Microsoft Surface Book vs. Apple MacBook Pro: Which one deserves your dollars?

Microsoft’s Surface Book has already drawn a great deal of press for its flexible design and crazy hinge. At its unveil, however, Microsoft took a number of potshots at one of the highest-regarded devices on the market today — Apple’s Macbook Pro. While we still don’t have all the specifications on Microsoft’s new laptop, we’ve got more than enough for a preliminary analysis of the two.

Since the Surface Book is only available in a 13.5-inch form factor, we’ll focus this comparison against the MacBook Pro 13-inch — though with some acknowledgements of the other Apple systems where appropriate. The Surface Book starts at $1,499, so we’ll compare primarily against the equivalent MacBook Pro.


Spec comparison

The Apple MacBook Pro uses a Core i5-5257U 28W CPU. Microsoft, of course, has said nothing about its basic CPU, although it showed off a Core i7 option (above), but we can make some educated guesses. If Microsoft chose to use a 15W chip in its basic model, we’re looking at something like a Core i5-6200U, with a 2.3GHz base clock and a 2.8GHz maximum frequency. That would put the Skylake CPU a touch behind Apple’s Broadwell, though maximum RAM frequencies and bandwidth would both be higher. The HD Graphics 520 onboard the 15W Skylake is also more advanced than Broadwell’s GPU. Alternately, MS could be using the Core i5-6300U, with a 2.4GHz base clock and 3GHz maximum frequency. Either way, CPU performance between Apple and MS should be very close.

At first glance, Microsoft seems to win the screen resolution comparison — the 13.5-inch display on the Surface Book has a PPI of 267 and becomes Retina-class at roughly 13 inches. The MacBook Pro has a PPI of 227 and becomes Retina-class at 15 inches. This comparison is undercut, however, by the fact that Microsoft and Apple take two very different approaches to font scaling — and Apple’s is typically perceived as better. Blurry text and improper application scaling are still problems under Windows 10, whereas Apple has largely solved these issues.


Surface Book, showing off the insane hinge

Poor DPI scaling in Windows is typically the fault of the app developers, not Microsoft, but I’ll wait and see how the two products look before calling this one. The 3000×2000 screen on the Surface Book isn’t going to do battery life any favors.

As for battery life, Microsoft is claiming 12 hours on standard battery with just 3 hours of useful run time if you detach the lid and use it like a tablet. Apple claims 10-11 hours for the MacBook Pro. The Surface Book, of course, will have touchscreen support, while Apple tends to offer better touchpads (and, as a rule, better support for multi-touch control).

What about that GPU?

Stepping up to the GPU-equipped model costs $1,899, minimum. Microsoft is charging $200 for the discrete graphics solution, and $200 more if you want a Core i7 vs. a Core i5. We don’t know anything about the GPU, save that it ships with 1GB of dedicated GDDR5. That’s not much RAM — even Nvidia’s lower-end mobile solutions, like the GTX 950M, typically pack at least 2GB. There are two possibilities here: Either Microsoft loaded the system with a low-end GPU that both it and Nvidia have refused to disclose because they don’t want to lose the hype around the idea of an ultrabook system offering any discrete cards at all, or Nvidia removed VRAM from a SKU that typically carries more in order to hit Microsoft’s preferred power target. NV is claiming that this chip is a “custom” design, but that doesn’t really mean much — a standard GPU + smaller VRAM loadout still qualifies as custom, even if it’s based on Maxwell.

A comparison that comes down to physics

People have been falling over each other since the Surface Book was announced to declare which system will end up being the winner. Such maneuvering is premature. We know Microsoft will offer a high-end Core i7 option in the Surface Book, and made mention of “two extra cores,” but has not disclosed which systems will carry a quad-core Core i7 (there are mobile dual-core Core i7s, too). Similarly, we don’t know how fast the GPU is, how well the battery holds up in each configuration, and how the various options will impact the weight of the system. Trying to judge which system is “better” than the other is impossible, because too much of that question relies on the look and feel of each computer.

Retina MacBook 13 Apple

What the specs tell us is this: Microsoft has pushed the envelope on screen resolution, GPU performance, and CPU capability farther than Apple has. But will that result in a better device that justifies its price point? That’s impossible to know right now. We’ve already seen evidence of how different manufacturer priorities resulted in Core M systems with significantly different performance characteristics. The big questions for devices like this will be how smoothly the hardware switches from tablet to docked mode, how well battery life on Windows 10 compares to OS X, and exactly which GPU Nvidia and Microsoft included. 1GB of RAM may make for a nice sound-byte, but it’s one of the least useful metrics for measuring GPU performance.

If I had to guess right now, I’d wager that both machines will be well-designed and capable, with which one is better coming down to what you want to use it for. I’d feel a bit better about the Surface Book if Microsoft could resolve its DPI problems, but that’s unlikely.


Light’s L16 is a DSLR-quality camera that fits in your pocket — for a stiff price

This week, startup Light announced plans to sell a 52MP camera that creates DSLR-quality images, but is only the size of a smartphone. The Light L16 will do this by combining images from multiple lower-resolution sensors — 10 of its 16 sensors fire for each shot — into a single high-resolution image. By cleverly combining images from some of its 35mm, 70mm, and 150mm fixed-focal-length lenses, the L16 can zoom from 35mm to 150mm (in 35mm equivalent terms). Light expects the L16 to be available in late summer of 2016 for $1,700, but it is currently accepting a limited number of pre-orders for units at $1,300.

The magic behind Light

The L16 has a touch-enabled screen that allows pinch to zoomWe’ve written in some detail how the multi-sensor Light module works, but the gist of it is that several overlapping images are fused in software into a single image. Because each sensor is small, the lenses needed are also small, resulting in a thin device. In addition, to provide telephoto capability, longer-focal-length lenses are placed in the unit sideways, and augmented with a mirror that allows them to capture images out of the front of the camera. Those mirrors are also movable, which helps the Light bring as many imagers as possible to bear for each zoom level.

Obviously this is easier to say than to do. The computational imaging work required to fuse the images is substantial, as the images have to be perfectly aligned in order to be merged successfully, and shutter and image readout timings have to be carefully controlled to avoid artifacts — especially with moving subjects.

Along with 52MP still images in JPEG, TIFF, and DNG (RAW), the L16 will be able to capture 4K video, with optical zoom from 35mm to 150mm. It will even be able to use its multiple sensors to capture single-shot HDR images, by setting them to differing exposures. It will feature a dual-tone built-in LED flash, although no hot shoe for an external flash.

When we first spoke with the Light founders, their vision was an $80 camera module that could be included in a smartphone. Obviously, creating and getting design wins for something like that takes years, so the idea of producing a standalone version entirely under their control as a first product makes a great deal of sense. If it sells well, it will help fund development of successive generations of the technology.

At the wide end the L16 has a 35mm-equivalent focal length

At the wide end the L16 has a 35mm-equivalent focal length

By selecting different sensors and moving small mirrors the L16 can zoom in to a 150mm-equivalent focal length

By selecting different sensors and moving small mirrors the L16 can zoom in to a 150mm-equivalent focal length

Will your next DSLR be a Light L16 instead?

A closeup of the 16 sensors on the appropriately-named Light L16Standalone camera sales have been tanking for several years now. The smartphone is by far the world’s most popular camera, with point-and-shoots and DSLRs falling out of favor quickly, and mirrorless fighting a rearguard action. But a smartphone-sized camera with all the features of a DSLR might just be enough to breathe life into the camera market. Light isn’t the first company to push in this direction. The DxO ONE augments an iPhone with a dedicated 1-inch-format sensor camera that produces 20MP images for $600, but the L16 aims much higher, with a much higher price tag to match.

The sample images from the L16 are gorgeous, and the promises Light has made about its functionality are impressive. But as other camera startups like Foveon, Lytro, and Pelican have learned, there is a big difference between prototypes and market success. I’m looking forward to when the L16 is out in the wild and can be evaluated for real.


ET deals: Dell Optiplex 3020 small form factor quad-core desktop for $529

Need a tiny PC that can still handle a heavy workload? Check out the OptiPlex 3020 small form factor desktop PC from Dell. It’s less than four inches thick, weighs roughly 13 pounds, and it doesn’t compromise on performance. The best part? You can save over $350 off the sticker price when you use today’s coupon code.

OptiPlex 3020 So, what kind of specs are we looking at? The OptiPlex 3020 has a fourth generation quad-core 3.3GHz Intel Core i5-4590 CPU, integrated Intel HD Graphics 4600, 4GB of DDR3 RAM (1600MHz), a 500GB 7200RPM hard drive, and a DVD burner. And since it has a VGA port and a DisplayPort on the rear, you’ll be able to plug into just about any monitor or HDTV. DisplayPort-to-HDMI adaptors cost 15 bucks or less when you buy on Amazon, so you won’t have to spend much to get up and running on your existing gear.

If you want to expand the capabilities of this tiny computer, it’s incredibly easy. Just open up the case, and you’ll see that it has a half height PCIe x16 slot, a half height PCIe x1 slot, two DIMM slots (for up to 16GB of RAM), a 3.5-inch drive bay, and a 5.25-inch slimline drive bay. So if you want to keep your PC running for a long time, rest easy knowing that the upgrade process is simple.

Even if you don’t particularly feel like dealing with the insides, it comes with two USB 3.0 ports and six USB 2.0 ports. Adding external storage or additional peripherals is a breeze.

Out of the box, this model has Windows 7 Professional (64-bit) installed, but a license for Windows 8.1 Pro (64-bit) is included too. Even better, you can upgrade to Windows 10 Pro (64-bit) for free until July of next year. If you want to take advantage of the latest and greatest features out of Redmond, it’s well worth a little bit of your time to go through the upgrade process.

Typically, this configuration retails for $898.57, but Dell is currently selling it online for just $549. And if you apply the coupon code “OPTI20OFF” during the checkout process, you’ll immediately save an additional 20 bucks. Take advantage of Dell’s free shipping option, and you’re only paying $529 (plus any applicable taxes). This is a limited time offer though, so don’t hesitate.

Our commerce group sources the best deals and products for the ET Deals posts. We operate independently of Editorial and Advertising and may earn a percentage of the sale, if you buy something via a link on the post. If you are interested in promoting your deals, please contact us at

For more great deals head over to TechBargains.


Volvo: We’ll take the blame if our self-driving cars crash

Volvo says “blame us” if any of its self-driving cars crashes when it’s in autonomous mode. The company hopes the assumption of liability takes away one concern of the public and spurs the development and adoption of self-driving cars. Volvo says self-driving cars will be here sooner than many believe.

Separately, Volvo called for a single federal set of standards for autonomous vehicles, also to speed up development of the vehicles. Volvo released the top lines on its corporate stance in advance of a seminar on self-driving it’s sponsoring Thursday in Washington.

Park Assist Pilot offers automatic reversing into a parking bay as well as entering and exiting a parallel parking spot.

One US standard … or cede leadership

In a release issued a day ahead of a seminar sponsored by Volvo and the Swedish embassy, Volvo president and chief executive Hakan Samuelsson said, “The US risks losing its leading position [on self-driving] due to the lack of federal guidelines for the testing and certification of autonomous vehicles. Europe has suffered to some extent by having a patchwork of rules and regulations. It would be a shame if the US took a similar path ….”

Perhaps unlike Sweden, the 50 US states are sometimes protective of their turf and might believe they have a better handle on what rules and regulations are best suited for self-driving cars. Not all of them see federal supremacy as a good thing. Nevada might see the importance of high-speed self-driving on open highways, where New Jersey would want to insure the cars could handle heavy traffic and Massachusetts might want to ensure autonomous cars don’t knock down pedestrians in Boston. It’s possible the lightly populated southwestern states without much snow would prefer softer rules — initially — to lure car testing.

And also … we accept liability (but crack down on hackers)

Then, in a Steve Jobs-like “one more thing,” at the end of the announcement the company added almost as an afterthought, “[Samuelsson Thursday] will say that Volvo will accept full liability whenever one if its cars is in autonomous mode, making it one of the first car makers in the world to make such a promise.”

Volvo also wants to see tougher rules on car hacking — “a criminal offense” — obviously a concern where self-driving cars are concerned. According to Samuelsson, “We are constantly evolving defensive software to counter the risks associated with hacking a car. We do not blame Apple or Microsoft for computer viruses or hackers.”

City Safety is now the umbrella name for all auto brake functions. The all-new Volvo XC90 is the first car in the world with technology that features automatic braking if the driver turns in front of an oncoming car.

Volvo’s plan for fewer deaths

In 2008, Volvo unveiled Vision 2020, a safety plan for driver assists and better safety features. “By 2020, nobody shall be seriously injured or killed in a new Volvo,” said Volvo’s top safety expert, Anders Eugensson, said at the time. “[Volvo has a] long-term vision to create cars that will not crash.” Because of “freak accidents” and the like, Volvo’s vision is moving toward zero deaths.

Most engineers and recognized auto safety experts — as opposed to the lay public with access to the Comment button — believe that driver assists up to and including autonomous driving will push the death rate down significantly. Already dozens of car models offer the components of self-driving 2015 style meaning as long as you keep your hands lightly on the wheel and pay some attention to the road, the combination of adaptive cruise control, blind spot detection, and lane departure warning or lane keep assist drives the car on limited access highways. The driver still has to watch out for other crazy drivers as well as the quirks in current-model driver assists, such as ACC that can’t pick up and react to a driver cutting in front of you if the other guy also lets off on the throttle.

The City Safety system in the all-new Volvo XC90 features Cyclist detection with full auto brake, day and night.

Volvo’s plan to regain the mantle of world’s safest brand

From the 1950s through the 1990s, Volvo was considered by many to be the automaker most invested in safety. By the 1990s, however, German automakers with more cash on hand invested heavily in safety, while Volvo suffered more than a decade of lagging technology and aging models after Ford bought the Swedish company in 1999. Under new ownership by China’s Geely Holding Group since 2010 and with a $10 billion infusion of cash, the company has remade itself with new models such as the Volvo XC90 SUV and added more safety features unique for the time being, such as auto-braking if another car turns left in front of you. Volvo was early with city safety (auto braking at low speed) and pedestrian and cyclist safety (auto braking for pedestrians and cyclists).

Volvo’s event Thursday is a reminder that Volvo is laying claim to being at the forefront of autonomous driving research and development.


How ‘unboiling an egg’ leads to better cancer treatments

There is a famous adage in physics that says, “your theory may be beautiful, but if it isn’t absolutely hilarious, you are probably just wasting your time.” The most astonishing scientific discoveries are often those that surprise us enough to make us laugh before we even have time to think. When our thoughts finally do catch up with our eyes and ears, we have been changed — inevitably, we come to know a little bit less, because some of the explanations we once held dear can no longer be true.

On the other hand, we also become a little bit wiser as those explanations that survive grow stronger. One thing most of us “know” is that you can’t unscramble an egg. Even if you could, you couldn’t possibly unboil an egg. One need do no more than trot out simple thermodynamics to show that the heat applied to the egg irreversibly denatures its proteins. As the saying goes, Humpty Dumpty could never be be put back together again.

But eggs, particularly their proteins and DNA, are not really so simple. When swaddled with more sober amounts of heat, they become chickens. In fact, if you treat a boiled egg right, it is even possible to unboil it. The guy who discovered that just received the Ig Nobel Prize last month for the method he published earlier this year. Colin Raston, of Flinders University in Adelaide, didn’t set out to unboil eggs or win an Ig Nobel. He wanted to find a general way to unravel and untangle proteins. To do that, he built a vortex machine capable of mechanically separating long strands of proteins that had been pre-processed with urea.

Urea not only chews up and unfolds proteins, but it also coats and protects them against re-aggregating in the vortex drive. When an egg is cooked, one of the first proteins to begin to gel is lysozyme. This multifunctional bactericide is naturally abundant in egg whites, and is also found in places like tears, saliva, milk, and mucus. When proteins such as lysozyme are heat-denatured, electrical charges that were originally ensconced away on the protein’s interior are exposed when it unfolds. That makes them available to bond into larger conglomerates that, incidentally, will scatter light more effectively.

Raston and his colleagues first perfected their methods with lysozyme and then moved on to larger proteins. They were even able to get proteins to refold back into their native forms within a few minutes. This is a huge improvement over the standard dialysis techniques now used, which more likely will take all day to do that. Refolding crystallized clumps of proteins is a bit more complicated than, for example, re-ordering grain structure in heat-treated metals. But it may be a good analogy for us here at a basic level. When proteins useful to humans are made on an industrial scale by coaxing bacterial into synthesizing them in huge vats, the main difficulty is that it requires more than just controlling the temperature to prevent them from crystalizing out into sticky clumps.

The way that a healthy cell controls its protein factories is to bind the growing tapes of amino acids with little protector molecules as they are being translated on the ribosome. This prevents the protein from folding prematurely before the full strand is done. If a human protein is instead expressed in bacteria and synthesized in a big vat, many of the essential accessory molecules and templates needed for proper folding are likely missing. To fully replicate all those cozy eukaryotic intangibles that our proteins have become accustomed to and rely on for proper assembly, within in an amorphous primitive bacterial slurry, is still a difficult challenge. If these so-called ‘recombinant’ proteins being synthesized are actually drugs for treating cancer, processing inefficiencies end up costing a lot of time and money.

Recombinant forms of insulin, for example, can alleviate the need to use inferior or inconvenient ‘natural’ sources (like cows) to make them for us. But insulin is a fairly simple peptide whose secondary folded structure is fairly well understood. Newer drugs like the ZMapp used to treat Ebola contain several antibody proteins which are only just beginning to be understood. The best way to produce ZMapp has been to splice the genes for it into a tobacco plant where the products could be later harvested. At the height of the Ebola scare there was simply no way to produce quality ZMapp in the quantities that would be needed in the case of an epidemic.


As far as cancer, understanding folding has important implications beyond just making drugs. Misfolding is a double-edged sword in that it can be both a cause and an effect of tumerogenicity. For example, the energetic deficits commonly observed in cancer cells can result in an oversupply of misfolded proteins. On the other hand, it is sometimes misfolded proteins themselves that can be the cause of the cancer. Much the same conundrum has been seen in the role of mitochondria as compared to genetic mutations in cancer. Although mutations can clearly result in over-expression of the so-called ‘oncogenes’ that make cells multiply uncontrollably, researchers now appreciate that energetically compromised mitochondria may be the more fundamental driver of tumor progression.

When mutation is understood to occur as the result of the energetic failure of normal repair mechanisms, or secondary to metabolic adjustments to that failure, the spectrum of cancer causes and effects comes full circle. As mentioned above, treating cancer can now be an expensive proposition, particularly some of the eclectic antibody drugs typically recognized by a fancy name ending with the suffix ‘mab’ (for monoclonal antibody). Antibodies are basically the universal computers of the immune system, in the sense that can be made on demand to recognize just about any molecule one can imagine. Anything from large viral coat proteins, to small metals, and perhaps even to things no less slippery than teflon itself.

Tales of $1,000-per-dose regimens for tough-to-treat tumors are no exaggeration. One drug commonly used as part of a cocktail elixir given for certain white-cell cancers — B-cell tumors or Hodgkins for example — is Rituximab. This is an interesting one, because it comes off of patent protection this year and can theoretically be opened up to the beneficial effects of greater competition. B cells are the cells which are responsible for making our own antibodies to pathogenic invaders. What we potentially have here is the prospect of treating faulty antibody-producing cells, crippled by their own misfolded proteins, with separate antibody drugs manufactured by controlling proper folding to target those proteins.

When the Ig Nobel prize givers say that their goal is as much to make us laugh as it is to help us learn, they are absolutely serious. Lest anyone doubt their success so far, we might note they are gaining in popularity compared to the ‘real’ Nobel prize. For example, who knows the recipients that won the Nobel yesterday in chemistry for their work on the DNA toolbox for cell repair? Perhaps a few, but at least now, you all know the chemist that won the Ig Nobel for unboiling an egg.

Check out our ExtremeTech Explains series for more in-depth coverage of today’s hottest tech topics.


Google, Porsche spar over how much data the search giant collects

Google is an advertising company that just happens to own a search engine and an operating system. But the company is exceptionally sensitive to how its data-hoovering practices are discussed in the media. Earlier this week, news broke that Porsche had chosen to use Apple’s CarPlay rather than Android Auto, due to the amount of data that Google gathers.

MotorTrend interviewed engineers at Porsche and reported the following: “As part of the agreement an automaker would have to enter with Google, Porsche said certain pieces of data must be collected and transmitted back to Mountain View, California. Stuff like vehicle speed, throttle position, coolant and oil temp, engine revs—basically Google wants a complete OBD2 dump whenever someone activates Android Auto… Apple, by way of stark contrast, only wants to know if the car is moving while Apple Play is in use.”

Google, meanwhile, is pushing back against these claims. A spokesperson for the company told MotorTrend that it does not collect data on coolant temperatures or throttle position. The company declined to provide a list of exactly what it does gather. Instead, the spokesperson emphasized that using Android Auto is an “opt-in” feature, and that certain data is used for safety, like restricting typing and only allowing voice input when the car is not in park. Other data is used to optimize the user’s “app experience.”


Yes, this comes down to a he-said / she-said situation, but there are reasons to believe Porsche over Mountain View.

Google’s abysmal privacy record

First, there’s the fact that Google openly acknowledges just how much data it gathers. As far back as 2010, Sergey Brin told The Atlantic, “With your permission you give us more information about you, about your friends, and we can improve the quality of our searches,” he said. “We don’t need you to type at all. We know where you are. We know where you’ve been. We can more or less now what you’re thinking about.” The company has been sued for numerous violations of user privacy across both the US and the EU.

Android’s privacy policies give Google enormous power to gather and share data or to sell it in advertising, which is why devices like Amazon’s Fire tablets and the Blackphone use custom OS forks without Google Play or other Google services. The company has faced multiple lawsuits for various privacy-violating schemes and systems and has generally shown NSA-like restraint when it comes to vacuuming data off devices.

The fact that users have to “opt-in” to share data with Android Auto is meaningless. The computing industry loves to talk about opt-in like it’s a separate checkbox or feature. In reality, Microsoft, Apple, and Google all regularly lock critical device features behind an “I Agree” button. If you want to use the capabilities of the Porsche, iPhone, or Surface you just purchased, you’re going to have to agree (though Linux is at least hypothetically a possibility for a computer).

The he-said / she-said nature of this kind of argument almost certainly turns on the kind of subtle technicalities PR firms love to exploit. It’s entirely possible that Porsche is telling the truth when it claims that Android Auto’s license agreement gives Google the ability to track large amounts of information, while Google could be telling the truth when it says it doesn’t currently track specific types of data. Porsche is likely concerned about the kinds of data Google could give itself permission to track if it chooses to do so, and what its legal options are in the event Android Auto begins tracking information Porsche doesn’t agree with.

The final reason I’m more inclined to agree with Porsche is that the auto manufacturer has no reason to lie. It could have explained the situation by claiming that its own internal research found that its customers preferred Apple, or that it felt Apple’s Car Play was simply a better fit for the kind of vehicle experience Porsche wanted to create. There are a hundred ways to explain the situation by emphasizing what makes Apple better as opposed to saying something negative about an Android service. Google, in contrast, has every reason to downplay the types of data Android Auto gathers, or can gather.


New Android adware spotted loaded with root exploits — but you’re safer than you think

A new piece of Android malware is reportedly making the rounds in as many as 20 different countries, and if security firm FireEye is to be believed, it’s quite a nasty bit of code. The exploit, known as Kemoge, was spotted masquerading as a number of legitimate apps, but upon installation it attempts to gain root access on the device, which could allow an attacker to gain complete control. It sounds bad, but as usual, the truth is a bit less sensational than they’d have you believe.

Kemoge is a form of malicious adware, according to FireEye. It borrows the icons from other apps the encourage a user to trust it. The first hurdle for the malware authors to clear is actually getting users to install the app, which is only possible via a third-party app store. That means the user has to download the APK, allow unknown sources in the security settings, then launch the package. Not exactly an easy process.

The way Kemoge functions when deployed on a vulnerable device is actually pretty clever. It copies device information and beams it back to a command and control server first, then it starts inserting ads into the UI, which can pop up in any app or even on the home screen. So that’s annoying, but what it does next is downright malicious. Kemoge contains as many as eight exploits, which uses in an attempt to root the device. This could give the attacker full control over an infected phone. If the infected device is rooted, Kemoge immediately uninstalls any antivirus apps it finds. The exception would be Google Play Services, which runs Google’s antivirus scans. It’s impossible to remove Play Services from a device (even with root) if you still want anything to work.


Are you sufficiently frightened now? What’s described above is really the worst case scenario. The adware aspect of Kemoge should work on almost any device, assuming you go to the trouble of manually installing it. However, the root angle is much less certain. FireEye lists several of the root exploits contained in Kemoge, and they’re all quite old. There’s Motochopper, mempodroid, and a few general Linux kernel vulnerabilities. These are relics from the days when an APK could be used to root your phone. All modern versions of Android should be patched to protect against these flaws. Testing was done with a Nexus 7 running Android 4.3 (software from more than two years ago).

Root exploits are hard to develop on Android these days, but they aren’t always designed to be malware. Many Android users want root access for their own use, and that’s where a lot of the exploits used by Kemoge come from — the enthusiast community. Many devices currently on the market don’t even have functional root exploits for people who want to root their phones, so it’s unlikely Kemoge has a magical unreleased exploit that can root your phone.

Bottom line — the old root methods employed by Kemoge don’t work on popular phones or people would be using them to intentionally root their devices. We’ve reached out to FireEye to get clarification on which versions of Android they’ve confirmed root access on and will update when and if they reply.

Your first line of defense from adware attacks like this is to simply get your apps from the Play Store or from a trusted source like F-Droid or APK Mirror. When you flip the unknown sources switch, you’re instantly less safe.

Update: FireEye got back to use and clarified all the exploits it detected in kemoge are public and several years old (2013 and earlier). They should be patched on all newer phones.


Samsung semiconductor profits rise 80 percent thanks to strong launches, prominent wins

Samsung’s semiconductor division has announced preliminary results ahead of its formal Q3 earnings call. 2015 has been a rough year for Samsung, overall, but the semiconductor unit’s performance should blunt those losses. Samsung claims to have earned an operating profit of $6.3 billion on semiconductor sales in Q3, up substantially from the $3.55 billion operating profit that it posted for the same quarter in 2014.

The vastly improved results were driven by two factors: favorable exchange rates and higher overall shipments. Samsung sold roughly 8% more volume this year compared to last, and its ASPs may have also improved, buoyed by the launch of the Galaxy Note 5 and the Apple iPhone 6s and 6s Plus. It shares the latter hardware with TSMC; Apple opted to dual-source its production last year after years of relying on a single vendor (usually Samsung) for its chip launches.

Some analysts have warned that the majority of this jump was driven by relatively pedestrian causes — namely, shifting foreign exchange rates. Still, we expect to see Samsung finish 2015 on a high note. Qualcomm is preparing to launch its Snapdragon 820, a 14nm chip built on Samsung’s semiconductor technology, and its share of the iPhone 6 business should stand it in good stead for the rest of the year.


Samsung’s Austin foundry

As we previously forecast, this advantage is coming at the expense of TSMC, which has already stated it expects much more modest Q3 2015 results. Back in July, the company stated that it expected revenue to grow by no more than 2.2% for Q3, up to $6.75 billion. Like Samsung, TSMC should see a boost in the end of the year, thanks to Christmas sales and a fresh crop of Android smartphones. It’s not clear which manufacturers are going to adopt TSMC’s 16nm FF+ node as compared to Samsung’s already-shipping 14nm hardware, but 2016 should give the Taiwanese foundry fresh ammunition with which to compete. Both firms have already announced that they intend to enter risk production on 10nm by the end of 2016, though that means something different in foundry land then it does when Intel or AMD say it. Even if both companies meet their goal of Q4 2016 for early 10nm ramps, we don’t expect to see shipping 10nm solutions until the back half of 2017 at the very earliest.

Meanwhile, Samsung is still struggling to turn its phone business around. The company has lost significant low-end market share in recent years, thanks to competition from companies Xiaomi and Huawei. Strong profits from the semiconductor division and overall technical leadership can definitely help, but the Galaxy S6 / S6 Plus simply wasn’t the winning solution that the company hoped it would be. Now that Apple competes directly with Samsung in the large-screen device space, Samsung may want to revisit some of the features that made its devices preferred options — like, say, microSD cards and replaceable batteries.


Verizon resurrects zombie cookie, combines personal tracking with AOL network

When news of Verizon’s unblockable “super cookie” broke last year, the wireless provider was initially defiant and refused to stop tracking users. Eventually, the company began offering opt-out options, after nearly six months, but only grudgingly. The telco obviously learned nothing from this series of events, because it’s now openly acknowledging it will share personally identifiable information about subscribers with its newly acquired AOL division.

First, some context. Most companies rely on privacy policies that at least pay lip service to the idea of anonymity. Despite the fact that research has shown how easy it is to turn anonymized information back into PII (personally identifying information), companies at least make a token effort to strip out details like names and addresses. Under the terms of Verizon’s new privacy policy, even this paper-thin protection is abolished.

The new policy states (emphasis added):

The best advertising is for something you might actually want, and that is what we want to give you. To help make this happen, starting in November, we will combine Verizon’s existing advertising programs–Relevant Mobile Advertising and Verizon Selects– into the AOL Advertising Network…”


How Verizon’s supercookie works.

The Relevant Mobile Advertising program uses your postal and email addresses, certain information about your Verizon products and services (such as device type), and information we obtain from other companies (such as gender, age range, and interests). The separate Verizon Selects program uses this same information plus additional information about your use of Verizon services including mobile Web browsing, app and feature usage and location of your device. The AOL Advertising Network uses information collected when you use AOL services and visit third-party websites where AOL provides advertising services (such as Web browsing, app usage, and location), as well as information that AOL obtains from third-party partners and advertisers.

We do not share information that identifies you personally as part of these programs other than with vendors and partners who do work for us.

So far, so bad. Verizon acknowledges that it uses PII to show you advertising and that it combines this information with AOL’s network. According to Pro Publica, AOL’s advertising platform runs on 40% of websites today. But wait — it gets worse:

These programs use online and device identifiers, including AOL browser cookies, ad IDs from Apple and Google, and one created by Verizon, known as a “Unique Identifier Header.” When the Verizon and AOL programs are combined, this Verizon identifier will be inserted in certain Web traffic that is sent only to Verizon companies (including AOL) and to certain partners. These partners will be authorized to use the Verizon identifier only as part of Verizon and AOL services.

You’re the product whether you pay for the service or not

There’s a common saying in the tech industry — “If you aren’t paying for the product, you are the product.” Verizon’s advertising program puts the lie to such pithy remarks. If you’re a Verizon customer, it’s because you pay a significant amount of money for the privilege — and in return, Verizon is now claiming it has the right to track you across all of your devices and share that information with anyone it chooses to work with. There’s no guarantee that it doesn’t include third-party data brokerages like Acxiom. AOL, in fact, notes that it works with Acxiom, Session M, Taboola, and Tapad on an ongoing basis, which means you should assume they have your data.

If you want to opt out of being tracked in this fashion, you can contact Verizon at 1-866-211-0874 or manage your privacy choices from within your account settings. You will also need to opt out of AOL’s interest-based advertising tracking as well. If you are a member of Verizon Selects, you will have to opt out of that program separately (again, this is handled in Verizon’s privacy section).


Microsoft Bing Predicts and the future of gambling

Like an 800 pound gorilla flailing wildly in a Victorian tea house, artificial intelligence has been disrupting one industry after another of late. Now the latest group to feel the burn is the gambling consortiums in Las Vegas. Microsoft’s AI engine, Bing Predicts, made headlines recently by beating the Las Vegas odds in predicting winners for week one of the NFL season. Its previous successes are even more breathtaking, correctly predicting the outcomes of all 15 games in the 2014 Brazil World Cup knockout round and almost all the results of the 2015 Academy Awards, including the winners of best picture, best director, best actor, and best actress. Which is all to say Microsoft’s AI is turning out to be an incredibly good gambler and the ramifications will go well beyond the world of sports betting.

Let’s take a look at how Bing Predicts was able to outwit the best sporting minds in Las Vegas, and in the process, explore how AI is poised to upend the world of professional gambling. The basic principle driving Microsoft’s success at gambling rests on the “wisdom of the crowd.” In regards to predicting NFL winners, not only does the AI algorithm take into account such diverse variables as a team’s previous margins of victory, player statistics (rushing yards and passing yards for example), stadium surfaces, weather conditions, and so on, the secret sauce that seems to give it an edge over the other experts is the ability to quantify aggregate sentiments on the social web.

Walter Sun and the Bing Predicts team at Microsoft

Walter Sun and the Bing Predicts team at Microsoft

By tapping into social media and digesting the opinions of thousands, if not millions, of Twitter and Facebook users, the AI can pick up intangibles that defy even the most hardcore of human statisticians. For instance, the model might detect a rumor among Twitter users that the Patriots starting quarterback just had a fight with his wife in the wee hours before Sunday’s game and hence is less likely to be at the top of his form. While such rumors may prove to be unfounded, they have a core of truth enough of the time that they give the model a statistical advantage. In precise terms, Walter Sun, who heads up the Bing Predicts team, found that analyzing this so-called “wisdom of the crowd” actually increases the accuracy of their predictions by 5%.

While 5% may seem like a small amount, when it comes to beating the Las Vegas odds, if one is consistently beating the experts 5% of the time, that equals a fortune in gambling earnings and a troubling turn of events for Vegas bookies. This raises a real question: can professional sports gambling survive in a world where a Silicon Valley corporation holds the highest card in the deck? But if Vegas bookies think they have a lot to worry about, they are just one among many. A whole slew of industries are essentially gambling houses, and any algorithm that could beat their models would pose a major threat to their very existence.

Notable among these are the fields of insurance and commodities trading. If Microsoft or another one of the Silicon Valley behemoths that are developing cutting edge AI can leverage their advantage in the prediction business to outgun the industry leaders in some of these fields, they wouldn’t have to wait long to achieve supremacy in the market. Brace yourself: We may be headed towards a world dominated by a handful of tech corporations vying with each other to develop the best AI prediction algorithm.