Start Up No.1345: scientists warn on airborne Covid, how Facebook flattens us, iOS 14’s password watch, MIT’s bad images, and more

Seems we have these folks – the Neanderthals – to blame for some susceptibility to SARS-Cov-2. CC-licensed photo by Allan Henderson 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. Restart! I’m @charlesarthur on Twitter. Observations and links welcome.

239 experts with one big claim: the coronavirus is airborne • The New York Times

Apoorva Mandavilli:


The coronavirus is finding new victims worldwide, in bars and restaurants, offices, markets and casinos, giving rise to frightening clusters of infection that increasingly confirm what many scientists have been saying for months: The virus lingers in the air indoors, infecting those nearby.

If airborne transmission is a significant factor in the pandemic, especially in crowded spaces with poor ventilation, the consequences for containment will be significant. Masks may be needed indoors, even in socially-distant settings. Health care workers may need N95 masks that filter out even the smallest respiratory droplets as they care for coronavirus patients.

Ventilation systems in schools, nursing homes, residences and businesses may need to minimize recirculating air and add powerful new filters. Ultraviolet lights may be needed to kill viral particles floating in tiny droplets indoors.

The World Health Organization has long held that the coronavirus is spread primarily by large respiratory droplets that, once expelled by infected people in coughs and sneezes, fall quickly to the floor.

But in an open letter to the W.H.O., 239 scientists in 32 countries have outlined the evidence showing that smaller particles can infect people, and are calling for the agency to revise its recommendations. The researchers plan to publish their letter in a scientific journal next week.


This surfaced over the weekend and has created a furore, science-wise. The implication is that if you’re not outdoors, you’re at risk (at least if people aren’t all using masks). Remember the restaurant that helped infect three different families? It’s essentially that.
unique link to this extract

DNA linked to Covid-19 was inherited from Neanderthals, study finds • The New York Times

Carl Zimmer:


“This interbreeding effect that happened 60,000 years ago is still having an impact today,” said Joshua Akey, a geneticist at Princeton University who was not involved in the new study.

This piece of the genome, which spans six genes on Chromosome 3, has had a puzzling journey through human history, the study found. The variant is now common in Bangladesh, where 63% of people carry at least one copy. Across all of South Asia, almost one-third of people have inherited the segment.

Elsewhere, however, the segment is far less common. Only 8% of Europeans carry it, and just 4% have it in East Asia. It is almost completely absent in Africa.

It’s not clear what evolutionary pattern produced this distribution over the past 60,000 years. “That’s the $10,000 question,” said Hugo Zeberg, a geneticist at the Karolinska Institute in Sweden who was one of the authors of the new study.

One possibility is that the Neanderthal version is harmful and has been getting rarer over all. It’s also possible that the segment improved people’s health in South Asia, perhaps providing a strong immune response to viruses in the region.


Un-peer-reviewed, it’s worth saying. Although the top line is that it makes you more vulnerable to severe illness, the much higher occurrence in Asia and lower in Europe argues the other way.

But if it is dangerous, then it’s the best example of revenge being a dish best served cold. Quite the move there, Neanderthals. (Thanks G for the link.)
unique link to this extract

How Facebook has flattened human communication • OneZero

David Auerbach, writing in 2018:


The conclusions and impact of data analyses more often flow from the classifications under which the data has been gathered than from the data itself. When Facebook groups people together in some category like “beer drinkers” or “fashion enthusiasts,” there isn’t some essential trait to what unifies the people in that group. Like Google’s secret recipe, Facebook’s classification has no actual secret to it. It is just an amalgam of all the individual factors that, when summed, happened to trip the category detector. Whatever it was that caused Facebook to decide I had an African-American “ethnic affinity” (was it my Sun Ra records?), it’s not anything that would clearly cause a human to decide that I have such an affinity.

What’s important, instead, is that such a category exists, because it dictates how I will be treated in the future. The name of the category — whether “African American,” “ethnic minority,” “African descent,” or “black” — is more important than the criteria for the category. Facebook’s learned criteria for these categories would significantly overlap, yet the ultimate classification possesses a distinctly different meaning in each case. But the distinction between criteria is obscured. We never see the criteria, and very frequently this criteria is arbitrary or flat-out wrong. The choice of classification is more important than how the classification is performed.


The way that we’re reduced to bits is easy to forget, but it’s part of why these systems distort our existences so much.
unique link to this extract

Spies, lies, and stonewalling: what it’s like to report on Facebook • Columbia Journalism Review

Jacob Silverman:


In conversations with more than fifteen journalists and industry observers, I tried to understand what it is like to cover Facebook. What I found was troublesome: operating with the secrecy of an intelligence agency and the authority of a state government, Facebook has arrogated to itself vast powers while enjoying, until recently, limited journalistic scrutiny. (Some journalists, like The Observer’s Carole Cadwalladr, have done important work linking Facebook data to political corruption in the UK and elsewhere.) Media organizations have stepped up their game, but they suffer from a lack of access, among other power asymmetries.

Many journalists contacted for this story declined to talk out of fear of hurting relationships with Facebook’s communications shop. A number of journalists agreed to be interviewed, only to pass after speaking to their editors and PR reps. Some spoke to me off the record.

Nearly everyone I talked to acknowledged that the relationship between Facebook and journalists had dramatically deteriorated in recent years. It wasn’t long ago, after all, that Facebook and its comms shop was, for many journalists, a valued source.


I can’t think of a time when I thought of Facebook as a valued source. Certainly though in my experience it moved, exactly like Google, from being eager to talk up new products and give on-the-record briefings about those things, to being secretive and reactive. The shift wasn’t just about going public; there was also a size element (harder to coordinate things), and numerous scandals (for Google and Facebook) which subsequently made them wary of answering questions.

It’s a great piece, this.
unique link to this extract

New Trump appointee puts global internet freedom at risk, critics say • The New York Times

Pranshu Verma and Edward Wong:


In less than a decade, the Open Technology Fund has quietly become integral to the world’s repressed communities. Over two billion people in 60 countries rely on tools developed and supported by the fund, like Signal and Tor, to connect to the internet securely and send encrypted messages in authoritarian societies.

After Mr. Pack was confirmed for his new post on June 4, following a personal campaign of support by President Trump, Mr. Pack fired the technology group’s top officials and bipartisan board, an action now being fought in the courts. A federal judge on Thursday ruled in Mr. Pack’s favor, a decision that plaintiffs will likely appeal.

On Friday, Mr. Pack appointed an interim chief executive, James M. Miles, to head the fund, according to a letter obtained by The New York Times. Mr. Miles is little known in the internet freedom community, and his appointment needs approval from the fund’s new board, which is stacked with Trump administration officials and chaired by Mr. Pack.

The move was a victory for a lobbying effort backed by religious freedom advocates displeased with the fund’s work and who are often allied with conservative political figures.

This battle revolves around software developed by Falun Gong, the secretive spiritual movement persecuted by the Chinese Communist Party.


At this point the story takes quite a weird turn, but the key point is that the OTF has always pushed open source software, because it’s easier to spread to countries, easier to verify. (Signal and Tor have been pushed by the OTF.) The bonkers group pushing at Pack wants him to boost a closed-source app. That creates all sorts of risks from fakes and man-in-the-middle attacks.

Every one of these political appointments makes the US that little bit worse.
unique link to this extract

iOS 14: iCloud Keychain now alerts users about leaked passwords, more • 9to5Mac

Filipe Espósito:


If you’re not familiar with iCloud Keychain, it stores and syncs all your passwords from different websites and apps through iCloud. Users can access the iCloud Keychain with an iPhone or iPad by opening the Settings app and then tapping the Passwords menu.

With iOS 14, Apple offers a new “Security Recommendations” menu that shows only your passwords that could put your accounts at risk for some reason. This includes passwords that are easy to guess and even those that may have leaked on the web.

iCloud Keychain now clarifies what the problem is with the password for each specific account saved there, so users can learn more about creating stronger passwords. Some of these features were already present in iOS 13, but they weren’t as prominent as in iOS 14.

Here are some examples of security alerts provided by iCloud Keychain on iOS 14:
• Many people use this password, which makes it easy to guess.
• This password is easy to guess.
• This password uses a sequence, “123”. Using commom patterns makes passwords easy to guess.

There’s a new alert in particular that’s also one of the most important for users. According to Apple, iCloud Keychain now verifies if your passwords are involved in a data breach.


Though that last one will probably mean you’ll get an alert all the time if you’ve signed in many places. Data breaches are more than an everyday occurrence.
unique link to this extract

Google-backed groups criticize Apple’s new warnings on user tracking • Reuters

Stephen Nellis and Paresh Dave:


Apple last week disclosed features in its forthcoming operating system for iPhones and iPads that will require apps to show a pop-up screen before they enable a form of tracking commonly needed to show personalized ads.

Sixteen marketing associations, some of which are backed by Facebook and Alphabet’s Google, faulted Apple for not adhering to an ad-industry system for seeking user consent under European privacy rules. Apps will now need to ask for permission twice, increasing the risk users will refuse, the associations argued.

Facebook and Google are the largest among thousands of companies that track online consumers to pick up on their habits and interests and serve them relevant ads.

Apple said the new feature was aimed at giving users greater transparency over how their information is being used. In training sessions at a developer conference last week, Apple showed that developers can present any number of additional screens beforehand to explain why permission is needed before triggering its pop-up.


Nobody’s going to say “yes, track me everywhere!” For the small adtech companies, it’s going to be a test of whether they really do need to do tracking to earn their money.
unique link to this extract

NewsGuard: The internet trust tool


Get detailed ratings of more than 4,500 news websites that account for 95% of online engagement with news. See ratings displayed as icons next to links on all the major search engines, social media sites, and platforms.

See who’s behind each site and whether it has a record of publishing accurate information. Learn how each site fares on the nine journalistic standards NewsGuard uses to assess each site.

Get warnings on new trending misinformation sites as they are flagged and rated by NewsGuard’s 24/7 rapid response SWAT team.

[But] NewsGuard is exclusively for personal use, for use by journalists reporting on misinformation, or for use by our school and library partners. Any use of NewsGuard’s data by researchers, or any commercial use, including by moderators, coders and others at technology platforms, search providers or other companies, is strictly prohibited unless expressly permitted…


Interesting project. Browser extensions for Edge, Explorer, Chrome, Firefox and Safari. Mobile apps for iOS and Android. It rates Gateway Pundit as “completely full of crap” (I paraphrase). Not free: it’s £2.95 per month.
unique link to this extract

MIT apologizes, permanently pulls offline huge dataset that taught AI systems to use racist, misogynistic slurs • The Register

Katyanna Quach:


MIT has taken offline its highly cited dataset that trained AI systems to potentially describe people using racist, misogynistic, and other problematic terms.

The database was removed this week after The Register alerted the American super-college. MIT also urged researchers and developers to stop using the training library, and to delete any copies. “We sincerely apologize,” a professor told us.

The training set, built by the university, has been used to teach machine-learning models to automatically identify and list the people and objects depicted in still images. For example, if you show one of these systems a photo of a park, it might tell you about the children, adults, pets, picnic spreads, grass, and trees present in the snap. Thanks to MIT’s cavalier approach when assembling its training set, though, these systems may also label women as whores or bitches, and Black and Asian people with derogatory language. The database also contained close-up pictures of female genitalia labeled with the C-word.

Applications, websites, and other products relying on neural networks trained using MIT’s dataset may therefore end up using these terms when analyzing photographs and camera footage.


An old but much-used AI picture training database, dating back to the early 2000s, where nobody had actually bothered (we have to assume) to check quite how the content was labelled. Until these researchers did.
unique link to this extract

Facebook admits Ben Shapiro is breaking its rules • Popular Information

Judd Legum:


Last week, Popular Information exposed how The Daily Wire has gained unprecedented distribution on Facebook through its relationship with Mad World News. Five large Facebook pages controlled by Mad World News expanded The Daily Wire’s audience by millions through the coordinated posting of dozens of links from The Daily Wire each day. 

Facebook previously said it had looked into the matter, found no evidence of a violation, and could not prove a financial relationship. The company now admits the two publishers are working together.

“After further investigation, we’ve found that these Pages violate our policies against undisclosed paid relationships between publishers. Our enforcement typically focuses on the Page distributing the cross-promoted content, which is why we are temporarily demoting Mad World News. We are also warning Daily Wire and will demote them if we see this behavior continue,” a Facebook spokesperson said.


Legum is chipping away, very gradually but thoroughly, at Shapiro’s little empire of paid-for Facebook echo chambers. (Facebook’s policy: “Facebook pages cannot ‘accept anything of value to post content that you did not create.'”) This is the first takedown, and more will surely follow.

Of course, far too difficult for Facebook, with its gigantic systems, to spot that someone’s posts always get echoed at precisely the same time across a whole ton of Pages. It needs one person asking difficult questions – and even then Facebook initially denied there was a problem.
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.