Start Up No.1788: why Musk still might not buy Twitter, HBO’s QAnon show shadowbanned, Apple unveils repair scheme, and more

the autocorrect function on iPhones annoys a lot of people, but there’s a reason why it can be persistently wrong, as its inventor explains. CC-licensed photo by Brett Jordan on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Twitter stuff starts four links in. I’m @charlesarthur on Twitter. Observations and links welcome.

Google may now remove search results that dox you • The Verge

Mitchell Clark:


According to a support page, Google will also remove things like “non-consensual explicit or intimate personal images,” pornographic deepfakes or Photoshops featuring your likeness, or links to sites with “exploitative removal practices.”

Making a request involves giving Google a list of URLs that link to the personal information, as well as the search pages that surface those links. After you submit a request, Google will evaluate it. Its FAQ says it tries to “preserve information access if the content is determined to be of public interest,” as in the case of content that’s “newsworthy,” “professionally-relevant,” or that came from a government. If Google does decide that the links should be removed, it says they’ll either not show up for any search query or that they won’t be surfaced for searches that include your name.

Google seems to be applying a relatively high bar for what counts as personally identifying information, which makes it a bit different from the systems it’s had to implement in places like the EU to comply with so-called right to be forgotten rules. Those laws let people request that links they deem unflattering or irrelevant be taken down, which isn’t the case here — the rules Google added today only cover links to very sensitive info.

It’s also unclear whether Google will remove sites that exist explicitly to sell people’s information. If you’ve ever searched for someone’s phone number, you may have ended up at one of these services, promising to give it to you if you subscribe. We asked Google about this and will let you know if we hear back.


Reason given by Google: “the internet is always evolving”. More like the threat landscape is the bit that keeps evolving: people calling in murderous attacks on others, using America’s weaponised police force, is almost commonplace. (In Europe I guess they just order lots of pizza? Killing you just the same, but over the course of decades, through arteriosclerosis.)

As with the RTBF, the original source remains; it’s just the Google index entry that goes.
unique link to this extract

Autocorrect explained: why your iPhone adds annoying typos while fixing others • WSJ

Joanna Stern speaks to Ken Kocienda, who created the iPhone autocorrect software:


Perhaps you’ve seen your phone fill in your intended word as someone’s last name, or the name of an app on your phone.

Here’s what’s going on. When you type, the autocorrect algorithms are trying to figure out what you mean by looking at various things, including where your fingers landed on the keyboard and the other words in the sentences, while comparing your word fragment to the words in two unseen dictionaries:

• Static Dictionary: Built into iOS, this contains dictionary words and common proper nouns, such as product names or sports teams. There were over 70,000 words in this when the first iPhone launched and it’s gotten bigger since then.
• Dynamic Dictionary: Built over time as you use your phone, this consists of words that are unique to you. The system looks at your contacts, emails, messages, Safari pages—even the names of installed apps.

It’s also where new words unique to your vocabulary get logged: By the third time you type an unknown word, the software will typically add it to the dynamic dictionary and stop trying to turn it into something different, said Mr. Kocienda and others.

“The static dictionary and the dynamic dictionary would be in a little bit of a battle with each other,” Mr. Kocienda said. The software is designed to break the tie, he added, but it doesn’t always pick what you would pick.

In my case, the static dictionary is saying “Hey, she’s trying to say ‘newsgirl’!” but the dynamic dictionary, having learned from me, now says, “No, you idiot, obviously she means ‘NewsGrid’!”

An Apple spokeswoman confirmed the learning rule and explained that with “NewsGrid,” the learning may have been delayed because of its unique capitalization. I would need to type the word the exact same way twice, and I often forget to capitalize.

What you can do: The surefire way to make sure your phone knows your personal vocabulary? On your iPhone, go to Settings → General → Keyboard → Text Replacement. Now add your words or phrases to both the Phrase and Shortcut fields, which will add them to the dynamic dictionary.


unique link to this extract

Apple’s self-repair vision is here, and it’s got a catch • iFixit News

Elizabeth Chamberlain:


We were cautiously optimistic about the program in November. Anything that enables more people to do repairs is great news! And there’s a lot to be excited about in the details Apple announced today: seven years of parts availability, retail sales of tools that only official Apple techs could get before, and free step-by-step visual repair manuals available for everyone on Apple’s site. But as the doors open on this new venue, we’re underwhelmed, and settling back into our usual skepticism.

The biggest problem? Apple is doubling down on their parts pairing strategy, enabling only very limited, serial number-authorized repairs. You cannot purchase key parts without a serial number or IMEI. If you use an aftermarket part, there’s an “unable to verify” warning waiting for you. This strategy hamstrings third-party repair with feature loss and scare tactics and could dramatically limit options for recyclers and refurbishers, short-circuiting the circular economy. 

As of today, you can buy an official Apple iPhone 12 screen and install it yourself, on your own device, with no fuss. Until now, DIY repairs relied on keeping the Face ID speaker and sensor assembly intact, then very carefully moving it to your new screen, and finally ignoring some gentle warnings. If your assembly was damaged or defective, you were out of luck. The new program will solve that problem—assuming you’ve bought an official Apple part.

To check out with that part, however, you’ve got to put in your phone’s serial number or IMEI. And when you’re done installing the part, you need to pair it with the phone you indicated in your purchase, via over-the-air configurator software Apple says they will make available through their parts store. Requiring parts pairing essentially puts an expiration date on iPhones.


I looked at this back in 2016. See it from a security point of view: what if someone malicious wanted to put a part in that would monitor everything you do? Maybe you’d want Apple to be able to forestall that.
unique link to this extract

Breakingviews: Elon Musk probably won’t buy Twitter • Reuters via NewsBreak

Lauren Silva Laughlin and Gina Chon:


Sure, the Tesla boss was clearly serious about acquiring Twitter as of recently. The financing from Morgan Stanley is shored up. The agreement includes a fee of $1 billion that he – or Twitter – would have to pay if they renege on the contract. And Twitter’s lawyers even wedged in a so-called “specific performance” clause, which could theoretically force Musk to buy the company if he threatens to back out, though in practice this could probably be settled by adding to the break fee.

There are good reasons for him to get cold feet. The biggest is Tesla. The electric-vehicle maker’s stock has fallen around a fifth since Musk first revealed his stake in Twitter, partly because Musk may sell shares to fund his new adventure. If Tesla’s stock bounces back – likely if the Twitter deal falls away – the $40 billion of recouped wealth would more than make up for the break fee.

China is a major sticking point too. Tesla produces half of its vehicles there, as well as a quarter of its revenue. But Twitter is no friend to the People’s Republic, most recently for defying Beijing in its handling of content related to Hong Kong protests. China could easily hold Tesla to ransom if a Musk-owned Twitter didn’t play ball. That’s uncomfortable for a self-professed “free speech absolutist.”

…One thing makes it easier for Musk to walk away before any of this becomes a problem: The market partly anticipates it already. Twitter’s stock is currently trading 11% below his offer price – a fairly wide spread for a deal with little antitrust pushback. Musk’s tweets criticizing some company actions – potentially flouting the merger agreement – already suggest he might be starting to lose interest. Most likely, Musk’s attention will wander elsewhere. It wouldn’t be the first time.


Well, someone had to offer the contrary position. (Tesla’s stock is down about 20% this month, but 10% up on a year ago.) He would be risking a huge amount on Twitter generating enough cash to pay the interest on margin loans of Tesla stock, which could be forfeit.
unique link to this extract

Twitter censored tweets on HBO QAnon show ‘Q: Into the Storm’ • Gizmodo

Dell Cameron:


For the past year, Twitter has censored tweets about a documentary exploring the origins of the QAnon movement.

The documentary, Q: Into the Storm, debuted as a six-part series on HBO Max in March 2021. Twitter decided to “limit the visibility” of the series on its social network shortly after the release, a Twitter spokesperson said.

Twitter admitted that it was restricting the reach of tweets about the series after the director, Cullen Hoback, tried paying to boost his own tweet publicizing the film’s iTunes debut on March 21. He was barred from buying promotion for his tweet. An email from Twitter’s ad department stated the film had been “manually reviewed” and deemed to be in violation of the social network’s “inappropriate content” policy. The documentary criticizes Twitter for the role it has played in the spread of QAnon.

Believing the response in error, Hoback’s production house, Hyrax Films, reached out to members of the Twitter communications team to request help. A response came three days later. To Hoback’s surprise, Twitter informed him the decision was intentional.

“In 2021, Twitter made the decision not to allow promotion of this documentary via advertising on the platform,” the company said. “This decision was aligned with the actions we took to suspend accounts dedicated to QAnon and to limit the visibility of QAnon-related content on the platform generally. As a result, the client will not be able to promote this content.”

It’s unclear what additional steps Twitter has taken to limit the visibility of Hoback’s account or others discussing the series. Since Jan. 2021, accounts sharing QAnon-related content have been excluded from features like “search” and algorithms that offer users personalized “suggestions,” the company said. According to Twitter, tweets about the series meet the definition of “related content” under this policy.


This is just barmy, and points to a company where there’s no coherent understanding of policy. This is clearly an editorial edict of “don’t let QAnon stuff be on the network”, but not enough comprehension to say “this is a trusted outlet, so suspend that rule”. Even after someone raises the objection. And on the topic of QAnon? Stable doors and horses.
unique link to this extract

Twitter says mass deactivations after Musk news were ‘organic’ • NBC News

Ben Collins:


Twitter has been flooded with user reports of high-profile accounts losing thousands of followers after news broke that Tesla CEO Elon Musk would purchase the social network. The company said Tuesday that the “fluctuations in follower counts” came from “organic” account closures.

Some accounts on the political right, including that of Rep. Marjorie Taylor Greene, R-Ga., saw their follower counts skyrocket. Greene, who boasted 539,000 followers the day before news of Musk’s takeover, had 632,000 followers by Tuesday evening.

Twitter did not provide an exact number of accounts that were shuttered or activated in the hours after the ownership announcement Monday. It said it was looking into the “recent fluctuations in follower counts.”

“While we continue to take action on accounts that violate our spam policy which can affect follower counts, these fluctuations appear to largely be a result of an increase in new account creation and deactivation,” Twitter said in a statement. A spokesperson at Twitter who spoke on the condition of anonymity said the accounts that experienced the most severe drop-offs in followers were “high-profile accounts.”

Former President Barack Obama, the most followed user on Twitter, whose number had increased every day in April, lost more than 300,000 followers after Monday’s announcement. Pop star Katy Perry, the third-most-followed user on Twitter, lost more than 200,000 after the announcement.

Some right-wing politicians noticed and lauded the increased follower counts Tuesday.


unique link to this extract

Some people need to lighten up. It’s Twitter! • Garbage Day

Ryan Broderick:


Modern social platforms have learned that compressing our internet experience at the expense of our sanity means we use their products longer, thus making them more money. And I actually tend think conservatives like Ben Shapiro are being genuine when they say they think Twitter is a left-wing website. For a lot of people, regardless of political affiliation, websites that run on centralized feeds of content, whether it’s algorithmic like Facebook or Instagram or chronological like Twitter and Tumblr, are inherently alienating. Maybe Shapiro looks at his feed and feels like he’s being piled on in the same way I do. For the last decade, researchers, journalists, and politicians have devoted a countless amount of energy to figuring out how the internet is radicalizing us, but what if it’s not the algorithms or the extremist groups, what if it’s just the feed? What if we just aren’t meant to consume an endless stream of content all jammed together into one place?

The other day, I put a shortened version of what I’ve written above in a Twitter thread. I speculated that there may just be a single cohort of around 5-10 million super-posters in the US who move from platform to platform causing trouble, and cataclysmic moments like Tumblr’s porn ban or Musk’s purchase of Twitter act as an inflection point where they start thinking about moving elsewhere.

And maybe that troublesome 5-10 million super-posters are also the people who have the biggest issues with feed-based internet platforms. Then my thread got a bunch of angry replies from people claiming I was saying we need to remove sex workers from the internet. Which I didn’t say or mean in the slightest, but is honestly a very helpful way to illustrate the point I’m trying to make here.


I mean, of course Ben Shapiro thinks Twitter is left-wing, because essentially the whole world is more left-wing than he is; he’s just too much of a dolt to realise it. (Let’s dream, for a moment, of the multiverse where everyone now in the US is transported to other countries, to discover that what passes for “left-wing” in the US is milquetoast right-wing everywhere else.) The idea of a flash mob that floats around causing disruption sounds very feasible. One wonders where all the people who’ve left Twitter have gone. (Broderick laughs off the idea that they’ve gone to Tumblr.)
unique link to this extract

Snickers dick vein: meet the Twitter shitposter behind a viral post • Rolling Stone

EJ Dickson:


The pinned tweet on Juniper’s account is “it’s [so] incredibly easy to create fake news it’s actually ridiculous lol.” But that hasn’t stopped right-wing media outlets from frothing at the mouth over her tweets, most recently a fake news headline indicating that Snickers was set to remove the beloved “dick vein” from its eponymous candy bars.

On April 16, @JUNlPER, whose display handle is Transgender Marx, tweeted a doctored version of a news story with the headline “Snickers are officially caving and removing the world-renowned dick vein from the candy bar.”

As Juniper told co-hosts Brittany Spanos and Ej Dickson on Rolling Stone‘s podcast Don’t Let This Flop, the “Snickers dick vein” had been a meme for quite some time before her tweet, but she was inspired to post the fake headline after watching the right-wing media have a collective freakout over the candy company Mars (which also owns Snickers) changing the green M&M’s iconic footwear from thigh-high boots to sneakers. The move, according to a press release from the company, was to “better reflect a “more dynamic, progressive world” and to show off the character’s “personality, rather than their gender.”

“Tucker Carlson actually talked about it on Fox News, that it was ‘wokeifying’ M&Ms or whatever,” Juniper says, chuckling. She decided to post about Snickers removing the dick vein as “a satirical way to point out how just ridiculous [the media] are about some of these things — how they’ll take these stories and kind of run with it and not even verify anything.”


There are two trends in modern media: a lack of interest in verifying anything, and a tendency to surf the waves created by other media who haven’t bothered to verify and thus gin up false stories. At least this one is explanatory.
unique link to this extract

Outside the US, Elon Musk’s vision of a rules-free Twitter is expected to unlock violence and civil strife • Coda Story

Elery Biddle:


What would it mean for the majority of Twitter users, who live outside the US?

“That just doesn’t work in a country like India,” said Nikhil Pahwa, a tech expert and founder of Medianama, an India-focused tech policy publication based in New Delhi. India [where it has more than 20 million users] is Twitter’s third-largest market after the US and Japan.

“We have real world consequences from the kind of speech that Twitter enables. Our political parties are really, really adept at understanding how the algorithms work, how to create trends, how to make something shareable,” Pahwa said. “What they excel at is essentially fueling hate.”

In recent years, India’s ruling Bharatiya Janata Party and other hardline Hindu nationalist groups have made Twitter, alongside Facebook and WhatsApp, an essential platform for promoting their agendas, sometimes inciting violence against religious minorities, Muslims in particular.

“I think we’re in a situation where we need more moderation of hateful content and not less. I don’t think Musk understands or cares for whether people are getting polarized or killed in India,” Pahwa told me.

…“Everyone thinks they know how to do content moderation until it becomes their job,” said Mishi Choudhary, founder of the Software Freedom Law Center, a tech policy group in New Delhi.

“I am not sure how [Musk] plans to address censorship by proxy that countries like India demand,” she wrote in a message.


India, Nigeria, Ethiopia: all examples of countries where this is far more complex than Musk imagines.
unique link to this extract

• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.

Errata, corrigenda and ai no corrida: none notified

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.