Start Up No.1317: Facebook knew it was polarising, ban Trump’s Twitter?, how Bill Gates became conspiracists’ villain, YouTube’s odd deletion, and more

In Turkey, an AI will tell your future based on this. Reliably? Well… CC-licensed photo by Jette chan 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 10 links for you. Enough? I’m @charlesarthur on Twitter. Observations and links welcome.

Facebook executives shut down efforts to make the site less divisive • WSJ

Jeff Horwitz and Deepa Seetharaman:


Facebook launched its research on divisive content and behavior at a moment when it was grappling with whether its mission to “connect the world” was good for society.

Fixing the polarization problem would be difficult, requiring Facebook to rethink some of its core products. Most notably, the project forced Facebook to consider how it prioritized “user engagement”—a metric involving time spent, likes, shares and comments that for years had been the lodestar of its system…

Even before the teams’ 2017 creation, Facebook researchers had found signs of trouble. A 2016 presentation that names as author a Facebook researcher and sociologist, Monica Lee, found extremist content thriving in more than one-third of large German political groups on the platform. Swamped with racist, conspiracy-minded and pro-Russian content, the groups were disproportionately influenced by a subset of hyperactive users, the presentation notes. Most of them were private or secret.

The high number of extremist groups was concerning, the presentation says. Worse was Facebook’s realization that its algorithms were responsible for their growth. The 2016 presentation states that “64% of all extremist group joins are due to our recommendation tools” and that most of the activity came from the platform’s “Groups You Should Join” and “Discover” algorithms: “Our recommendation systems grow the problem.”

…Asked to combat fake news, spam, clickbait and inauthentic users, the employees looked for ways to diminish the reach of such ills. One early discovery: bad behavior came disproportionately from a small pool of hyperpartisan users.

…Under Facebook’s engagement-based metrics, a user who likes, shares or comments on 1,500 pieces of content has more influence on the platform and its algorithms than one who interacts with just 15 posts, allowing “super-sharers” to drown out less-active users. Accounts with hyperactive engagement were far more partisan on average than normal Facebook users, and they were more likely to behave suspiciously, sometimes appearing on the platform as much as 20 hours a day and engaging in spam-like behavior. The behaviour suggested some were either people working in shifts or bots.


This is, partly, the 1% rule. But then Facebook amplifies it. Joel Kaplan, a former deputy chief of staff for George W Bush, comes out looking like a particular villain. (The Verge has a non-paywalled rewrite.)
unique link to this extract

Twitter must cleanse the Trump stain • The New York Times

Kara Swisher:


The company tends to be hands-off when a Trump controversy erupts, relying on a tenet that he is a public figure and also that it cannot sort out what is truth and a lie and is therefore better off letting its community argue it out. While that might work when it comes to some issues, it has broken down here [after Trump tweeted a conspiracy theory about the death of a woman who worked with Joe Scarborough; her deeply upset husband wrote to Twitter CEO Jack Dorsey requesting Trump’s tweets be deleted].

How to fix it is the digital equivalent of a Gordian knot, except there is no cybersword of Alexander the Great to slice it in half. Banning Mr. Trump outright, the most extreme move, seems to be a nonstarter, given Mr. Dorsey’s belief that less is more when it comes to governing. While it worked when [Alex] Jones was tossed off, a move that Mr. Dorsey came to last among the social media giants, doing the same to Mr. Trump would be quite different.

While I had thought throwing Mr. Trump off Twitter was not the worst idea — after all, what would the president do without his raging addiction to Twitter? — I have come to believe that a Trump ban would be pointless and too drastic. The firestorm it would set off would alone be disastrous for Twitter to manage and probably come with deep financial repercussions. If you think that is not a good enough reason, I invite you to visit the reality of living as a public company in the digital age.
Another solution being discussed inside Twitter is to label the tweets as false and link to myriad high-quality information and reporting that refute the tweets’ sinister insinuations. Sources told me that after initial hesitance in dealing with Mr. Trump’s tweets about Ms. Klausutis, the company has accelerated work on a more robust rubric around labeling and dealing with such falsehoods.
Again, top company executives hope that this placement of truth against lies will serve to cleanse the stain. I think this is both naïve and will be ineffective…


Swisher reckons Trump’s tweets about the woman should be deleted, as a one-off action to show that there are boundaries to taste. But Trump’s tweets don’t mention the woman’s name (though his idiot son’s tweets do). When you start wading into this, the water gets very deep very fast. The labelling option looks far, far better to me; which Twitter is now doing.
unique link to this extract

Social media has turned Bill Gates into the coronavirus pandemic’s fake villain • Buzzfeed News

Ryan Broderick:


The paranoia around the former Microsoft CEO has been building for months, festering in Facebook Groups and YouTube comment sections. Here’s how the conspiracy theorists, panicked and ignorant people, and technology platforms that allowed the hoaxes to grow turned Bill Gates into the villain of the coronavirus pandemic.

The most popular version of the rumor stems from a tabloid in Ghana.

In 2010, a former staffer with a government health initiative in Ghana claimed that a community health initiative, partially funded by the Gates Foundation, had tested the contraceptive Depo-Provera on unsuspecting villagers in Navrongo, a remote town in the country, as part of an illicit “population experiment.”

The woman making the charge, Mame-Yaa Bosumtwi, was the Ghanian-born, US-educated communications officer for a separate Gates-funded initiative by the Ghanaian government and Columbia University. The program used cellphones to improve healthcare access for women and children in rural areas. Bosumtwi had clashed with another team member, James Phillips, a demographer at Columbia University’s Mailman School of Public Health; when her contract was not renewed, she took her professional gripes with Phillips to the Ghanaian press and filed a lawsuit against Columbia for millions of dollars in damages.

After the lawsuit was dismissed, Bosumtwi went back to the press with a much more shocking claim. Without evidence, she said that Phillips’s project in rural Ghana had experimented with Depo-Provera on women as a test run for a broader population control campaign. Patients had been abused. Some had died.

Wanted posters with Phillips’s face sprouted across the country. Protesters mobilized outside Columbia’s research center in Navrongo. Ghanaian health officials called her claims libel, and community leaders and women from the rural area condemned them as false. But death threats escalated so badly that two members of Phillips’s team had to be evacuated across the border to Burkina Faso.


And things, uh, kinda spiralled from there. Your long read for the day. It’s a doozy.
unique link to this extract

Sero-surveillance of COVID-19 • GOV.UK


A number of serological collections have been established by PHE [Public Health England] to provide an age-stratified geographically representative sample across England over time. These include samples from healthy adult blood donors, supplied by the NHS Blood and Transplant (NHS BT). Donor samples from different geographic regions (approximately 1000 samples per region) in England are tested each week. The results presented here are based on testing using the Euroimmun assay. Figure 1 shows the overall prevalence in each region over time which has been adjusted for the accuracy of the Euroimmun assay (sensitivity and specificity).


So this is showing the proportion of people with antibodies. That graphic is pretty dramatic. London has had the highest case numbers (over 27,000) and deaths (nearly 6,000). That shows it being at best one-third (20%) of the way to low herd immunity (60%), at worst under one-fifth (13%) of the way to full herd (80%). A long way from being there. Also, I didn’t think blood donation was still going on.
unique link to this extract

Can a former model predict your future? A million Turkish users say yes • Rest of World

Kaya Genç:


In Apple’s App Store, Faladdin describes itself as “far beyond a fortune telling app.” The description states that it can predict one’s destiny “by evaluating a person’s past.” It does this by bringing the Turkish tradition of coffee fortune-telling into the Internet age. For centuries, Turks have boiled coffee grinds and water in the same pot, which leaves a residue that practitioners can read like a Rorschach inkblot. Thanks to Faladdin — and at the expense of Turkey’s traditional tellers — in-person consultations are no longer necessary. Every day, more than one million Faladdin users upload photos of their coffee cup grinds, and Taşdelen’s team provides personalized “readings” of them within 15 minutes. Of these readings, 700,000 are in Turkish, 200,000 are in Arabic, and 100,000 are in English…

…Faladdin collects data from users and then draws language from a pool of preformulated interpretations. These readings are produced by a group of 30 contributors, including a dramatist, a psychologist, an ad director, and an author. “They are like the precogs in ‘Minority Report,’” Taşdelen said of his writers. “They possess the psychic ability to see the future, but of course we don’t house them in pools.”

Last year, Taşdelen began feeding these texts to OpenAI, the artificial intelligence platform cofounded by Elon Musk, in order to train Faladdin to autonomously produce fortunes. Taşdelen refused to elaborate when I asked about the technical details, but he did say that the AI was initially a massive failure.


Fortune-telling by AI! As it happens, Janelle Shane, author of the AI Weirdness blog, has already trained a machine learning system to do western astrology. Read them here, with forecasts such as “come to your joyful love journey and ride your uptight, stoned tiny horses”. (Thanks Jim for the link.)
unique link to this extract

YouTube is deleting comments with two phrases that insult China’s Communist Party • The Verge

James Vincent:


YouTube is automatically deleting comments that contain certain Chinese-language phrases related to criticism of the country’s ruling Communist Party (CCP). The company confirmed to The Verge this was happening in error and that it was looking into the issue.

“This appears to be an error in our enforcement systems and we are investigating,” said a YouTube spokesperson. The company did not elaborate on how or why this error came to be, but said it was not the result of any change in its moderation policy.

But if the deletions are the result of a simple mistake, then it’s one that’s gone unnoticed for six months. The Verge found evidence that comments were being deleted as early as October 2019, when the issue was raised on YouTube’s official help pages and multiple users confirmed that they had experienced the same problem.

Comments left under videos or in live streams that contain the words “共匪” (“communist bandit”) or “五毛” (“50-cent party”) are automatically deleted in around 15 seconds, though their English language translations and Romanized Pinyin equivalents are not.

The term “共匪” is an insult that dates back to China’s Nationalist government, while “五毛,” (or “wu mao”) is a derogatory slang term for internet users paid to direct online discussion away from criticism of the CCP. The name comes from claims that such commenters are paid 50 Chinese cents per post.

These phrases seem to have been accidentally added to YouTube’s comment filters, which automatically remove spam and offensive text. The comments are removed too quickly for human moderation and are deleted even if the banned phrases are used positively (e.g., “The 五毛 are doing a fantastic job”). YouTube says it’s been relying more on its automated filters in recent months due changes to its workforce brought about by the pandemic.


I don’t believe for a second that they were added to the comment filters by accident. The example of Saudi infiltration of Twitter makes me suspect a similar infiltration here. It’s such a particular thing.
unique link to this extract

Early tests of vaccine for COVID-19 pass peer review, look promising • Ars Technica

John Timmer:


The first indication of progress toward a vaccine that we’re aware of came in the form of a company press release. This new one comes in the form of a peer-reviewed article in the prestigious medical journal The Lancet. Most of its authors are academic researchers or public health authorities; only two have affiliations with a company.

The two reports also differ significantly in terms of their approach to generating an immune response. The earlier announcement, from a company called Moderna, involved injecting carefully packed RNAs that encode the spike protein that normally resides on the surface of the virus. The RNAs transit inside a person’s cells and induce them to produce the spike protein, thereby exposing the immune system to it.

The Chinese researchers used a very different approach to inducing immunity. In their case, they engineered the gene that encodes the spike protein into a harmless virus called Adenovirus 5. They then produced large quantities of the engineered virus and injected that into people. Even though adenoviruses are essentially unrelated to coronaviruses (they use DNA as their genetic material, rather than RNA), the cells that the engineered viruses infect will produce the coronavirus spike protein, again exposing the immune system to it.


Promising start. Many Manhattan projects running in parallel, but to stop a bomb going off.
unique link to this extract

Formula E driver disqualified for getting impostor to race for him • Reuters

Alan Baldwin:


Audi Formula E driver Daniel Abt was disqualified and ordered to pay €10,000 ($10,900) to charity on Sunday for getting a professional gamer to compete under his name in an official esports race.

The German, who apologised for “having called in outside help”, was also stripped of all points won to date in the all-electric series’ Race at Home Challenge which features drivers using simulators remotely.

“I did not take it as seriously as I should have,” said the 27-year-old, accepting the punishment for sporting misconduct.

“I am especially sorry about this because I know how much work has gone into this project on the part of the Formula E organisation. I am aware that my offence has a bitter aftertaste but it was never meant with any bad intention.”

Pro gamer Lorenz Hoerzing, Abt’s ‘ringer’, was disqualified from all future rounds of the separate Challenge Grid competition.

The 15-lap race around a virtual Berlin Tempelhof track was won by Britain’s Oliver Rowland for Nissan e.dams with Belgian Stoffel Vandoorne second for Mercedes.


Pretty soon – next race? – you’ll have race administrators who will be there beside the racers to make sure they are actually the ones controlling it.
unique link to this extract

Facebook Messenger adds safety alerts—even in encrypted chats • WIRED

Andy Greenberg:


For the past year, governments around the world have pressured Facebook to abandon its plans for end-to-end encryption across its apps, arguing that the feature provides cover for criminals and, above all, child predators. Today Facebook is rolling out new abuse-detection and alert tools in Messenger that may help address that criticism—without weakening its protections.

Facebook today announced new features for Messenger that will alert you when messages appear to come from financial scammers or potential child abusers, displaying warnings in the Messenger app that provide tips and suggest you block the offenders. The feature, which Facebook started rolling out on Android in March and is now bringing to iOS, uses machine learning analysis of communications across Facebook Messenger’s billion-plus users to identify shady behaviors. But crucially, Facebook says that the detection will occur only based on metadata—not analysis of the content of messages—so that it doesn’t undermine the end-to-end encryption that Messenger offers in its Secret Conversations feature. Facebook has said it will eventually roll out that end-to-end encryption to all Messenger chats by default.


A good thing in its way, while also telling us that Facebook is monitoring everyone’s metadata (which we know in the back of our minds anyway, after all.)
unique link to this extract

Sony Mobile smartphone shipments hit low in 1Q20 • Digitimes

Max Wang and Steve Shen:


Sony Mobile Communications saw its quarterly smartphone shipments hit a 10-year low of 400,000 units in the first quarter of 2020 despite robust sales of new models, including Xperia 1 II and Xperia 10 II, in Japan.

The vendor has been keen on promoting its smartphones in Europe and Latin America, but its shipment prospects for the second quarter remain flat as demand in these two regions has been crippled by the coronavirus pandemic, according to sources from Taiwan’s handset supply chain.

Although it was the second largest vendor with a 10% share in Japan in the first quarter, the fact that it lagged top vendor Apple by 50pp in terms of market share is making it hard to further push its sales in its home market, said the sources.

In addition to the impacts of the pandemic, Sony will continue to face keen competition from Samsung, Motorola (Lenovo), Huawei, Xiaomi and LG in Latin America, the sources added, noting that Sony Mobile’s share in the region has fallen to around 1% recently.


Revenues shrank by 25% year-on-year, though losses narrowed a lot. Sony’s mobile business is – I keep saying – dead on its feet; it sold 4m phones in the whole of 2019, and you can’t blame coronavirus for that. Along with LG, it seems to be a sort of vanity publishing.
unique link to this extract

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 )

Twitter picture

You are commenting using your Twitter 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.