing that bots aren’t the problem. What people want to be rid of on Twitter is hate speech, harassment, antisemitism, and misinformation, just for a start. The problem is that the owner now indulges in all of those. Why would you pay to not have your problems fixed, and have them made worse?
Only two questions remain. The first: at what level should we expect Musk to set the price for admission? The second: how quickly should we then expect people to drain away from the site?
On the first, the current price for Twitter Blue—sorry, X Premium (page 404’d for me) is $8 per month, for which you get a blue tick, editing, long posts, priority in the algorithm, and so on. (It’s also variably priced by country.) What’s a reasonable amount to expect to pay to post? $1, $2 per month? But that’s barely going to cover the transaction fees; it would effectively lose money because of the administration required. I’d suspect something like $20 per year, as a lump sum. Even so, the administration will be a nightmare, and a few things will happen.
• First, the payment system will be the target for scammers looking to use stolen credit card details to buy accounts
•Second, actual paid-for accounts will become the target of intensified hacking attempts (as already happens, and as used to happen a lot with “verified real” accounts)
• Third, X/Twitter’s payments system will become the target of some very rigorous pentesting by hackers eager to get their hands on a few million working authorised credit card numbers. And Twitter hasn’t been immune to hacks, both from inside and outside, in the past. While it didn’t take users’ financial details, it was relatively safe. Once it does, bad things are sure to happen.
What about the question of how quickly, and how many, people will shift away? On this, I think it’s Mark Zuckerberg’s game to lose. If Threads can improve quickly enough, and add the features that people like on Twitter in time, then it becomes the obvious, free, port in a storm. But that 5%-10% retention figure is pretty solid, observed across multiple instances of paywall introduction, and the only thing that could make some people decide to pay up is that they don’t want to have to recreate their social graph somewhere else; there’s a sunk cost in using Twitter, so why not sink some more there too. We know that there was very little takeup for Twitter Blue (as X Premium was known previously): only 0.2% of users took up that generous offer.
Meanwhile I’ve already seen plenty of power users abandon Twitter; it’s absolutely not the service that it was even a year ago. Once it loses users at the margins (which it will) and from the power user group (who post more than others), the utility one is getting from that putative $20 per year looks worse and worse value. Which leads to non-renewal, and so to fewer users as they defect elsewhere, and so on. Eventually, Musk owns MySpace 2.
But at least he’ll have got rid of the bots. Well, apart from the millions of accounts that hackers will have taken over, whose owners will either be furious but unable to contact Twitter support because, um, their account has been taken over, or indifferent because they’ve stopped using it. Musk and the bots. Together forever, ad infinitum.
Glimpses of the AI tsunami
(Of the what? Read here. And then the update.)
• Announcing Microsoft Copilot, your everyday AI companion. It’s going to be integrated into Windows, search, everything Microsoft. The AI tide is rising.
• Game of Thrones author (and other authors) sue OpenAI. Deep sigh. The sigh is because…
• Why George R.R. Martins’s lawsuit against generative AI will cost authors even if they win. Sure, it’s a blogpost by a company that makes open source software for deepfakes, but correct in pointing out that if GRRM et al win, we all lose.
• Amazon tells self-publishers to reveal if AI wrote their book. Sure, that’ll solve it.
• New Alexa will be more chatty, powered by an LLM. An obvious development, really, but do we really want chatty rooms?
• 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.