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 13 links for you. Or 1101 in binary. I’m @charlesarthur on Twitter. Observations and links welcome.
After [the former DARPA Grand Challenge for self-driving vehicles participant, Anthony] Levandowski arrived at Google, his plan was to send out hundreds of cars, equipped with cameras, to photograph America’s roads. Then he encountered Google’s bureaucracy.
The company was less than a decade old, but it had almost seventeen thousand employees, including a thick layer of middle managers. Levandowski recently told me, “One of the reasons they wanted us was because Larry Page knew we were scrappy—we would cut through red tape.” Page, Google’s co-founder and chief executive, often complained that the company had become bloated, and had lost the hacker mentality that had fuelled its initial success. By the time Levandowski arrived, Google’s apparatchiks were in ascent.
“Hiring could take months,” Levandowski told me. “There was a program called WorkforceLogic, and just getting people into the system was super-complicated. And so, one day, I put ads on Craigslist looking for drivers, and basically hired anyone who seemed competent, and then paid them out of my own pocket. It became known as AnthonyforceLogic.” Around this time, Levandowski went to an auto dealership and bought more than a hundred cars. One of his managers from that period told me, “When we got his expense report, it was equal to something like all the travel expenses of every other Google employee in his division combined. The accountants were, like, ‘What the hell?’ But Larry said, ‘Pay it,’ and so we did. Larry wanted people who could ignore obstacles and could show everyone that you could do something that seemed impossible if you looked for work-arounds.”
Levandowski and his team were asked to map a million miles of U.S. roads within a year. They finished in nine months, and then set up an enormous office in Hyderabad, India, to begin mapping every street on earth.
This isn’t the heart of the story – this is back in 2007 – but it illustrates something pertinent about both Levandowski and Page, particularly the latter: he’ll forgive if you get the results.
It also goes into Silicon Valley’s culture, which it says is built on one big idea: betrayal.
link to this extract
Apple apologized over the hacking of some Chinese accounts in phishing scams, almost a week after it emerged that stolen Apple IDs had been used to swipe customer funds.
In its English statement Tuesday, Apple said it found “a small number of our users’ accounts” had been accessed through phishing scams. “We are deeply apologetic about the inconvenience caused to our customers by these phishing scams,” Apple said in its Chinese statement.
The incident came to light last week when Chinese mobile-payment giants Alipay and WeChat Pay said some customers had lost money.
The victims of the scams, Apple said Tuesday, hadn’t enabled so-called two-factor authentication—a setting that requires a user to log in with a password and a freshly-generated code to verify their identity.
The Cupertino, Calif.-based company didn’t specify how many users were hit or how much money was stolen, nor did it offer details about how the hackers acquired the users’ Apple IDs and passwords. To help prevent unauthorized access to their accounts, Apple said, people should enable two-factor authentication.
It was a pretty safe bet that the people who got phished hadn’t enabled 2FA. (And that it was phishing rather than hacking.) Strange, since Apple pushes a reminder in the Settings app. This is interesting PR, though: apologising for something the customer got wrong and that Apple couldn’t control.
link to this extract
Google CEO Sundar Pichai says Project Dragonfly, the censored Chinese search engine, works • The Washington Post
“If Google were to operate in China, what would it look like? What queries will we be able to serve?” chief executive Sundar Pichai said during an event hosted by Wired on Monday night. “It turns out we’ll be able to serve well over 99% of the queries.”
The announcement could prompt more questions from U.S. policymakers, some of whom have accused Google of being evasive about Project Dragonfly. Meanwhile, Google and its peers in the tech industry are facing intense scrutiny over its approach to user privacy and data, with some federal lawmakers proposing legislation that could impose new restrictions on tech companies’ handling of customer information.
Like many other firms, Google is eyeing China as a massive market opportunity. China, which has an estimated population of 1.4 billion, is already heavily dependent on Google’s Android operating system; in 2013, 9 out of 10 smartphones in China were running Android. But Google’s position in mobile could eventually erode as Chinese competitors have sought to develop alternatives to Android. Gaining broader access to Chinese audiences could give Google more opportunities to serve online advertising and sell mobile apps.
As many have noted, the hardware is still extremely limiting. The technology underpinning these experiences seems genuinely advanced, and if it were not for a multi-year blitzkrieg marketing campaign insisting a reality where pixels blend seamlessly with IRL physics was imminent, it might have felt truly impressive. (Whether or not it’s advanced enough to eventually give rise to Leap’s prior promises is an entirely open question at this point.) For now, the field of vision is fairly small and unwieldy, so images are constantly vanishing from view as you look around. If you get too close to them, objects will get chopped up or move awkwardly. And if you do get a good view, some objects appear low res and transparent; some looked like cheap holograms from an old sci-fi film. Text was bleary and often doubled up in layers that made it hard to read, and white screens looked harsh—I loaded Google on the Helio browser and immediately had to shut my eyes.
According to Magic Leap, over 1,000 people had signed up to be here. Why?, I wanted to ask all of them at once. Do you think this is the future? Do you really?
I’ll reiterate my prediction that pretty soon Magic Leap will pivot to industrial applications, which might exist.
link to this extract
It turns out that Facebook could in fact use data collected from its Portal in-home video device to target you with ads • Recode
Last Monday, we wrote: “No data collected through Portal — even call log data or app usage data, like the fact that you listened to Spotify — will be used to target users with ads on Facebook.”
We wrote that because that’s what we were told by Facebook executives.
But Facebook has since reached out to change its answer: Portal doesn’t have ads, but data about who you call and data about which apps you use on Portal can be used to target you with ads on other Facebook-owned properties.
“Portal voice calling is built on the Messenger infrastructure, so when you make a video call on Portal, we collect the same types of information (i.e. usage data such as length of calls, frequency of calls) that we collect on other Messenger-enabled devices. We may use this information to inform the ads we show you across our platforms. Other general usage data, such as aggregate usage of apps, etc., may also feed into the information that we use to serve ads,” a spokesperson said in an email to Recode.
That isn’t very surprising, considering Facebook’s business model. The biggest benefit of Facebook owning a device in your home is that it provides the company with another data stream for its ad-targeting business.
I’m shocked, shocked to learn that data collection for targeting ads is going on in this Facebook device.
link to this extract
Warden used to be chief technology officer for a company called Jetpac, which used neural networks to do interesting stuff with Instagram photos; then Google bought it, and he’s working on machine learning there:
One of the other reasons I think ML is such a good fit for compression is how many interesting results we’ve had recently with natural language. If you squint, you can see captioning as a way of radically compressing an image. One of the projects I’ve long wanted to create is a camera that runs captioning at one frame per second, and then writes each one out as a series of lines in a log file. That would create a very simplistic story of what the camera sees over time, I think of it as a narrative sensor.
The reason I think of this as compression is that you can then apply a generative neural network to each caption to recreate images. The images won’t be literal matches to the inputs, but they should carry the same meaning. If you want results that are closer to the originals, you can also look at stylization, for example to create a line drawing of each scene. What these techniques have in common is that they identify parts of the input that are most important to us as people, and ignore the rest.
It’s not just images.
There’s a similar trend in the speech world. Voice recognition is improving rapidly, and so is the ability to synthesize speech. Recognition can be seen as the process of compressing audio into natural language text, and synthesis as the reverse. You could imagine being able to highly compress conversations down to transmitting written representations rather than audio. I can’t imagine a need to go that far, but it does seem likely that we’ll be able to achieve much better quality and lower bandwidth by exploiting our new understanding of the patterns in speech.
With more than 80% of the world’s smartphones running on the Android operating system, the product is vital to Google’s future revenues and profitability.
Google denied any wrongdoing and has appealed against the EU’s decision to the European Court of Justice. But on Tuesday a company spokesperson said that from October 29, Android phonemakers “wishing to distribute Google apps” would also be able to build “non-compatible, or forked, smartphones and tablets for the EEA”.
The spokesperson added that phonemakers would also be able to able to license Google Play separately from Google’s search engine and Chrome for an unspecified fee.
With Tuesday’s announcement, Google addressed each of the practices that Ms Vestager deemed illegal. However, critics say the changes are unlikely to upend the global smartphone industry.
Thomas Vinje, a lawyer at Clifford Chance whose clients have raised competition concerns over Google’s Android contracts, said: “The bottom line is that Google’s so-called remedies would mean that both Android and Google’s other dominant mobile products will remain immune from effective competition.
“No manufacturer will produce a device based on a forked version of Android only for Europe,” he added.
Vinje is probably correct.
link to this extract
Five ways Google Pixel 3 camera pushes the boundaries of computational photography • Digital Photography Review
With the launch of the Google Pixel 3, smartphone cameras have taken yet another leap in capability. I had the opportunity to sit down with Isaac Reynolds, Product Manager for Camera on Pixel, and Marc Levoy, Distinguished Engineer and Computational Photography Lead at Google, to learn more about the technology behind the new camera in the Pixel 3.
One of the first things you might notice about the Pixel 3 is the single rear camera. At a time when we’re seeing companies add dual, triple, even quad-camera setups, one main camera seems at first an odd choice.
But after speaking to Marc and Isaac I think that the Pixel camera team is taking the correct approach – at least for now. Any technology that makes a single camera better will make multiple cameras in future models that much better, and we’ve seen in the past that a single camera approach can outperform a dual camera approach in Portrait Mode, particularly when the telephoto camera module has a smaller sensor and slower lens, or lacks reliable autofocus [like the Galaxy S9].
This isn’t actually a test of the Pixel 3. Plenty of interesting things here; will they come to the wider range of Android, though? The Pixel is a fraction of a fraction of Android sales.
We’re also approaching the point where it’s only the low-light pictures that show substantial differences between generations. (Thanks stormyparis for the link.)
link to this extract
The world is on fire but the new Google Pixel 3 — a Good Phone, which I do recommend you buy if you like Android and can afford it, although its updates are mostly incremental — in my pocket is cool to the touch. A dark slab of metal and glass. It comes alive when I rub my finger across the back of it.
“We’re doomed,” a colleague texts me on Signal*. A push alert from a well-regarded news site has more details on the alleged murder and dismemberment of a Saudi journalist. On Nextdoor, several neighbors report that their drinking water has tested positive for unsafe levels of pesticides. The Citizen app prompts me to record video of an angry naked man rampaging in the shit-strewn streets of San Francisco. Facebook is hacked and our information is out there. Everyone on Twitter is angry, you fucking cuck. You idiot. You tender, triggered snowflake. Everyone on Instagram is posturing, posing. You are less beautiful than they. The places you go are not as interesting. You should feel bad because you are worse in every way. The world is dying; come see it, come see it.
I don’t recall exactly when my phone became such a festival of stress and psychological trauma, but here we are.
If you haven’t read – or had forgotten – Honan’s piece from CES Las Vegas, called “Fever Dream of a Guilt-Ridden Gadget Reporter“, it’s time to enjoy that too. Sample paragraph:
I try to remember all the products I’ve talked about that I won’t even bother to cover—and that nobody’s going to buy. There were some Bluetooth speakers. Or maybe they were WiFi. But there was definitely a helmet cam. And a waterproof phone. And a tablet and an ultrabook and an OLED TV. There was ennui upon ennui upon ennui set in this amazing temple to technology.
That was January 2012. Never change, Mat.
link to this extract
Turkey releases passport scans of men it says were involved in journalist’s killing • Washington Post
Turkish officials have provided The Washington Post with scans of passports that they say were carried by seven men who were part of a Saudi team involved in the killing of journalist Jamal Khashoggi inside the Saudi Consulate in Istanbul on Oct. 2.
These passport scans add to the information made public by Turkey as it seeks to fill out the narrative of what happened to Khashoggi, a Post contributor who vanished after entering the consulate to obtain a document he needed for his upcoming wedding.
The Post is publishing the passport scans but obscuring the faces and names of the men because it has not independently verified their identities.
Within days of Khashoggi’s disappearance, Turkish investigators said they had pieced together most of the mystery, concluding that he had been killed inside the consulate and dismembered.
Turkey said a 15-member team dispatched from Saudi Arabia played a role in the killing. Turkish officials have confirmed that the 15 names reported in the Turkish media are those of the suspected team members, and their alleged involvement is part of the evidence cited by Turkey that Saudi Arabia was responsible for Khashoggi’s death.
Turkey’s playing an interesting game here. “Sources close to the investigation” have also released security camera footage, which claims to show a big people carrier with blacked-out windows leaving the consulate and then arriving at the consul’s home.
Turkey knows it can make Saudi Arabia uncomfortable, and embarrass the US if Trump says it’s fine, and then it releases video or audio. Saudi Arabia knows this; the US knows it. Turkey can keep dripping out this stuff for ages, to keep the story in the headlines.
So what does Turkey want in exchange for not doing this? Something political, of course. But what?
link to this extract
With the benefit of hindsight, however, there is one aspect of the company’s figures which looks odd: average sales per store barely changed in five years, even as the number of them doubled. Expansion, the addition of different brands, economic vagaries – through it all a Patisserie Valerie cafe took sales of about £600k a year.
In the year to September 2014, when there were 128 stores on average, each contributed revenues of £598k. Last year, 192 stores contributed an average £596k each.
Here’s the progression of sales, to £114m last year:
And here’s the average revenue per store, as the group’s total number of sites went from 89 to 206:
The metric was remarkably stable, suspiciously so we might now say. Business is rarely that smooth, as weather, the ebb and flow of competition, and even politics (a Brexit effect?) play a role.
This is part of the “Someone is wrong on the internet” series – a series title too wonderful for words. Patisserie Valerie is a chain of retail cake shops (so, as the story says, pretty much zero inventory) which a week ago discovered it has £20m less than it thought.
link to this extract
July 2017: Russian national and bitcoin exchange charged in 21-count indictment for operating alleged international money laundering scheme and allegedly laundering funds from hack of Mt. Gox • USAO-NDCA | Department of Justice
A grand jury in the Northern District of California has indicted a Russian national and an organization he allegedly operated, BTC-e, for operating an unlicensed money service business, money laundering, and related crimes…
…“Mr. Vinnik is alleged to have committed and facilitated a wide range of crimes that go far beyond the lack of regulation of the bitcoin exchange he operated. Through his actions, it is alleged that he stole identities, facilitated drug trafficking, and helped to launder criminal proceeds from syndicates around the world,” said Chief Don Fort, IRS Criminal Investigation. “Exchanges like this are not only illegal, but they are a breeding ground for stolen identity refund fraud schemes and other types of tax fraud. When there is no regulation and criminals are left unchecked, this scenario is all too common. The takedown of this large virtual currency exchange should send a strong message to cyber-criminals and other unregulated exchanges across the globe.”
“BTC-e was noted for its role in numerous ransomware and other cyber-criminal activity; its take-down is a significant accomplishment, and should serve as a reminder of our global reach in combating transnational cyber crime,” said Special Agent in Charge of the USSS Criminal Investigative Division Michael D’Ambrosio. “We are grateful for the efforts of our law enforcement partners in achieving this significant result.”
“The arrest of Alexander Vinnik is the result of a multi-national effort and clearly displays the benefits of global cooperation among US and international law enforcement,” said FBI Special Agent in Charge Hess.
OK, so that was more than a year ago. But you can bet that if there’s money laundering on one bitcoin exchange, then given how many there are around, it will be happening on others. Which brings us to…
link to this extract
The white paper that heralded Tether’s creation explicitly calls for regular audits. Without them, anyone buying Tether is effectively operating on faith. Think about it: you can barely rent an apartment without going through a credit check and proving you can cover the cost. You’d think the market would demand some concrete assurances about the issuance of $2.7bn worth of currency.
Let’s assume, though, that Tether really does have $2.7bn sitting in a safe somewhere. Where did it all come from? The most innocent answer is that some deep-pocketed investors decided they wanted to invest in cryptocurrency, but rather than simply buy some with dollars, they instead opted to buy Tether first and then use that to purchase the crypto.
Just why anyone would do that remains unclear, especially since, as UC Berkeley computer science researcher Nicholas Weaver has pointed out on Lawfareblog.com, “[O]ne has to believe that they did this even though these unregulated exchanges have a history of getting hacked, with customers losing their investments.”
A less innocent answer is that the investors couldn’t go to a banked exchange because their funds came from illegal activity, so they used Tether to turn their ill-gotten gains into untraceable crypto loot. In other words, money laundering.
Perhaps the most troubling answer for crypto investors is that Tether minted currency out of thin air, used it to buy other cryptocurrency, sold that cryptocurrency, and used the proceeds to create its reserves. That is, assuming the reserves actually exist at all.
In a sense, though, it doesn’t matter whether the money is in the bank or not. Tether’s terms of service state, “We do not guarantee any right of redemption or exchange of tethers by us for money.” Even if the money is in the vault, Tether holders have no claim to it.
Increasingly I suspect that Tether/Bitfinex’s official location in Panama means that it is a gigantic money laundering operation for, eh, shall we say drug cartel money? This would explain its occasional gigantic wafts of money, and its desperate search for a bank that will actually hold its reserves. And why it persists.
link to this extract
Errata, corrigenda and ai no corrida: none notified