Facebook can tweak its algorithm so that better-quality news becomes more prominent. Yet won’t make it permanent. CC-licensed photo by vhines200 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. Why are there concession stands and concession speeches? I’m @charlesarthur on Twitter. Observations and links welcome.
Facebook struggles to balance civility and growth • The New York Times
Kevin Roose, Mike Isaac and Sheera Frenkel:
»
[In the days following the US election, as Trump falsely claimed widespread electoral fraud] the employees proposed an emergency change to the site’s news feed algorithm, which helps determine what more than two billion people see every day. It involved emphasizing the importance of what Facebook calls “news ecosystem quality” scores, or N.E.Q., a secret internal ranking it assigns to news publishers based on signals about the quality of their journalism.
Typically, N.E.Q. scores play a minor role in determining what appears on users’ feeds. But several days after the election, Mr. Zuckerberg agreed to increase the weight that Facebook’s algorithm gave to N.E.Q. scores to make sure authoritative news appeared more prominently, said three people with knowledge of the decision, who were not authorized to discuss internal deliberations.
The change was part of the “break glass” plans Facebook had spent months developing for the aftermath of a contested election. It resulted in a spike in visibility for big, mainstream publishers like CNN, The New York Times and NPR, while posts from highly engaged hyperpartisan pages, such as Breitbart and Occupy Democrats, became less visible, the employees said.
It was a vision of what a calmer, less divisive Facebook might look like. Some employees argued the change should become permanent, even if it was unclear how that might affect the amount of time people spent on Facebook. In an employee meeting the week after the election, workers asked whether the “nicer news feed” could stay, said two people who attended.
…The trade-offs came into focus this month, when Facebook engineers and data scientists posted the results of a series of experiments called “P(Bad for the World).”
The company had surveyed users about whether certain posts they had seen were “good for the world” or “bad for the world.” They found that high-reach posts — posts seen by many users — were more likely to be considered “bad for the world,” a finding that some employees said alarmed them.
So the team trained a machine-learning algorithm to predict posts that users would consider “bad for the world” and demote them in news feeds. In early tests, the new algorithm successfully reduced the visibility of objectionable content. But it also lowered the number of times users opened Facebook, an internal metric known as “sessions” that executives monitor closely.
«
At this point it’s becoming trivial to say that Facebook is utterly toxic, but it truly is. My forthcoming book will go into more detail. (The cover there is a placeholder, and the publication date might – will? – come forward. The topic is what it’s all about, though.)
unique link to this extract
New Honor expected to capture 2% smartphone market share in 2021 due to limited foundry capacity • TrendForce
»
Despite the change of ownership [sold by Huawei to a Shenzhen company], however, “new Honor” still has to cope with the shortage of foundry capacity in 2021, leading to a forecasted market share of 2%, while Huawei’s market share is expected to reach 4%. It should be pointed out that Apple is expected to capture some demand that was previously aimed at Huawei’s high-end smartphones. At the same time, Huawei’s Chinese competitors Xiaomi, OPPO, and Vivo are expected to ramp up device production. Hence, the volume of new smartphones coming from these sources will exceed the estimated market share gap left by Huawei. Also, if the smartphone market does not have sufficient demand to accommodate the overly inflated production plans in 2021, then brands may have to readjust their production targets.
Huawei had been adopting a coopetition strategy with Honor, which comprised of resource sharing and independent operation, and the latter is expected to return swiftly to the smartphone market through cooperation with channels under its preexisting model of independent operation. However, US sanctions remain as the most pressing concern for new Honor, as they affect its component procurement, R&D, product design, and GMS (Google Mobile Services) integration. Whether new Honor will be free to undertake these activities due to its split from Huawei remains to be seen.
«
A great deal now hinges on what Joe Biden’s administration does in January. Reverse the sanctions on Huawei as an olive branch? Or hold them as a bargaining chip against China for who knows what in return?
unique link to this extract
This Week in Apps: Apple slashes commissions, Twitter launches Fleets, warnings about Parler • TechCrunch
Sarah Perez:
»
According to App Annie data, around 98% of all iOS developers in 2019 (meaning, unique publisher accounts) fell under the $1m annual consumer spend threshold. This supports Apple’s claims that the “vast majority” of developers would benefit. This group of developers accounts for 567,000 unique apps, or 93% of all apps generating revenue through in-app purchases.
Combined, their revenues represented just under 8% of the overall App Store revenue share — in other words, it’s money Apple could stand to lose.
App Annie also found that the group of mid-range developers who are “nearing” that $1 million threshold is really small. The data indicates roughly 0.5% of developers are making between $800,000 and $1 million. And just over 1% are in the $500,000-$800,000 range.
Most developers have much smaller revenue streams, with 87.7% making less than $100,000 in 2019.
«
Useful for understanding that Apple decision to cut commissions (for developers who apply) to 15% from 30% if they have revenues of less than $1m. It sounds like app revenue is extremely bimodal – most of it under $100,000 and then a significant number of really big companies doing well over $1m. Apple will have known this for absolutely ages, meaning it could pick its time to introduce this. It could probably announce a fresh “under $500k” lower-commission tier at some future point without doing much harm to revenues, and gaining the PR benefit again.
unique link to this extract
UK government running ‘Orwellian’ unit to block release of ‘sensitive’ information • openDemocracy
Peter Geoghegan, Jenna Corderoy and Lucas Amin:
»
The British government has been accused of running an ‘Orwellian’ unit in Michael Gove’s office that instructs Whitehall departments on how to respond to Freedom of Information requests and shares personal information about journalists, openDemocracy can reveal today.
Experts warn that the practice could be breaking the law – and openDemocracy is now working with the law firm Leigh Day on a legal bid to force Gove’s Cabinet Office to reveal full details of how its secretive ‘Clearing House’ unit operates.
Freedom of Information (FOI) requests are supposed to be ‘applicant-blind’: meaning who makes the request should not matter. But it now emerges that government departments and non-departmental public bodies have been referring ‘sensitive’ FOI requests from journalists and researchers to the Clearing House in Gove’s department in a move described by a shadow cabinet minister as “blacklisting”.
This secretive FOI unit gives advice to other departments “to protect sensitive information”, and collates lists of journalists with details about their work. These lists have included journalists from openDemocracy, The Guardian, The Times, the BBC, and many more, as well as researchers from Privacy International and Big Brother Watch and elsewhere.
The unit has also signed off on FOI responses from other Whitehall departments – effectively centralising control within Gove’s office over what information is released to the public.
«
They could always use the excellent WhatDoTheyKnow site, which lets you file FOI claims; though they’re all publicly visible, there are so many that you could effectively hide. And would the Cabinet Office know that they were coming from a journalist? Again, you could adopt an identity. If you’re going to fight Orwellian methods, you have to fight fire with fire.
unique link to this extract
Parler is growing but conservatives are not ready to leave Twitter • The Washington Post
Drew Harwell and Rachel Lerman:
»
“Stop the Digital Inquisition! JOIN PARLER,” tweeted Dan Bongino, a Parler investor and right-wing star who consistently ranks among Facebook’s top-performing link posts nationwide, on Nov. 11, one of his 90 tweets that day — the same day he posted on Parler 51 times.
Madison Gesiotto, a pro-Trump commentator who tweeted to her 190,000 followers that she was “sick of big tech censorship,” has posted five times to Parler but 95 times to Twitter since declaring (in a tweet) that social media is “worse than ever before!”
Perhaps that’s understandable. Conservative provocateurs have mastered the art of getting attention and amplifying opinions on the very social networks they so roundly criticize. But Parler’s rise highlights how the polarized national debate could even further splinter the American Internet, in the same way that news sources and digital social circles have split into parallel partisan realities.
…Parler and Gab now average about 5 million views a month, which makes them, in social media terms, microscopic. Their combined traffic worldwide last month was 0.05% of Facebook’s and 0.22% of Twitter’s, SimilarWeb data show. But Andrew Torba, Gab’s chief executive, said Parler’s growth further validates the “alt-tech ecosystem.”
«
The numbers underplay the effect of those networks being small. Metcalfe’s Law says that the effect of a network grows proportionally to the square of the number of users. On Twitter, you can have colossal effect; on Parler, barely any. Smaller social networks have less impact. Simple as that.
unique link to this extract
Right-wing social media had to divorce from reality • The Atlantic
Renée DiResta:
»
The president himself, while tweeting about how the election was being stolen, amplified accounts that touted OANN and Newsmax as places to find accurate reporting on the truth about his election victory.
And on Parler, the conspiracy-mongering has grown only more frenzied as Trump makes state-by-state fraud allegations. In addition to concerns about Sharpies, the social network abounds with rumors of CIA supercomputers with secret programs to change votes, allegations of massive numbers of dead people voting, claims of backdated ballots, and assorted other speculations that users attempt to coalesce into a grand unified theory of election theft.
How far these ideas spread depends in part on whether mainstream social-media outlets keep moderating content as closely as they did during this election season. For most of Trump’s term, Facebook and others had been loath to crack down on even baseless conspiracy theories, including those repeated by the president himself.
Freedom of expression, the argument went, covers the right to think and say even floridly false things, which were best addressed through corrections and counter-speech. Yet the major platforms concluded that misleading theories about the election were a distinct class of misinformation because of their potential to cause significant harm to the body politic. As the split between reality-based information outlets and those catering to pro-Trump bitter-enders has widened, the distribution of their content is becoming significantly siloed.
This could have two major effects: It may limit the spread of conspiracy theories and reduce the possibility that Facebook and YouTube recommendation algorithms will draw casual users into the world of QAnon. But the bifurcation also raises the possibility that, among those who gravitate to niche platforms like Parler, the discussion may grow even more extreme.
«
As above, I think that if there is a move away from the big social networks to smaller ones, then that will actually be a good thing: less fertile space, and though the echo might be intense, it won’t reach anyone who hasn’t already bought into the nonsense.
unique link to this extract
We will not understand Covid until we give up debating it • Tim Harford
»
We all have a tendency to think with our hearts rather than our heads, and that tendency is sharpened, not dulled, by a vociferous argument. Wishful thinking, tribal loyalty, and tortured logic are ever-present pitfalls, but the pits yawn wider and deeper once a few alpha chimps are yelling at each other about “covidiots” and “face-nappies.”
A disheartening autumn provides us with an interesting case in point. At the end of August, the virus seemed to be in retreat. The prevalence survey published by the Office for National Statistics on 4th September, covering late August, suggested that infections had fallen to 36 per million people per day in England. Even for the highly vulnerable, the risk of taking a day out was looking small. But then each new week showed a large increase, and by 25th September, the estimate of infections was up to 175 per million people per day—mostly in the under-35s, and mostly in London and the north of England.
Those are the facts. But the facts were not of much interest: cabinet ministers blamed the public, lockdown sceptics blamed false positives, and newspaper columnists mocked the government for reversing its stance from “get back to the office” to “actually, stay at home.”
Everyone got their zingers in, but an ordinary citizen, trying to weigh up the health risks she faces, her responsibility to keep others safe, and the threats to her livelihood, is none the wiser. The personal risk remains low for most people, but the fact that cases have risen so rapidly suggests that we have a real challenge on our hands.
The truth, it turns out, is complicated. But complicated is no way to win a shouting match. If we want to understand the virus—and, for that matter, anything else in a complex world—we must first give up on the illusion that what passes for public “debate” is about anything more than scoring cheap points, which inevitably come at the cost of the whole truth.
«
Harford is arguably the UK’s Zeynep Tufekci, and I say that as a marker of great respect. (He’s got a new book out: How To Make The World Add Up. It might make a good Christmas present for a friend.)
unique link to this extract
Why Newsmax supports Trump’s false voter-fraud claims • The New Yorker
Isaac Chotiner interviews Chris Ruddy, CEO of Newsmax, the deluded media outlet which reckons it’s good policy to pretend Trump hasn’t lost the election:
»
Chotiner: I think we know that America’s voting systems are not pristine and shouldn’t be unquestioned. I think the problem is just saying that it was fraudulent, and that the election was stolen. I think that’s the main problem.
Ruddy: I think in very close elections, you might be able to make that claim, if it’s not properly done.
IC: But you’re the head of a news organization, and so if you want to just send reporters to investigate this and prove this, fine, but the Trump Administration in court is not actually showing any of these things. They’re just making the claims, and then not producing any evidence.
CR: Well, I think, for instance, they’ve been showing that there were issues with the mail-in ballots, the process by which they were counted. There was a story that indicated the rejection rates on absentee ballots and mail-in ballots was a lot lower in this election, in many of these swing states. That’s unusual. Now, is that indicative of fraud? I would say it’s not a direct indication, but it’s suggestive. And it’s something that we should look at. [A Times story on the low rejection rates suggested that there were multiple factors, including simplified voting requirements and greater enthusiasm and attentiveness among voters.]
IC: But that’s something that journalists could investigate, rather than making broad claims about it.
CR: Yeah. Well, I think before we even make the claim, we should say, “Hey, look at this anomaly. Why is this the case?” And we start asking about it. But you know what? At the end of the day, it’s great for news. The news cycle is red-hot, and Newsmax is getting one million people per minute, according to Nielsen, tuning into Newsmax TV. I think it’s good.
«
Basically: claiming fraud is great for clicks! Who cares if it’s a lie and we don’t actually bother to investigate it? Essentially, a form of free riding on Trump’s delusional nonsense. Useful to know what democracy is up against, though.
unique link to this extract
Early coronavirus mutation made it harder to stop, evidence suggests • The New York Times
James Glanz, Benedict Carey and Hannah Beech:
»
one mutation near the beginning of the pandemic did make a difference, multiple new findings suggest, helping the virus spread more easily from person to person and making the pandemic harder to stop.
The mutation, known as 614G, was first spotted in eastern China in January and then spread quickly throughout Europe and New York City. Within months, the variant took over much of the world, displacing other variants.
For months, scientists have been fiercely debating why. Researchers at Los Alamos National Laboratory argued in May that the variant had probably evolved the ability to infect people more efficiently. Many were skeptical, arguing that the variant may have been simply lucky, appearing more often by chance in large epidemics, like Northern Italy’s, that seeded outbreaks elsewhere.
But a host of new research — including close genetic analysis of outbreaks and lab work with hamsters and human lung tissue — has supported the view that the mutated virus did in fact have a distinct advantage, infecting people more easily than the original variant detected in Wuhan, China.«
Outbreaks grew faster, infections happened more quickly. That might, partially, explain why China was less affected if it managed an efficient lockdown. (Though I thought the mutation was called D614G – the replacement at locus 614 of glycine where previously there was aspartic acid residue; it’s on the spike. (Thanks G for the link.)
unique link to this extract
Thread by @BethanyAllenEbr on Thread Reader App • Thread Reader App
Bethany Allen-Ebrahimian is a reporter for Axios:
»
According to sources I have spoken to with knowledge of the matter, this Washington Post story does not accurately characterize Apple’s position on the Uyghur Forced Labor Prevention Act.
It is not accurate to say that Apple’s aim is to water down key provisions of the bill, and it is not accurate to characterize Apple as lobbying against the bill.
(And no, the sources I am citing are not a strident email from Apple’s PR department).
«
You’ll recall that I was a bit dubious about that article: it seemed to accuse Apple of things that it said it was diametrically opposed to, and it wasn’t named in the Act, unlike a number of other companies. (Thread Reader App lets you create a single page from a Twitter thread: to get it to happen, respond to any of the tweets in the thread you want “unrolled” with “@threadreaderapp unroll please” and it will respond so that your off-Twitter friends, or those who just want a simpler life, can read it in one gulp.)
unique link to this extract
Errata, corrigenda and ai no corrida: none notified
The Overspill: when there's more that I want to say 