Start Up: Facebook beats privacy, DeepMind’s wrist slapped, the satellite revolution, and more

Face recognition unlocking looks likely to be included in the next iPhone. Photo by nicolasnova on Flickr.

You can now 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 8 links for you. Use them wisely. I’m @charlesarthur on Twitter. Observations and links welcome.

Facebook beats privacy lawsuit in US over user tracking • Yahoo Finance

Jonathan Stempel:


A US judge has dismissed nationwide litigation accusing Facebook Inc of tracking users’ internet activity even after they logged out of the social media website.

In a decision late on Friday, US District Judge Edward Davila in San Jose, California said the plaintiffs failed to show they had a reasonable expectation of privacy, or that they suffered any “realistic” economic harm or loss.

The plaintiffs claimed that Facebook violated federal and California privacy and wiretapping laws by storing cookies on their browsers that tracked when they visited outside websites containing Facebook “like” buttons.

But the judge said the plaintiffs could have taken steps to keep their browsing histories private, and failed to show that Menlo Park, California-based Facebook illegally “intercepted” or eavesdropped on their communications.

“The fact that a user’s web browser automatically sends the same information to both parties,” meaning Facebook and an outside website, “does not establish that one party intercepted the user’s communication with the other,” Davila wrote.


It’s your own fault, and also you should have known about it.
link to this extract

Apple tests 3D face scanning to unlock next iPhone • Bloomberg

Mark Gurman:


Apple Inc. is working on a feature that will let you unlock your iPhone using your face instead of a fingerprint. 

For its redesigned iPhone, set to go on sale later this year, Apple is testing an improved security system that allows users to log in, authenticate payments, and launch secure apps by scanning their face, according to people familiar with the product. This is powered by a new 3-D sensor, added the people, who asked not to be identified discussing technology that’s still in development. The company is also testing eye scanning to augment the system, one of the people said.

The sensor’s speed and accuracy are focal points of the feature. It can scan a user’s face and unlock the iPhone within a few hundred milliseconds, the person said. It is designed to work even if the device is laying flat on a table, rather than just close up to the face. The feature is still being tested and may not appear with the new device. However, the intent is for it to replace the Touch ID fingerprint scanner, according to the person. An Apple spokesman declined to comment.


Google has had face unlocking for a while in Android; it’s quite weird how this article doesn’t mention it, but does mention Samsung’s iris unlock. The 3D scan is going to be fun: will you have to waggle the phone? And – crucially not answered – will there still be TouchID unlock on the front face? If not, what happens to the Apple Pay double-click interaction? Welll…


The new device will have slimmer side bezels around the screen and eliminate the physical home button in favor of a virtual software-based button. Apple has faced challenges integrating the Touch ID fingerprint scanner into this new screen, people familiar with Apple’s work have said. Apple is also testing additional gestures, such as swiping across the center of the screen to launch actions, to replace the home button.


More on this story as it develops.
link to this extract

Jelmer Verhoog on Twitter


@elonmusk Couldn’t wait 4 my #Model3, so made this AR app, what do you think?


This is astonishing – see what your new car will look like in your drive. Shadows too (though they’re not quite congruent with the sun’s position). ARKit is already looking like the most significant thing to happen on iOS for a while – though perhaps it gives good demo. But you can see it for architecture, kitchen/house design, and so on.
link to this extract

The Information Commissioner, the Royal Free, and what we’ve learned • DeepMind

Mustafa Suleyman (co-founder) and Dominic King, clinical lead on Deepmind health:


Today, dozens of people in UK hospitals will die preventably from conditions like sepsis and acute kidney injury (AKI) when their warning signs aren’t picked up and acted on in time. To help address this, we built the Streams app with clinicians at the Royal Free London NHS Foundation Trust, using mobile technology to automatically review test results for serious issues starting with AKI. If one is found, Streams sends a secure smartphone alert to the right clinician, along with information about previous conditions so they can make an immediate diagnosis. 

We’re proud that, within a few weeks of Streams being deployed at the Royal Free, nurses said that it was saving them up to two hours each day, and we’ve already heard examples of patients with serious conditions being seen more quickly thanks to the instant alerts. Because Streams is designed to be ready for more advanced technology in the future, including AI-powered clinical alerts, we hope that it will help bring even more benefits to patients and clinicians in time.

 The Information Commissioner (ICO) has now concluded a year-long investigation that focused on the Royal Free’s clinical testing of Streams in late 2015 and 2016, which was intended to guarantee that the service could be deployed safely at the hospital. The ICO wasn’t satisfied that there was a legal basis for this use of patient data in testing (as the National Data Guardian said too), and raised concerns about how much patients knew about what was happening. The ICO recognised that many of these issues have already been addressed by the Royal Free, and has asked the Trust to sign a formal undertaking to ensure compliance in future…

…Ultimately, if we want to build technology to support a vital social institution like the NHS, then we have to make sure we serve society’s priorities and not outrun them. There’s a fine line between finding exciting new ways to improve care, and moving ahead of patients’ expectations. We know that we fell short at this when our work in health began, and we’ll keep listening and learning about how to get better at this.


DeepMind, as a reminder, is Google’s AI subsidiary – a British company based in King’s Cross, London. This is quite a mea culpa. (Note too how it fits into the Silicon Valley paradigm: better to ask for forgiveness than for permission.)

The ICO report begins bluntly: “The ICO has ruled the Royal Free NHS Foundation Trust failed to comply with the Data Protection Act when it provided patient details to Google DeepMind.”
link to this extract

How to build a simple neural network in 9 lines of Python code • Medium

Milo Spencer-Harper:


As part of my quest to learn about AI, I set myself the goal of building a simple neural network in Python. To ensure I truly understand it, I had to build it from scratch without using a neural network library. Thanks to an excellent blog post by Andrew Trask I achieved my goal. Here it is in just 9 lines of code.


And yes, it is just nine lines. The explanation is longer, but if you’re looking for a primer on how to build a simple neural network, this is it.
link to this extract

The time I got recruited to collude with the Russians • Lawfare

Matt Tait, who isn’t an American nor based in the US but who is a security expert who looked at public data about Hillary Clinton’s emails to analyse whether the Russians might have hacked her private email server, was contacted last summer by a someone claiming to be well-connected with the Trump campaign:


Towards the end of one of our conversations, [Republican activist Peter] Smith made his pitch. He said that his team had been contacted by someone on the “dark web”; that this person had the emails from Hillary Clinton’s private email server (which she had subsequently deleted), and that Smith wanted to establish if the emails were genuine. If so, he wanted to ensure that they became public prior to the election. What he wanted from me was to determine if the emails were genuine or not.

It is no overstatement to say that my conversations with Smith shocked me. Given the amount of media attention given at the time to the likely involvement of the Russian government in the DNC hack, it seemed mind-boggling for the Trump campaign—or for this offshoot of it—to be actively seeking those emails. To me this felt really wrong.

In my conversations with Smith and his colleague, I tried to stress this point: if this dark web contact is a front for the Russian government, you really don’t want to play this game. But they were not discouraged. They appeared to be convinced of the need to obtain Clinton’s private emails and make them public, and they had a reckless lack of interest in whether the emails came from a Russian cut-out. Indeed, they made it quite clear to me that it made no difference to them who hacked the emails or why they did so, only that the emails be found and made public before the election.


Some of the detail is quite eye-opening: the company to do this is closely associated with people you’ll have heard of, which tried to set up as a Delaware company to avoid being linked to campaign financing rules. Tait has been quoted in the WSJ in stories about how Smith was in contact with Mike Flynn.

Short version: if there’s a Russian operative on the other end of the “dark web” contact, Smith is in deep trouble, and so is Mike Flynn – and possibly some others close to Trump. The question is whether the FBI or others can pin that down. Of course, this article is how Tait is inviting the special prosecutor Robert Mueller to interview him.
link to this extract

The tiny satellites ushering in the new space revolution • Bloomberg

Ashlee Vance:


Once aligned, Planet’s attitude and determination control system, which sets the satellite’s orientation, takes over. Gyroscopes and sensors on the Dove [satellite, about the size of a shoebox] look for magnetic fields and seek out the Earth’s horizon, the sun, and other stars. Magnetorquers and reaction wheels then adjust the satellite’s movement until it reaches the desired alignment. “It’s not that difficult to make a system that does this,” says Ben Howard, Planet’s chief spacecraft architect. “It’s difficult to make it as cheaply as we have and to make it tuned so well for a specific application.”

Each Dove is responsible for collecting 10,000 images covering 2 million square kilometers per day, an area the size of Mexico. The pictures—40 gigabytes’ worth—are relayed during 10 daily eight-minute sessions on custom-built radios between the satellites and a dozen ground stations built by Planet in Antarctica, Chile, Hawaii, Iceland, and other places.

Once the images reach Earth, Planet’s software compiles them, cleans them up, and deletes photos marred by clouds and shadows. Customers can then log on to an application and browse the pictures as they please. Planet’s largest clients include the Mexican government, the German space agency, and the agricultural companies Monsanto, Wilbur-Ellis, and Bayer Crop Science. They pay millions or even tens of millions of dollars per year for access to the most recent, highest-quality images. Nonprofits, students, and news organizations receive the same access for free, while the public at large can see older, lower-quality pictures gratis. Planet refuses to say how much revenue it draws, but it appears to be enough to keep investors interested. The company has raised more than $180m in venture capital to date, and its valuation has been widely reported to exceed $1bn.

Planet’s 88 new satellites, which will give it the only daily view of Earth, at least for now, promise to be even better for the bottom line.


The examples of what can be done with systems like this are remarkable. For example:


Crawford can call up an image of a port in Shanghai that’s been broken down like a puzzle, with cylindrical oil storage tanks color-coded green, ships in red, and buildings in blue. Hit a button, and the software shows that eight new buildings have gone up in a few months. Hit another button, and the software will calculate how much oil is in a given tank. “There are floating lids that sit on top of the tanks,” Crawford says. “If the lid is all the way up, there’s no shadow, and we know it’s full.” If there’s a shadow, Orbital Insight measures its angle and the dimensions of the tank to calculate the volume of liquid inside. What Crawford’s company is after, he says, is “observational truth.”


link to this extract

Demographics of mobile device ownership and adoption in the US • Pew Research Center


A substantial majority of Americans are cellphone owners across a wide range of demographic groups. By contrast, smartphone ownership exhibits greater variation based on age, household income and educational attainment.


The demographics are the second group here – you have to scroll down the page past the smartphone ownership growth data. Notable falloff among those over 65, those who didn’t graduate from high school (equivalent to secondary school in UK, ie like leaving school after GCSE), those earning under $30,000pa. The latter two – or all three – groups might intersect substantially. Rural ownership is also comparatively low.
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 )

Google+ photo

You are commenting using your Google+ 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.