23 October 2017

The Atlantic: “Inside Waymo’s Secret World for Training Self-Driving Cars”

Originally developed as a way to “play back” scenes that the cars experienced while driving on public roads, Carcraft, and simulation generally, have taken on an ever-larger role within the self-driving program.

At any time, there are now 25,000 virtual self-driving cars making their way through fully modeled versions of Austin, Mountain View, and Phoenix, as well as test-track scenarios. Waymo might simulate driving down a particularly tricky road hundreds of thousands of times in a single day. Collectively, they now drive 8 million miles per day in the virtual world. In 2016, they logged 2.5 billion virtual miles versus a little over 3 million miles by Google’s IRL self-driving cars that run on public roads. And crucially, the virtual miles focus on what Waymo people invariably call “interesting” miles in which they might learn something new. These are not boring highway commuter miles.

The simulations are part of an intricate process that Waymo has developed. They’ve tightly interwoven the millions of miles their cars have traveled on public roads with a “structured testing” program they conduct at a secret base in the Central Valley they call Castle.

Alexis C. Madrigal

Fascinating inside look at the Waymo operation tasked with making autonomous cars a reality. The comparison with competitors in terms of real and simulated miles makes it evident how far ahead Google is on the software side. In my opinion, employing simulation is definitely the better strategy here, as physical cars would never be able to match the amount of simulated hours on the road; this also allows engineers to replay the same situation again and again and measure if and how the driving algorithm improves. Tesla is mentioned in the article as a strong competitor, because they are actively collecting data from their existing vehicles, but the company is limited by the number of cars sold (where they continue to have manufacturing issues) and the situations Tesla drivers are encountering.

Monday Note: “Tesla’s New Car Smell”

My first serious doubts about Tesla didn’t stem from missed schedules, I’ve been guilty of too many of these, they’re part of tech life. What seriously worried me was a July 2016 visit to Tesla’s manufacturing plant in Fremont, California. In taking delivery of my wife’s Model S, we were treated to a group tour of the site. Everyone marveled at the robot porn, at the activity on the assembly line, at the endless stores of spare parts piled to the ceiling.

Everyone but yours truly.

I couldn’t help check off the sins against the “Toyota Bible”, prescriptions for car manufacturers that are lucidly detailed in The Machine That Changed The World (a great and, in parts, sad read). In particular, one mustn’t stockpile parts on the floor, they must be fed in small quantities at small time intervals. If a part has a problem, only a small quantity needs to be shipped back to the supplier who can inspect, correct, and quickly adjust their own production process.

Jean-Louis Gassée

Sounds like the Tesla myth is starting to crack. From random firings to handmade parts in a plant that is supposed to be automated, and now the company has missed production goals for the Model 3 – by a lot! Somebody in their management team should really look up ‘Lean manufacturing’ when they’re not being yelled at by Elon Musk. This doesn’t make me any more confident that Musk’s other high-profile bet, SpaceX, can deliver on their plans for future space launches.

20 October 2017

Photoshop Blog: “Introducing: Lightroom CC, Lightroom Classic CC and more”

What’s the Difference between Lightroom and Lightroom Classic?

Going forward, the product you’ve known as Lightroom will be rebranded “Lightroom Classic CC”. Why change? We have introduced a new photography service that will now be called “Lightroom CC”. It is designed to be a cloud-based ecosystem of apps that are deeply integrated and work together seamlessly across desktop, mobile, and web. Lightroom Classic CC is designed for desktop-based (file/folder) digital photography workflows. It’s a well-established workflow solution that is distinct and separate from our new cloud-native service.  By separating the two products, we’re allowing Lightroom Classic to focus on the strengths of a file/folder based workflow that many of you enjoy today, while Lightroom CC addresses the cloud/mobile-oriented workflow.

Tom Hogarty

Big news for Lightroom users, but slightly confusing on several levels. Basically, there are now two applications named Lightroom: the ‘Classic’ is an update to the previous software with new features, and there’s a new, cloud-based Lightroom. It’s not mentioned clearly in the announcement, but there are now two Creative Cloud photography plans to choose from: one includes Photoshop and the two Lightroom flavors like before, and the other offers 1TB cloud storage space for RAW file backup, but only the new Lightroom as desktop/mobile editing app.

15 October 2017

The Guardian: “Israel-Palestine: the real reason there’s still no peace”

The real explanation for the past decades of failed peace negotiations is not mistaken tactics or imperfect circumstances, but that no strategy can succeed if it is premised on Israel behaving irrationally. Most arguments put to Israel for agreeing to a partition are that it is preferable to an imagined, frightening future in which the country ceases to be either a Jewish state or a democracy, or both. Israel is constantly warned that if it does not soon decide to grant Palestinians citizenship or sovereignty, it will become, at some never-defined future date, an apartheid state. But these assertions contain the implicit acknowledgment that it makes no sense for Israel to strike a deal today rather than wait to see if such imagined threats actually materialise. If and when they do come to be, Israel can then make a deal. Perhaps in the interim, the hardship of Palestinian life will cause enough emigration that Israel may annex the West Bank without giving up the state’s Jewish majority. Or, perhaps, the West Bank will be absorbed by Jordan, and Gaza by Egypt, a better outcome than Palestinian statehood, in the view of many Israeli officials.

Nathan Thrall

That’s a damn valid reason! Israel (as a state) has little to lose from the status quo, given the unwillingness of the international community to impose tough sanctions. This lack of resolve is somewhat understandable, seeing how Israel is the only real long-term ally of Western democracies in a region largely controlled by military and religious autocracies – including Turkey lately. On the other hand, it exposes a dangerous double-standard, if we compare it, for example, with Russia’s incursion in Crimea, where sanctions were swiftly imposed.

13 October 2017

Vanity Fair: “How Elizabeth Holmes’s House of Cards came Tumbling Down”

Like Apple, Theranos was secretive, even internally. Just as Jobs had famously insisted at 1 Infinite Loop, 10 minutes away, that departments were generally siloed, Holmes largely forbade her employees from communicating with one another about what they were working on—a culture that resulted in a rare form of executive omniscience. At Theranos, Holmes was founder, C.E.O., and chairwoman. There wasn’t a decision—from the number of American flags framed in the company’s hallway (they are ubiquitous) to the compensation of each new hire—that didn’t cross her desk.

And like Jobs, crucially, Holmes also paid indefatigable attention to her company’s story, its “narrative”. Theranos was not simply endeavoring to make a product that sold off the shelves and lined investors’ pockets; rather, it was attempting something far more poignant. In interviews, Holmes reiterated that Theranos’s proprietary technology could take a pinprick’s worth of blood, extracted from the tip of a finger, instead of intravenously, and test for hundreds of diseases—a remarkable innovation that was going to save millions of lives and, in a phrase she often repeated, “change the world”. In a technology sector populated by innumerable food-delivery apps, her quixotic ambition was applauded.

Nick Bilton

Silicon Valley arrogance at it’s finest. While some startup failure stories have an almost comic quality, this one verges on tragedy. Not for the founder, who seems so entangled in her illusions of grandeur that she completely lost touch with reality, but for the patients who used her fake blood tests.

12 October 2017

War on the Rocks: “The Rise and Fall of Erdoganocracy: Why Victory May Defeat Turkey’s President”

Eventually, the alliance fell apart in 2013, as Gulenists turned against Erdogan. The war between the two culminated in the coup attempt, which allowed Erdogan to purge most known Gulenists and opponents of all other ideological stripes from the state institutions and beyond.

Having relied on Gulenists as a substitute for secularists in the bureaucracy, this presents Erdogan with a human resource challenge. Erdogan and the AKP’s best bet in the short term is rewarding loyalty, not necessarily merit. The long term impact of this strategy will be dire: increased corruption and nepotism, decaying institutional effectiveness, and a flailing economy. Erdogan will likely blame all of this on Turkey’s opposition, but populist rhetoric has its limits, usually defined in terms of what everyday people experience in their own lives.

There will be a short term impact, too. Within the AKP, Erdogan will increasingly favor his die-hard loyalists. It is unlikely that the resulting resentment against unadulterated patronage within AKP ranks, as some commentators argue, will lead to the implosion of the party. More likely is a future where the AKP, marginalizing whatever is left of its own talent, will cannibalize what made it a big success story in the first place: institutional coherence and discipline. Put simply, one cannot have the cake and eat it too. As sociologist Max Weber recognized a century ago, “charismatic authority” and “bureaucratic authority” cannot easily co-exist. Erdogan’s rise as the “one man”, not only in Turkey but also within his own party, will also mean that AKP will start to decay as an organization.

Burak Kadercan

Interesting insights into the political situation in Turkey, which is becoming increasingly tense and polarized under the current president’s authoritarian measures. The secular, democratic state that his predecessors strived to build seems further and further from reality. I found the paragraphs above especially striking, as I suspect a similar situation is developing in the United States under the presidency of Donald Trump.

10 October 2017

The Verge: “Apple now sells an iPhone dongle with a headphone jack and charging port”

It took an entire year after the release of the iPhone 7 for Apple to start selling a dongle that lets you plug in headphones (or your car’s AUX cable) and charge at the same time. But now it’s here. Apple is just selling the thing, mind you — not making it. In September, Belkin quietly announced a new, second version of its “Rockstar” adapter that now includes both a 3.5mm jack and Lightning port.

Chris Welch

More courage!

Backchannel: “The Myth of a Superhuman AI”

There is no doubt that a super AI can accelerate the process of science. We can make computer simulations of atoms or cells and we can keep speeding them up by many factors, but two issues limit the usefulness of simulations in obtaining instant progress. First, simulations and models can only be faster than their subjects because they leave something out. That is the nature of a model or simulation. Also worth noting: The testing, vetting and proving of those models also has to take place in calendar time to match the rate of their subjects. The testing of ground truth can’t be sped up.


To be useful, artificial intelligences have to be embodied in the world, and that world will often set their pace of innovations. Without conducting experiments, building prototypes, having failures, and engaging in reality, an intelligence can have thoughts but not results. There won’t be instant discoveries the minute, hour, day or year a so-called “smarter-than-human” AI appears. Certainly the rate of discovery will be significantly accelerated by AI advances, in part because alien-ish AI will ask questions no human would ask, but even a vastly powerful (compared to us) intelligence doesn’t mean instant progress. Problems need far more than just intelligence to be solved.

Kevin Kelly

I’ve written about this in the past, but I think it’s worth repeating: AI doesn’t exist somewhere outside the physical restrictions of the real world, nor will it magically solve all human problems. And this article does a wonderful job of exposing and taking apart five assumptions most people make when discussing AI – well worth reading!

09 October 2017

Gizmodo: “Uber’s iOS App had Secret Permissions that allowed it to Copy your Phone Screen”

The screen recording capability comes from what’s called an “entitlement”—a bit of code that app developers can use for anything from setting up push notifications to interacting with Apple systems like iCloud or Apple Pay. This particular entitlement, however, was intended to improve memory management for the Apple Watch. The entitlement isn’t common and would require Apple’s explicit permission to use, the researchers explained. Will Strafach, a security researcher and CEO of Sudo Security Group, said he couldn’t find any other apps with the entitlement live on the App Store.

It looks like no other third-party developer has been able to get Apple to grant them a private sensitive entitlement of this nature, Strafach said. Considering Uber’s past privacy issues I am very curious how they convinced Apple to allow this.

Kate Conger

Tell me again how Apple is the ultimate defender of user privacy! Until they need help promoting one of their products, then all principles fly out of the window.

What if the meeting between Tim Cook and Travis Kalanick in early 2015 wasn’t about warning Kalanick to stop circumventing Apple’s rules, but instead to ask for Uber’s help in testing Apple Watch apps?!

The New Yorker: “A.I. versus M.D.”

The most powerful element in these clinical encounters, I realized, was not knowing that or knowing how—not mastering the facts of the case, or perceiving the patterns they formed. It lay in yet a third realm of knowledge: knowing why.


Knowing why—asking why—is our conduit to every kind of explanation, and explanation, increasingly, is what powers medical advances. Hinton spoke about baseball players and physicists. Diagnosticians, artificial or human, would be the baseball players—proficient but opaque. Medical researchers would be the physicists, as removed from the clinical field as theorists are from the baseball field, but with a desire to know “why”. It’s a convenient division of responsibilities—yet might it represent a loss?

A deep-learning system doesn’t have any explanatory power, as Hinton put it flatly. A black box cannot investigate cause. Indeed, he said, the more powerful the deep-learning system becomes, the more opaque it can become. As more features are extracted, the diagnosis becomes increasingly accurate. Why these features were extracted out of millions of other features, however, remains an unanswerable question. The algorithm can solve a case. It cannot build a case.

Siddhartha Mukherjee

Interesting overview of the promises and challenges of applying deep-learning techniques in medical diagnosis. There are certainly high expectations in the area, most notably IBM’s Watson, but recent reporting shows it is still far from fulfilling the promises. As I (and others) commented before, the big challenge for AI in any area is evolving past basic pattern recognition (factual knowledge) to a deeper understanding of the world, knowing how and why things work.

08 October 2017

The Guardian: “Total recall: the people who never forget”

Price was the first person ever to be diagnosed with what is now known as highly superior autobiographical memory, or HSAM, a condition she shares with around 60 other known people. She can remember most of the days of her life as clearly as the rest of us remember the recent past, with a mixture of broad strokes and sharp detail. Now 51, Price remembers the day of the week for every date since 1980; she remembers what she was doing, who she was with, where she was on each of these days. She can actively recall a memory of 20 years ago as easily as a memory of two days ago, but her memories are also triggered involuntarily.


The findings suggest that no one, not even a “memory wizard”, is immune to the reconstructive mechanisms that enable memory distortions. When people with average memory recall an experience, it is formed not only by what they think happened and how they felt at the time, but by what they know and feel now. We’re pulling together everything in the present to come up with an approximation of the past, and that’s the same with HSAM people, Patihis said. The findings were not popular with some of the HSAM subjects because, as Stark, a co-author on the paper, pointed out, having accurate memories is central to their identities.

Linda Rodriguez McRobbie

How would your life look like if you couldn’t (or wouldn’t) forget anything? Tricky question, since very few people actually experience this strange medical condition – and, as discovered, even they are subject to the imperfections of human memory. It’s a fascinating subject to study, but I fear, with so little cases, it will remain a medical curiosity, rather than uncovering new insights into how memory works.

03 October 2017

Scientific American: “Will Democracy Survive Big Data and Artificial Intelligence?”

Let us suppose there was a super-intelligent machine with godlike knowledge and superhuman abilities: would we follow its instructions? This seems possible. But if we did that, then the warnings expressed by Elon Musk, Bill Gates, Steve Wozniak, Stephen Hawking and others would have become true: computers would have taken control of the world. We must be clear that a super-intelligence could also make mistakes, lie, pursue selfish interests or be manipulated. Above all, it could not be compared with the distributed, collective intelligence of the entire population.

The idea of replacing the thinking of all citizens by a computer cluster would be absurd, because that would dramatically lower the diversity and quality of the solutions achievable. It is already clear that the problems of the world have not decreased despite the recent flood of data and the use of personalized information—on the contrary!


In a rapidly changing world a super-intelligence can never make perfect decisions (see Fig. 1): systemic complexity is increasing faster than data volumes, which are growing faster than the ability to process them, and data transfer rates are limited. This results in disregarding local knowledge and facts, which are important to reach good solutions. Distributed, local control methods are often superior to centralized approaches, especially in complex systems whose behaviors are highly variable, hardly predictable and not capable of real-time optimization. This is already true for traffic control in cities, but even more so for the social and economic systems of our highly networked, globalized world.

Dirk Helbing et al.

Interesting perspective on the ongoing debate about AI and its consequences on human society. The vision of a world carefully guided by a super-intelligence vaguely reminds me of Asimov’s Robot series – but at least his positronic brains had a clear set of hardcoded ethical guidelines. The article above makes a compelling counterpoint to this vision of increasingly offloading decisions to AI: however benevolent, ceding control to an artificial intelligence would mean nothing less than to erode the idea of democracy and self-determination, to roll back the social advancements of the enlightenment and return to a form of digital despotism and conformism.

02 October 2017

WebKit Blog: “Designing Websites for iPhone X”

Out of the box, Safari displays your existing websites beautifully on the edge-to-edge display of the new iPhone X. Content is automatically inset within the display’s safe area so it is not obscured by the rounded corners, or the device’s sensor housing.

The inset area is filled with the page’s background-color (as specified on the <body> or <html> elements) to blend in with the rest of the page. For many websites, this is enough. If your page has only text and images above a solid background color, the default insets will look great.

Timothy Horton

If you haven’t read this so-called guide yet, this is as far as you need to go. The solutions that follow are increasingly worse:

  1. Using the Whole Screen causes some of the page’s content to be hidden under the notch.
  2. Respecting the Safe Areas recommends a new WebKit-only CSS function that the standardization body decided to adopt under a different name – expect incompatibility problems as this rolls out to other browsers. On top of that, this may break layout on regular iPhones, and probably other smartphones.
  3. Bringing it All Together – and finally for the cherry on top: CSS functions no current browser supports, not even iOS 11 Safari!

01 October 2017

Ars Technica: “Finding North America’s lost medieval city”

My first find was a slab of red pottery that crumbled to dust in my fingers. “Don’t worry about that,” Baires assured me. “It’s just unfired clay and it won’t hold up.” Later, I found nuggets of charcoal, blobs of yellow pigment, a few jagged pieces of fired pottery, and several burned deer bones.

The bones were the worst, because there were so many of them that it halted our digging dozens of times. We had to be careful to determine that these weren't human bones, because human remains must be reported immediately. Though we'd already identified these as deer bones, archaeologists will sometimes do a lick check to be sure. Lick check? I stared at Baires in bewilderment. “Do you want to lick it?” she asked. “Bones are porous, so your tongue will stick to it.” The students looked at me. Would the weird journalist do it? Hell yeah, I would. I brought a small fragment of bone to my mouth, tasted salt, and felt my tongue adhere lightly to the surface. “Yep, it’s deer”, Baires shrugged.

After I’d been shovel scraping for an hour, blisters started coming up and popping on my fingers. When I fell exhausted into bed at 8:30pm, I could feel the exact part of my thigh that I used to push the shovel handle. I couldn't stop thinking about how I licked the bones of a deer that was cooked for a feast in Cahokia 900 years ago. I wish I could have been there to see the party, but this might be the next best thing.

Annalee Newitz

If you’ve ever played Civilization 5, you’re certainly familiar with the name Cahokia as one of the in-game city states. It’s fascinating to discover more about its history, another case of mysterious civilization going through the inevitable cycles of rise and fall in the middle of the American continent.