Start Up No.1620: Twitter’s broken verification, the $1bn chip machines, China’s “privacy” law, UK pauses NVidia-ARM, and more

Let battle commence: who has paid for the space into which the passenger in front will recline their seat – you, or them? Or might it be both of you?CC-licensed photo by Matthew Hurst 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 11 links for you. Back again! I’m @charlesarthur on Twitter. Observations and links welcome.

Rethinking Twitter verification • Terence Eden’s Blog

Eden thinks:


I have my Twitter Tick™ because I’m a mediocre white man in the tech industry. No joke. During one of Twitter’s periodic bouts of opening up their verification programme, I applied and basically said “Don’t you know who I am??!” and got it. Sucks to be anyone other than me, I guess.

Twitter has paused the Verification process for now. In part, I suspect, due to overwhelming false positives and negatives.

The main problem, I think, is that no one knows what “Verified” means. Is the user a celebrity? A journalist? A spoof account? A friend of Twitter’s CEO? It’s all so unclear.

It seems to me that the answer is expanding Verification to take into account the different types of users of Twitter.

The Mastodon social network has the concept of “Bot” accounts. They are clearly marked as robots and that lets users know they’re interacting with a non-person.

Lots of journalists get verified by Twitter – because Twitter wants to integrate with newsrooms. Should a junior fashion reporter at the South Nowhere Bugle get the same sort of tick as the Chief Political Correspondent of the world’s biggest news show? It’s egalitarian, to be sure, but it doesn’t help a user determine if the Verified account Tweeting about Afghanistan is a credible source. Once the journo leaves the industry and becomes a pizza delivery driver, do they keep their tick?

…I think it’s time Twitter spoke to its users and understood what they want out of a Verification programme.


This is definitely the problem. As much as anything you want un-verification or anti-verification: to be told something is a bot, or not the real person. That would be just as valuable.
unique link to this extract

Reclining airline seat fights highlight an age-old legal battle over ownership • Slate

Dahlia Lithwick spoke to the law professor Michael Heller, whose new book “Mine! How the hidden rules of ownership control our lives” was inspired by his experiences on airplanes when people lean their seat back:


there’s a guy named James Beach who was flying from Boston to Denver, and he had actually a little plastic clamp called a Knee Defender, which you can buy online. It’s really very effective. You stick it on the seat in front of you, on the little tray table, and it keeps the seat in front of you from leaning back. On this particular flight, the woman in front of him tried to lean back. She couldn’t; she realized what was wrong. She asked him to take them off. He didn’t comply. She turned around and threw her water at him. The pilot did an emergency landing right away. They were taken off the flight. The plane went on to Denver an hour and 38 minutes late.

But those little Knee Defenders turn out to reveal a tremendous amount about the ownership conflicts that are all through our lives. The woman in front is saying, “That space behind my seat, it’s mine, because the little button reclines the seat.” And the guy behind, like the kids in the playground, he’s saying, “No, it was mine. I had it first, for my laptop,” or “I possessed it first with my knees.” So that wedge of space is an ownership battle, it turns out, between attachment and possession and first-in-time.

When I talk to audiences about that conflict, I always poll them, and it’s amazing to me that invariably half say the person in front is in the right, and half say the person in back is in the right. What’s most amazing is how each side is just amazed that anybody else could have a different view. It feels and looks and seems so obvious, what’s mine, the same way it is to toddlers on a playground. But that little conflict on the airplane seat is not just an accident, it turns out. It’s deliberately engineered by the airlines so they can sell that same space twice. Most of us are just polite; we try to work it out, and that’s true in all of the ownership conflicts we go through throughout our day, throughout our lives, in the Starbucks line, to line up at Disney World. Anywhere that we’re trying to make something mine, our experience is being engineered and designed by some owner to shape our behavior.


Selling the same space twice. Very neat explanation of why people get annoyed. (Via Ian Leslie’s weekly Substack.)
unique link to this extract

Facebook unveiled VR work meetings, which, no thanks • Buzzfeed News

Katie Notopoulos in a cannot-miss followup on Friday’s story:


This raises two grim possibilities. The first is the sad realization that science fiction icon Neal Stephenson’s metaverse — a collision of the physical and virtual in a shared online space — is a sad little office veal-penned in by floating whiteboards. The metaverse is the officeverse, and office work is boring. Meetings are boring. A large corpus of popular art is devoted to this concept.

The second is that like so many innovations touted as magnificent, world-changing shifts, this “embodied internet” that Zuckerberg is peddling is more of a sad-trombone “neat” than a Jobsian “BOOM.” We were promised flying cars and a VR whale jumping out of a basketball court. What we got is just another way to attend the work meetings we’re already sick of attending.

It’s “cool,” it looks pretty fun, I bet it’s nice to use. Is it life-changing? Mind-blowing? I don’t know. “A different kind of productivity experience” is not the first thing that comes to mind when I think of “metaverse.”

Earlier this spring Facebook revealed a wearable wrist device and glasses that could decipher neuron impulses from your brain to your hand. This was some very cool sci-fi stuff! We want the full experience! I want Facebook to steal my DNA and do something actually fucked up and weird and bad! Clone my ass, Zuck! Send my roboclones to fight in a space war against the Boston Dynamics dog robots while I bathe in a pod of goo! I want THAT. I don’t want more work meetings. No one wants more meetings. Please, sometimes just a phone call works.


“Veal-penned” is the most murderous phrase.
unique link to this extract

The company that makes the machines that makes the chips • Trung Phan


ASML is the most important company you’ve never heard of.

The $300+ bn Dutch firm makes the machines that make semiconductors. Each one costs $150m and access to them are a huge geopolitical flashpoint.


Fascinating thread (on one page, off Twitter) about this. Each machine also weighs 180 tonnes and shipping them requires three 747s and 20 trucks; or 40 shipping containers. To you for $1bn. Except, of course, if you’re Chinese – in which case the US will block the shipment.
unique link to this extract

Have you already had a breakthrough Covid infection? • The New Yorker

Condé Nast:


A new cohort study from Israel—conducted during the reign of Alpha, not Delta—provides perhaps the most rigorous evidence on the frequency and severity of breakthrough infections. Researchers examined what happened after Sheba Medical Center, Israel’s largest hospital, vaccinated more than 11,000 health-care workers between December, 2020, and April, 2021. During that period, around 1,500 workers experienced either a known coronavirus exposure or developed suspicious symptoms; of that number, 39—less than 3%—tested positive for the coronavirus. Those who got infected tended to have lower antibody levels. Most had mild symptoms; a third were asymptomatic; no one had to be hospitalized; and no one passed the virus on to others. At the same time, 19% of those who experienced a breakthrough infection—seven people—continued to have symptoms, such as cough, fatigue, or loss of smell, six weeks later. These findings were widely publicized, sometimes in ways that focussed on this final, alarming statistic. “Study: 20% of vaccinated health workers who test positive suffer from long covid,” one headline read. “One in five breakthrough cases among health care workers in Israel resulted in long covid,” announced another.

Such headlines, of course, fail to take account of the whole picture. In statistical terms, they neglect the “base rates” of infection. Take 19% of 3%, and you’ll find that the people tested in the study had about a one in 200 chance of developing long Covid. Even this number is too high: of the more than 11,000 vaccinated people in the hospital, the researchers tested only those who had symptoms or a known coronavirus exposure.

…The 19% figure itself may be higher than it should be. Farzad Mostashari, a former epidemic intelligence service officer at the C.D.C., has argued that a form of recall bias—a phenomenon in which people may attribute current symptoms to salient but unrelated prior events—could inflate the apparent prevalence of long covid. Suppose that you provide someone who’s been sick with a questionnaire, asking him to indicate which of a list of symptoms he has continued to suffer over the past few months. Faced with such a list, Mostashari wrote, on Twitter, “respondents often feel like they’re supposed to say ‘yes’ ” to some of the symptoms. In his view, this approach, which was taken in the Israeli study, could have led people to highlight symptoms needlessly.


In short: nobody knows anything. (Thanks G for the link.)
unique link to this extract

China just passed a major data privacy law—with a big, government-sized loophole • Gizmodo

Shoshana Wodinsky:


Like most privacy laws, the full PIPL is wordy and dense. But in a nutshell, it mandates that those who operate apps, sites, or any other tech doing data collection—obtain consent from their users in order to collect that data, just like we’ve seen with the GDPR. In cases where that app or device handles “sensitive” data like a person’s fingerprint or financial details, it’s required to ask for consent again before collecting those specific details, even asking operators to get “written consent” from users if the law requires it.

On top of that, the law also requires that users are given different options for how their data is allowed to be handled. Users must be allowed to, say, tell an app it can track their data, but not use that data to target them with ads. And when they give that consent, the app is required to give those users an easy way to withdraw it at any time. If you’ve seen the way Apple rolled out app tracking choices in iOS 14, what the law’s asking for sounds pretty similar. Only in this case, it won’t be Apple taking your app down if you’re caught flouting these requirements—it’s China’s government.

The PIPL also has pretty strict guidelines for foreign companies doing business in the region—and that includes data-hoovering giants like Facebook that offer services to Chinese customers through obscure subsidiaries. The PIPL states that any of these outfits aren’t only required to abide by the new law but that they need to “pass a security assessment organized by the State cybersecurity and information department” before they get a pass to operate in the country.

When companies get caught flouting privacy laws in the US, companies like Facebook are slapped with the same sort of punishment they’d get if they were caught violating those rules in then EU: thousands (sometimes millions) of dollars worth of fines. As you’d probably expect, the consequences for companies in China is much more severe.


It’s OK, they don’t kill the executives (yet). The fines are a lot bigger than the US, but about in line with Europe.

Oh, and user privacy doesn’t extend to protection from state oversight. Obviously.
unique link to this extract

We built a system like Apple’s to flag child sexual abuse material — and concluded the tech was dangerous • The Washington Post

Jonathan Mayer and Anunay Kulshrestha:


Our research project began two years ago, as an experimental system to identify CSAM in end-to-end-encrypted online services. As security researchers, we know the value of end-to-end encryption, which protects data from third-party access. But we’re also horrified that CSAM is proliferating on encrypted platforms. And we worry online services are reluctant to use encryption without additional tools to combat CSAM.

We sought to explore a possible middle ground, where online services could identify harmful content while otherwise preserving end-to-end encryption. The concept was straightforward: If someone shared material that matched a database of known harmful content, the service would be alerted. If a person shared innocent content, the service would learn nothing. People couldn’t read the database or learn whether content matched, since that information could reveal law enforcement methods and help criminals evade detection.

Knowledgeable observers argued a system like ours was far from feasible. After many false starts, we built a working prototype. But we encountered a glaring problem.

Our system could be easily repurposed for surveillance and censorship. The design wasn’t restricted to a specific category of content; a service could simply swap in any content-matching database, and the person using that service would be none the wiser.

…China is Apple’s second-largest market, with probably hundreds of millions of devices. What stops the Chinese government from demanding Apple scan those devices for pro-democracy materials? Absolutely nothing, except Apple’s solemn promise. This is the same Apple that blocked Chinese citizens from apps that allow access to censored material, that acceded to China’s demand to store user data in state-owned data centers and whose chief executive infamously declared, “We follow the law wherever we do business.”


Feels like it predates the “two intersecting databases” requirement.
unique link to this extract

Obligatory boost for my book Social Warming: why social media is driving everyone (even those who don’t use it) a bit mad.

San Francisco doctor charged with possessing child pornography in iCloud • AppleInsider

Mike Peterson:


Amid controversy surrounding Apple’s CSAM detection system, a San Francisco Bay Area doctor has been charged with possessing child pornography in his Apple iCloud account, according to federal authorities.

The U.S. Department of Justice announced Thursday that Andrew Mollick, 58, had at least 2,000 sexually exploitative images and videos of children stored in his iCloud account. Mollick is an oncology specialist affiliated with several Bay Area medical facilitates, as well as an associate professor at UCSF School of Medicine.

Additionally, he uploaded one of the images to social media app Kik, according to the recently unsealed federal complaint.


The distinguishing thing about these people is that they collect in large numbers. Every report I see suggests that they collect hundreds or, as in this case, thousands of pictures. (I’d imagine this was initially discovered through the postings to Kik, not Apple proactively scanning his iCloud account.)
unique link to this extract

Only small cuts in flying and driving needed to beat climate change, says Blair institute • The Times

Ben Webster:


Climate change can be tackled with small reductions in flying and driving and we can continue eating meat and dairy, according to a report by Tony Blair’s think tank.

It rebuts claims that meeting the UK’s legally binding target of net zero by 2050 will require a “total transformation” of people’s daily lives and says the number of behaviour changes needed over the next 15 years is “relatively limited”.

It adds that the emissions savings needed to hit the target would not require a reduction in living standards.

Average kilometres travelled per person by plane would need to fall by only about 6% between 2019 and 2035, the report by the Tony Blair Institute for Global Change says.

Kilometres travelled per driver would need to decline by about 4% over the same period, although 60% of cars by 2035 would need to be electric.

Meat and dairy consumption would have to be cut by 20% but “we do not all need to become vegetarian”, the report adds.


That all sounds feasible, but you’d need some hefty price signalling to make those things happen – all the trends are in the opposite direction.
unique link to this extract

Google shutting down health division as chief David Feinberg leaves • Business Insider

Hugh Langley and Blake Dodge:


Google is dismantling its health division, Insider has learned, as its chief, Dr. David Feinberg, departs the company.

Feinberg is joining Cerner, the electronic-medical-records behemoth, as its CEO and president, Cerner announced on Thursday. Meanwhile, the projects and teams that make up Google Health are being spread across different parts of Google, according to an internal memo obtained by Insider. Feinberg will leave on September 1.

Google Heath came together in 2018 as a way to organize the tech giant’s health efforts under one roof. The company poached Feinberg, a well-respected healthcare leader and the CEO of Geisinger Health, to lead it.

The group battled fallout and public distrust from a controversial deal with the health system Ascension, struggled to hammer out its roadmap, and let major deals fizzle out along the way.

Google’s healthcare challenges aren’t unusual for large companies with minimal healthcare experience trying to change the industry on a short timeline. In 2012, Google shuttered its first serious healthcare project, a service to let people store their medical records online. It was also named Google Health.

In a memo sent to employees on Thursday, Jeff Dean, the head of Google’s research division and Feinberg’s boss, said Google Health would no longer function as a unified organization.

Google’s chief medical officer, Dr. Karen DeSalvo, who leads the team focused on regulatory and clinical matters, will now report to Chief Legal Officer Kent Walker, the memo said.


Always a great time when you learn that your health boss is going to be reporting to the head of legal. Sure sign of deep interest by the company in, ah, what is it you do again? (But why has Health failed so badly when Google has just picked up Fitbit?)

And the day before, also at Business Insider:


Apple is scaling back a key project in its health division, four people familiar with the matter told Insider.

It’s an app called HealthHabit that Apple employees can use to log fitness goals, manage hypertension , and talk to clinicians and coaches at AC Wellness, the doctors’ group that Apple works with.

More than 50 employees were spending a significant amount of time working on the app. Some of them will be laid off with severance if they’re unable to find other roles inside Apple in the next few weeks, two of the people said.


unique link to this extract

CMA finds competition concerns with NVIDIA’s purchase of Arm • GOV.UK


The CMA [UK Competitions and Markets Authority] has determined that an in-depth investigation into the deal between NVIDIA and Arm is warranted on competition grounds.

Should the deal go ahead, the CMA is concerned that the merged business would have the ability and incentive to harm the competitiveness of NVIDIA’s rivals by restricting access to Arm’s intellectual property (IP). Arm’s IP is used by companies that produce semiconductor chips and related products, in competition with NVIDIA.

Ultimately, the CMA is concerned this loss of competition could stifle innovation across a number of markets, including data centres, gaming, the ‘internet of things’, and self-driving cars. This could result in more expensive or lower quality products for businesses and consumers.

NVIDIA offered a behavioural remedy – a measure which regulates the ongoing behaviour of a business – but the CMA found that this type of remedy would not alleviate its concerns. Therefore, the CMA found that the merger should be progressed to an in-depth Phase 2 investigation on competition grounds.


This is going to take quite a while, isn’t it. Can’t fault the CMA logic though. AMD for example would be worried about Nvidia getting early access to Arm improvements, or driving them in a direction that favours something about Nvidia rather than AMD.
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 )

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.