In London on October 18th? I’ll be giving a talk: “Social Networks and the Truth“:
how social media is both polarising our opinion and weakening the ways in which we hear contrary views.
Think about the claims made around Brexit (“£350m per week to go back to the NHS!”), Donald Trump’s ability to make wild claims, believed by his followers without question, and the difficulty of getting anyone to agree even on what seem like simple facts – the disappearance of MH370, 9/11, climate change, the list goes on.
So how did we get here? And what will happen next?
Tickets are limited; book now. £10 secures your place.
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 14 links for you. Not flammable. I’m charlesarthur on Twitter. Observations and links welcome.
Southwest Airlines flight 994 from Louisville to Baltimore was evacuated this morning while still at the gate because of a smoking Samsung Galaxy Note 7 smartphone. All passengers and crew exited the plane via the main cabin door and no injuries were reported, a Southwest Airlines spokesperson told The Verge.
More worryingly, the phone in question was a replacement Galaxy Note 7, one that was deemed to be safe by Samsung. The Verge spoke to Brian Green, owner of the Note 7, on the phone earlier today and he confirmed that he had picked up the new phone at an AT&T store on September 21st. A photograph of the box shows the black square symbol that indicates a replacement Note 7 and Green said it had a green battery icon.
Green said that he had powered down the phone as requested by the flight crew and put it in his pocket when it began smoking. He dropped it on the floor of the plane and a “thick grey-green angry smoke” was pouring out of the device. Green’s colleague went back onto the plane to retrieve some personal belongings and said that the phone had burned through the carpet and scorched the subfloor of the plane.
Credit to Golson, who actually got hold of the photo and checked it. Hurrah for real journalism. Meanwhile, Samsung suddenly has a much bigger problem than it had a few days ago, in its second-biggest market.
link to this extract
First, we needed a test subject. We planned to buy a small app with no active marketing activities, but significant and steady download numbers. Then we’d increase the app’s size, leaving everything else constant, and observe the impact on the app’s install rate. This would simulate the impact of app bloat on downloads.
So we bought the Mortgage Calculator Free iOS app. It was a minuscule 3MB, with a steady pattern of organic installs (about 50 installs per day for several years) and no active marketing activities. It was the perfect test case.
Over the course of the experiment, we increased the app size from 3MB to 99MB, 123MB and 150MB. We kept everything else constant to observe the isolated impact on install rate with each change in app size.
App sizes can increase substantially with the addition of seemingly simple things, like an explainer video, a bunch of fonts, an SDK or a background picture for your loading screen. For the purposes of our experiment, we bloated our app with a ton of hidden Taylor Swift album art.
To measure the impact of each successive bloating, we looked at data provided directly by Apple in iTunes Analytics. We specifically tracked conversion from “Product Page Views” to “App Units,” better known as “installs” to ”install rate.”
Hit 100MB and you doom your installs. (That’s a limit Apple sets for Wi-Fi only downloads.) But there’s more.
link to this extract
Back in 2013, a handful of researchers at Google set loose a neural network on a corpus of three million words taken from Google News texts. The neural net’s goal was to look for patterns in the way words appear next to each other.
What it found was complex but the Google team discovered it could represent these patterns using vectors in a vector space with some 300 dimensions.
It turned out that words with similar meanings occupied similar parts of this vector space. And the relationships between words could be captured by simple vector algebra. For example, “man is to king as woman is to queen” or, using the common notation, “man : king :: woman : queen.” Other relationships quickly emerged too such as “sister : woman :: brother : man,” and so on. These relationships are known as word embeddings.
This data set is called Word2vec and is hugely powerful. Numerous researchers have begun to use it to better understand everything from machine translation to intelligent web searching.
But today Tolga Bolukbasi at Boston University and a few pals from Microsoft Research say there is a problem with this database: it is blatantly sexist.
This is a remarkable study (and the de-sexisation of the corpus is even more impressive); it does make one wonder the extent to which news headlines continue ages-old tropes. (Here’s the original paper.)
link to this extract
UBS analysts think this means there’s a lot of upside to Apple stock that investors aren’t factoring in. Sure, the iPhone is a huge hit and a commercial success, but they see Apple laying the groundwork for “the next era of personal technology — the Ambient Paradigm.”
“We consider the Apple Watch and AirPods similar transition products today on the way to an integrated user experience based on multiple products seamlessly connected. We call it the Ambient (present on all sides) Paradigm. It is Tim Cook’s ‘iOS everywhere,'” Milunovich wrote.
He sees Apple’s “other products” like the Apple Watch and AirPods evolving in coming years, with Siri acting as the glue, and potentially affecting industries like healthcare and education.
If you’re constantly surrounded by Apple products and services, that presents a huge revenue opportunity for the computer maker, and also increases the possibility that users will feel locked into Apple’s ecosystem.
The ruling, issued on Friday by the U.S. Court of Appeals for the Federal Circuit, found that three patents asserted against anti-virus companies Symantec SYMC 0.63% and Trend Micro were invalid because they did not describe a patentable invention. The patents were owned by Intellectual Ventures, which has a notorious reputation in the tech world as a so-called “patent troll,” a phrase that describes firms that buy up old patents and wage lawsuits in order to demand payments from productive companies.
The most important part of the decision, which has created a stir among the patent bar, is a concurrence by Circuit Judge Haldane Mayer. In striking down a key claim from U.S. Patent 5987610, which claims a monopoly on using anti-virus tools within a phone network, Mayer says it is time to acknowledge that a famous Supreme Court 2014 decision known as “Alice” basically ended software patents altogether.
There are many other ways to interpret this story. For example, the government may simply be demanding that when Yahoo satisfies demands for emails (based on email addresses), that it does so from the raw incoming stream, before it hits spam/malware filters. Or, they may be demanding that Yahoo satisfies their demands with more secrecy, so that the entire company doesn’t learn of the email addresses that a FISA order demands. Or, the government may be demanding that the normal collection happen in real time, in the seconds that emails arrive, instead of minutes later.
Or maybe this isn’t an NSA/FISA story at all. Maybe the DHS has a cybersecurity information sharing program that distributes IoCs (indicators of compromise) to companies under NDA. Because it’s a separate program under NDA, Yahoo would need to setup a email malware scanning system separate from their existing malware system in order to use those IoCs. (@declanm’s stream has further variations on this scenario).
My point is this: the story is full of mangled details that really tell us nothing. I can come up with multiple, unrelated scenarios that are consistent with the content in the story. The story certainly doesn’t say that Yahoo did anything wrong, or that the government is doing anything wrong (at least, wronger than we already know).
Declan McCullagh offers a scenario where the Department of Homeland Security wanted to pick out emails which had particular malware attachments (foreign spearphishing attacks?), and Yahoo’s legal team threw together an engineering team but couldn’t clear it with the Yahoo security team. And now Yahoo can’t correct the reporting because it’s all classified.
link to this extract
In 1939 there were about 140,000 Jews living in the Netherlands, including about 25,000 German Jews who had fled from Germany. By 1945, only about 35,000 of these people were alive. The Nazi extermination of Dutch Jews was remarkably efficient, mainly because Holland had been a well-administered state which kept very good records of its citizens, their addresses and their religions. So when the Nazis arrived, their genocidal task was easier than it was in some other occupied countries.
This horrific story neatly encapsulates the dilemma of the data-driven state. On the one hand, good governance requires that a state knows a lot about its citizens — where they live, what they do for a living, what taxes they pay, which schools their children can attend, and so on. Since 9/11, Western democracies have determined that the ‘war’ on terror (or the need to keep us safe, depending on your point of view) requires that the state needs to know an awful lot more about its citizens, and so comprehensive surveillance of their online and mobile communications, movements and financial transactions has been added to the government’s shopping list.
As we know from Edward Snowden and other sources, the scale and intrusiveness of this surveillance is now staggering. And — as the UK Investigatory Powers Bill shows, the state’s appetite for fine-grained personal data seems insatiable and is destined to grow.
Imagine if Yahoo had operated in Nazi Germany. Or if a fascist was elected to run the US. One would require time travel. The other..
link to this extract
The gimbal, or camera-stabilizing device, faced production delays and the first Solos hit the market [in June 2015] without this add-on, making it unsuitable for photos and video, the chief use of most consumer drones. “Making the gimbal was harder than making the drone,” said Guinn, who noted that the devices didn’t get to customers until August, a full two months after Solo’s launch.
Still, 3D Robotics executives remained bullish on Solo’s potential, forecasting huge sales for the holiday season. According to one employee, CFO John Rex and Anderson, who had already committed to make 60,000 of the quadcopters with contract manufacturer PCH International, decided in mid-June with less than a month of sales data that an additional 40,000 devices should be built. That represented a significant commitment, said another person who helped engineer Solo, because each drone and its gimbal cost more than $750 to manufacture and ship to retailers. Though the company was able to raise $64m in 2015, most of that was sunk into manufacturing costs, sources told Torbes.
Multiple people blamed the 3D Robotics’ bold projection for Solo’s failure, including one former employee who said that the fatal mistake was in basing predictions off of “sell in” versus “sell through” figures. The company forecasted Solo sales erroneously based on the inventory it was distributing to retail channels like Best Buy–a poor indicator of consumer demand because retailers can send back unsold inventory–and not on the number of devices actually purchased by customers from those stores.
A person, who worked for 3D Robotics’ marketing team, also questioned the company’s practices when displaying the drone to the press. The demo with The Verge in the spring of 2015, for example, featured a drone that was “worked over and souped up” and did not feature the typical parts you’d find in an off-the-shelf Solo. “We knew the drone would work,” he said, noting that there was an improved GPS component that wasn’t shipped in regulars Solos.
Showing hyped-up designs to visiting journalists (and others) is a common ploy. Shop-bought ones, now, that’s a different thing. (Casey Newton, who was the person it was demonstrated to, is now mad as hell – but also wiser in the ways of the world.)
link to this extract
I gave Daydream a go at a special launch event in London on Tuesday — but it failed to blow me away.
The best virtual reality experiences have an element of transcendence to them. It’s the moment when the fact that you’re sitting there with a screen awkwardly strapped to your head just melts away, and you really feel there. It doesn’t matter if the graphics are cartoony, or if the screen is a bit pixellated — all of a sudden, you’re transported into an entirely new world. Done right, it can be magical.
At no point using Daydream did I feel this. I got the chance to try out two demos — an “experience” based around the forthcoming “Fantastic Beasts and Where to Find Them” JK Rowling movie, and a stylised game that has you roll a ball around a race track by tilting the remote.
Both times, it felt like I was using Cardboard. Sure, it was clearly more polished than the DIY headset, with less lag and a great new input device — but the underlying experience was similar. Head movements felt unnatural, the images didn’t feel “real”: There was no transcendence. I came away with a headache.
I was, inescapably, sitting in a chair with a smartphone stuck to my face.
Daydream, is evolutionary, not revolutionary — and that’s a problem
Mobile-only virtual reality simply isn’t ready for prime time. Alexander Koerner/Getty Images
Daydream is an evolutionary upgrade to Cardboard, rather than anything revolutionary. And I came away feeling like the technology behind mobile-powered virtual reality just isn’t there yet.
On Tuesday, as a Google event began in San Francisco, new hardware chief Rick Osterloh [who used to be the hardware chief when Google owned Motorola] reassured the audience that the company was serious about the move. “We’re in it for the long run,” he said.
About nine months after initial talks with carriers, Google rolled out its flavor of technical support: Pixels have built-in chat support where customer-service reps can take over smartphone screens to identify problems.
Google will be responsible for returns and recycling and is building a supply chain that can re-absorb faulty and rejected devices, Osterloh said in a recent interview.
Osterloh will be partly judged by how many devices Google sells, a contrast to the Nexus program which showcased Android features for other handset makers to adopt. But the executive was still cagey about the company’s sales aspirations.
“In markets where we do business, we’re definitely going to want meaningful share,” he said. “But it’s highly unlikely that the primary driver will be to be in every market with as high as possible volume.”
That seems to imply the US and some bits of Europe as “markets where [Google does] business”. Those markets are stagnant. Quite how Google expects to get “meaningful” share (whatever that means), I don’t know. But Osterloh has a long history of blinding journalists with vague words. When Motorola was losing money hand over fist, he’d insist that “we make money on every phone we sell”. He meant gross margin – the sale price compared to raw cost of goods – and handily omitted everything else, such as marketing, administration, R D, and so on.
link to this extract
perhaps the most fascinating implication of Spectacles is what it says about the potential of a long-term rivalry between Snapchat and Apple. Snapchat CEO Evan Spiegel has said that Snap née Snapchat is a camera company, not a social network. Or, perhaps more accurately, the company is both: it is a fully contained ecosystem that is more perfectly optimized for the continual creation and circulation of content than even Facebook. What matters from Apple’s perspective is that Snapchat, like Facebook or WeChat or other apps that users live in, is one layer closer to their customers. For now that is not a threat — you still need an actual device to run those apps — but then again most people used Google on Windows, which made Microsoft a lot of money even as it froze them out of the future.
This is exactly why Apple is right to push forward into the wearable space even though it is an area, thanks to the important role of services like Siri, in which they have less of an advantage. Modern moats are not about controlling distribution but about owning consumer touch points — in the case of wearables, quite literally.
I’ve linked to this a little after its publication, but the general point Thompson is making – that groundbreaking products need to be able to slot into ecosystems around them – is crucial.
link to this extract
Google’s Pixel smartphones target the most profitable segment, hurting Google’s partners • IHS Technology
Google is pursuing a similar integrated hardware-software strategy to Apple with Pixel smartphones, Daydream View, and the other new hardware Google has announced. This is the final defeat for the operating system licensing model which Microsoft pioneered, and everyone tried to copy before Apple’s iPhone success.
But Google’s culture and deep learning, intelligence, organizational, and software competitive strengths are very different to Apple. Yet, with Pixel smartphones Google is aiming at the same competitive areas which Samsung and Apple are: camera quality, cloud storage, and the ease of experience. Google needs to differentiate based on its competitive strengths around AI, but Google Assistant needs to be on as many smartphones as possible to support Google strategy and so cannot be a long term differentiator for Pixel smartphones.
Google still has many smartphone hardware partners, unlike Apple, and it continues to need them. Because if not, Samsung may ramp its fall-back TIzen OS strategy, and more significantly Google’s many China headquartered smartphone maker partners may fork Android and take their more proprietary Chinese Android variants into international markets.
Android may be dominant now, but it’s not invincible if Google makes the wrong strategic moves and undermines its ecosystem partners.
“It’s still early days, but when all of that works together, the Google Assistant allows you to get things done, bringing you the information you need, when you need it, wherever you are,” writes Pichai, in a caveated phrase that scores extremely high on the underwhelming / vague promises index.
He adds that he has “confidence” of being able to “do some amazing things for users over the next 10 years”.
So, in other words, trust us with all your data!
This week the EFF also excoriated Google for how AI is impinging on user privacy, focusing on another of its recent products: the Allo messaging app. That app also bakes in Google Assistant, and because Allo does AI by default the app does not offer end-to-end encryption by default — only as an ‘optional extra’ — because of course Google’s AI can’t function when Google’s AI can’t read your messages…
Criticizing the way Allo silos end-to-end encryption within an ‘incognito’ mode, which the EFF argues risks confusing users and risks sensitive data leaking out, it accuses Google of “training users to use encryption as an occasional measure” — going on on to conclude that: “A more responsible messaging app would make security and privacy — not machine learning and AI — the default.”
So whether it’s Google Home or Google Allo, Google is promising consumers a magical, AI-powered experience of unrivaled convenience. But it pays to ask tougher questions.
The adtech giant is trying to control the narrative, just as it controls the product experience. So while Google’s CEO talks only about the “amazing things” coming down the pipe in a world where everyone trusts Google with all their data — failing entirely to concede the Big Brother aspect of surveillance-powered AIs — Google’s products are similarly disingenuous; in that they are designed to nudge users to share more and think less.
Very nifty: postcode-based search if you want it (and if your broadband speed is less than you wanted, you’ll want to); and some dramatic graphs of how 4G and broadband speeds are moving.
link to this extract
Errata, corrigenda and ai no corrida: none notified