A report says Apple has put off a spring launch of an iPhone. So what will it do instead? CC-licensed photo by Mark Mathosian 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. Now wash someone else’s hands. I’m @charlesarthur on Twitter. Observations and links welcome.
Two and a half years ago, my dad and his wife took an Uber from the airport to my house in Washington, D.C., as they’ve done many times before. Only this time, it took them to the wrong place: an empty lot by a cemetery, two miles from my house. My dad told the Uber driver it was not my house, but the driver didn’t believe him. His app had led him here, so it must be right!
They all sat there for a few minutes, staring at the driver’s phone, paralyzed by the startling gap that had opened between the app and reality.
This is how we discovered that Google Maps had two locations listed for our home. One was right, one was wrong. This seemed like a pretty minor problem in the scheme of things, and it was. For a while, I even thought it was kind of wonderful. We could be anonymous! Even Google didn’t know where we lived!
But over time, as Google Maps got embedded in more and more apps, the problem worsened. Google Maps is used by Uber, Instacart, Lyft, Door Dash and even something called the Zombie Outbreak Simulator.
The second-most-popular maps app in the United States is Waze. Guess who owns Waze? Google Maps again! Soon Waze had our address wrong, too. Eventually, almost everyone who tried to find our house was directed to the wrong place.
I once dated a woman whose road was a mountweazel – it was intentionally not included in the index of the London A-Z (we used to use paper-bound maps, kids). Finding her house the first time sure was a challenge.
unique link to this extract
When I told my children, both “Harry Potter” fans, that I was going to check out an invisibility cloak, they were excited. I’d learned of Goldstein’s cloak in a scientific paper that he and his students produced about their work. But when I saw Goldstein in his sweatshirt, which featured a foreground of blurry organic shapes in orange, like a display of horribly irradiated vegetables, with dark, vaguely human shapes above, I couldn’t imagine Harry or Hermione wizarding with one. The only recognizable shape (to me) was what appeared to be a traffic light just below the neckline. Considered more generously, the pattern loosely evoked Georges Seurat’s “A Sunday Afternoon on the Island of La Grande Jatte,” as it might appear at the bottom of a swimming pool painted by David Hockney.
Then Goldstein stepped in front of the camera, and the yolo detector did a double take. It couldn’t see him at all. The computer saw the chair behind him (“Chair,” the bounding box was labelled) but not the six-foot-tall, thirty-six-year-old man standing right in front of it—Goldstein Unbound. I, in my supposedly anonymous city duds, was instantly detected and labelled. It was like a conceit from William Gibson’s 2010 science-fiction novel, “Zero History,” in which a character wears a T-shirt so ugly that CCTV cameras can’t see it.
The pattern on the sweatshirt was an “adversarial image”—a kind of deep-learning optical illusion that stopped the algorithm from seeing the person wearing it. Unlike poison attacks, which seek to subvert surveillance systems with bad data, adversarial attacks are images that have been engineered to take advantage of flaws in the way computers see. They are like hacks, but for artificial intelligence. The security vulnerabilities of operating systems and computer networks are widely known, but deep-learning A.I. systems are still new and so complex that scientists don’t yet fully understand the kinds of hacks they are vulnerable to.
By nightfall, the Ukrainian town of Novi Sanzhary, population 8,300, would be turned upside down, thrust into a state of panic and chaos stemming from residents’ fears that the novel coronavirus was going to bring death to this previously unheralded backwater.
Violence would soon break out, leading to nine police officers being injured, 24 people being arrested for rioting, with five officially charged, and a statement from President Volodymyr Zelensky describing the melee as “medieval” behavior. The government response would include visits not only by ministers but also by a celebrity TV doctor who tried to calm the town’s frayed nerves.
What caused it all? A toxic mix of limited information released by Ukraine’s authorities, disinformation spread by the public on social media, and a targeted fake news campaign from yet-to-be-identified malign actors, according to more than a dozen government officials, medical experts, and local residents interviewed by BuzzFeed News.
It all happened in the way that decline generally happens in American culture, which is one anxious, hopeful, cynical capitulation at a time. We have compressed and corroded and finally collapsed what used to be the core of a publication—its relationship with its readers, and the basic notion that one should not make it hard for them to read.
It goes without saying that everyone involved is perpetually maxed-out and stressed and scrabbling for a dwindling and finite amount of money in an arbitrary and artificially constricted ad economy that runs on wobbly, untrustable, and easily manipulated data. (A friend who works in advertising operations described the work as “a game of catching falling knives.”)
In the last half-decade, ads have rapidly migrated from the sides and top of the page into the actual text. This is the result of pressures created by the transition from desktop computers to mobile devices. The ads need to get seen on a screen with no margins.
The ads that stalk you down the page reflect advertisers’ demands that their ads remain “in view.” And all the clammy unbidden video stuff is exactly as desperate as it looks. Not many people will watch video ads if given any choice in the matter. Taking choice out of the equation helps a lot.
Some sites have deliberately made the experience of reading them for free more assaultive, in order to bully readers into buying subscriptions. For the price of a small monthly indulgence on your end, it can all go back to normal and your laptop’s fan can finally turn off.
It’s a rolling, desperate, iterative exercise in seeing how bad things can become before readers finally stop coming at all. The pseudonymous author of the newsletter “Last Week In Ad Ops” describes it as “a decision that we know is bad, bad for our product, users, brand… but we do it anyway[,] often telling ourselves it’s ‘just for this quarter/client/order.’”
That point about how newspapers tried to make their content easier to read is completely forgotten. (Though I’d say The Guardian’s website does make that effort.)
unique link to this extract
A couple of years ago, WeChat seemed like an unstoppable force. It was completely dominating the Chinese social media landscape, was beating Alibaba in payments and was making forays into e-commerce.
Fast-forward to 2020, and users are already spending more time on Douyin than they are on WeChat, swiping for hours between short videos. What went wrong?
It can seem easy to paint WeChat’s current demise by their inability to spot the short-video craze early.
I would, however, argue that WeChat started losing the battle earlier. WeChat neglected incremental changes that could have improved the platform.
Ten years ago, Facebook would frequently revamp its interface, each time creating an uproar among its users. Yet it kept going at it, constantly refining its user experience and design.
That’s something that WeChat didn’t do. The result: WeChat moments interface seems old fashioned compared to its Western competitors.
The App barely uses the phone screen, keeping pictures cluttered and forcing users to click in order to see them. Engagement buttons are also hidden in a sub-menu, making it harder for users to like photos or comments.
Facebook takes a very different approach, displaying engaging full-screen pictures and prominent actions calls to like and comment.
The truth is that WeChat has been innovating quite a bit: they launched payment, Mini-programs, and many other disruptive features that have changed the behavior of both users and businesses.
But WeChat has been looking down at small incremental changes. Tencent was apparently too afraid to make any modifications to its hit product which might anger or confuse users.
The inertia of success.
unique link to this extract
Ring will require two-factor authentication starting today (Update: Blink verifying emails) • Android Police
Ring has dealt with its fair share of privacy snafus (and then some), but its latest move might allay some of your fears. The Amazon-owned smart home company has instituted a new login policy, effective immediately. Now, you’ll need to enter a two-factor authentication (2FA) code every time you log into your account.
Ring follows in the footsteps of Google’s Nest, which made a similar move earlier this month. When you enter your username and password in the Ring app, you’ll get a one-time code via either email or SMS (your choice). Ring doesn’t mention authenticator apps as an option, unfortunately.
Two steps forward, one step back.
unique link to this extract
Apple is widely expected to unveil a slew of new products this month, including the affordable iPhone 9 (or iPhone SE 2, as it’s sometimes called) device that kept popping up in all sorts of rumors in the past few weeks, as well as a new iPad Pro generation. However, the new novel coronavirus threat has prompted Apple to revise is plans, and a leaker now says that not only is the March press event canceled, but Apple is also considering postponing the iPhone 9 announcement and release date. That all has to do with the COVID-19 outbreak that’s currently sweeping the globe with no end or even slowdown in sight.
Santa Clara County has just banned mass gatherings of people due to the increasing number of coronavirus cases in the region, and the ban will lift in early April, according to authorities.
As a result, Apple has reportedly decided to cancel the March press conference, which should have taken place on March 31st. That’s what sources familiar with the matter told FrontPageTech’s Jon Prosser. Apple has reportedly considered holding the event elsewhere or announcing the new products via an online-only press event. It’s unclear what will happen, but Apple still plans to launch a few new products in the coming months.
Now, for the really bad news: five separate sources told Prosser that the iPhone 9 won’t be among the products that are launched. Apple will not announce the new iPhone via a streaming event or a press release right now, and it won’t start selling it in early April as it was previously rumored.
I’d imagine it’s difficult for Apple to sell products that it probably doesn’t have in-country. China will have stopped production in the past couple of months. It’ll want to know when the product is ready before it announces it.
unique link to this extract
Mark Bergen and Gerrit De Vynck:
On YouTube, Google’s video service, the company is trying to quickly remove videos claiming to prevent the virus in place of seeking medical treatment. And some apps related to the virus have been banned from the Google Play app store, prompting complaints from developers who say they just want to help. An Iranian government app built to keep track of infections was also removed from the Play Store, ZDNet reported.
The company is also giving up revenue. Pichai said in another recent memo that Google has blocked tens of thousands of ads “capitalizing” on the virus. It’s also pulled ads from YouTube videos that discuss Covid-19, while giving governments and NGOs free ad space on the video service.
“In a highly uncertain, fearful moment there will naturally be more disinformation,” said Thomas Rid, a professor of strategic studies at Johns Hopkins University. “Right now, Google should absolutely emphasize results from the government agencies that can be trusted here, from research-based, evidence-based data.”
Despite such hands-on responses to the virus outbreak, Google stressed that it wasn’t manually changing search results. “Our systems are designed to automatically detect searches that may be related to topics like health and apply the same treatment of elevating reliable and authoritative sources in the results,” a company spokeswoman said.
As one of the first stops for people seeking all kinds of medical information online, the Alphabet Inc. unit has earned the nickname Dr. Google. This has been an uncomfortable role for the company. Medical experts have for years cautioned against going to the search engine for answers. And Google ads have sometimes been abused by unscrupulous businesses looking to take advantage of vulnerable people seeking medical help online.
Isn’t feeding things to not be shown into the AI system that drives search essentially the same thing as manually changing the results? Also, it takes this sort of crisis to make search elevate reliable and authoritative sources?
unique link to this extract
Juliette Kayyem is a former assistant secretary at the Department of Homeland Security:
the officials now urging citizens to keep calm understand far more acutely than the general public how much else can go wrong. A municipal police chief in the Boston area recently urged me to imagine that a school district closed for even three weeks. Take just one child, raised by a single parent who is a police officer. The child is home, so the parent must stay home. Other officers in the same patrol will be affected even if they don’t have kids in school. Shifts will change, nonessential functions will be put off, and the department will have less flexibility to respond to problems unrelated to the epidemic—even as, with more teens unsupervised, rates of car accidents and certain crimes could well increase.
Emergency-response officials are hesitant to play out these dangers in public. This police chief asked me not to identify him because, like so many others in positions of responsibility, he worries that misgivings like his will become self-fulfilling prophecies—that citizens will panic if their local authorities give voice to their own doubts.
Meanwhile, President Donald Trump and his administration have vacillated between ignoring the threat and making wildly unrealistic promises about it. On Wednesday, Vice President Mike Pence promised 1.5 million coronavirus tests, but The Atlantic reported Friday that, according to all available evidence, fewer than 2,000 had been conducted in the United States. Trump himself is simply lying about basic facts about the COVID-19 response; despite the testing kit shortfall, he has publicly stated that everyone who wants to get tested can get tested.
Errata, corrigenda and ai no corrida: none notified