It’s probably a decade or so since I first tried out a virtual reality headset. It was a pretty unforgettable experience: like getting drunk, putting on an antique brass diving helmet, climbing onto a spinning merry-go-round and trying to play a Commodore 64 game.
But a lot less fun.
VR isn’t like that anymore: the latest headsets are light(er), the leap in processing power means the images are crisp and there’s little vertigo. After many, many false dawns it’s finally a technology that is almost ready for the mass market. Sure, gaming and (adult) entertainment will lead the way, as they usually do, but there are going to be plenty of other uses too.
Give it another 10 years and that weekly catch-up with the boss (or your mum) won’t be on the phone it will be in VR.
Still, the technology isn’t without it’s controversies. Last week an inadvertently hilarious/terrifying image of Mark Zuckerberg striding past rows of oblivious journalists strapped into their VR headsets launched a thousand memes about our future VR overlord, which quite overshadowed what he was actually at MWC to say.
Zuckerberg’s argument: that far from being an isolating experience and the harbinger of the end of human civilisation as some have feared, VR is actually a social experience.
Going back 10 years we could share text online he said, then photos and now video, and “pretty soon we are going to live in a world where everyone has the power to share and experience whole scenes as if you there, right there in person,” he said.
“VR is going to be the most social platform,” he said, adding that when his daughter takes her first steps he wants to be able to record it in a 360 VR video. Facebook is assembling a team to work on social VR.
Most new technologies are accused of making us less sociable (all those people on your train every morning buried in their smartphones) but that’s often unfair (those people on the train might be being very social – texting, tweeting, Facebooking or chatting – but just not with you).
VR will come in for exactly the same kind of criticism so it’s interesting that Zuckerberg is trying to get out in front of it, if not entirely succesfully. Yes VR may be social, but social on who’s term?
What makes VR different – and scary – is the immersive nature: you can’t glance away from a headset in the way you can from a smartphone: it’s an all-or-nothing deal which is exactly what made that photo of Zuckerberg and the sightless hacks so powerful. Perhaps in the world of VR the headset-less man is king.
VR has been a joke technology for so long that we haven’t really thought through the implications of it. Now it is really happening we have to start thinking about the right and wrong ways to use it, and what it means for consumers, businesses and society. But I think Zuckerberg is right – VR is the next big platform: now we have to decide just what that means.
A Japanese company is opening the first farm run almost entirely by robots.
Spread, a sustainable vegetable producer, plans to allocate almost all of the labor involved in growing thousands of heads of lettuce at its new facility in Kyoto to robots. The only task that will still be performed by humans is planting the seeds.
The move comes amid rising concerns in Japan that an aging population will soon result in a potentially crippling labor shortage. The government and tech industries are pushing the development of robotics to take over from humans.
It’s an interesting inversion of the more typical conversation that surrounds robots and jobs–the perception that robotic automation is stealing jobs from humans.
Spread’s robot farm will be watched closely outside Japan, as well. There’s growing consensus that our current food production paradigm is unsustainable. The U.N. estimates that by 2050 we’ll need to sustainably produce 70 percent more food by calories than we do today to keep up with global population growth.
It’s a tricky problem because there will be less land for farming as more people arrive on the planet. Farmers are also aging globally as younger generations migrate to cities. That’s largely because a productivity boom over the last century has kept food prices low, which makes farming unattractive economically.
But global agricultural productivity is suddenly slowing for the first time in decades. No one is quite sure why, but it’s likely a systemic problem related to the rise of monocultures and the overuse of fertilizers, which add harmful salts to soils.
All of that suggests the need for more efficiency in the way food is farmed in order to maximize yield. Robots are likely one part of the solution. Spread estimates that its new farm will allow it it to more than double its lettuce production from 21,000 head of lettuce per day to 50,000 per day. The company plans to raise that number to as much as a half-million daily within five years using robots and other automated growing systems.
The majority of Spread’s robots will be articulated arms that work around a conveyer belt system, which will run throughout the 4,400 square meter farm. The installation will look more like a logistics fulfillment center than the family spread, with floor-to-ceiling shelves growing pesticide-free lettuce.
The robotic arms will transfer and replant seedlings and perform all harvesting. The smart farm will automatically optimize temperature, humidity, and CO2 levels, as well as recycle 98% of water and control lighting for optimal growth.
This is very likely the future of farming. Other Japanese technology firms like Panasonic, Toshiba, and Sharp are also experimenting with robotic farming solutions, including automated pickers for easily-bruised fruits and vegetables.
The social media network has launched a new feature in the United Kingdom to offer support to users hesitating with their own lives. Facebook announced a new feature called Suicide Prevention tool developed with support from the Samaritans charity. The instrument asks Facebook users to flag and report posts from their friends that cause concern. The company also allocated a special team to review such posts and provide help options sent to the publishers of the posts deemed to be struggling. The tool works in the following way. When the reported user next logs in, they would see a message saying: “Hi, a friend thinks you might be going through something difficult and asked us to look at your recent post.” The message will also offer the user to talk to someone about it – either a friend or helpline worker. They can also be sent tips and support directly. Of course, you can ignore all the options.
Facebook announced that keeping its community safe is its most important responsibility; this is why the company cooperates with various organizations including Samaritans in order to develop the tools like this. In this case, Samaritans told Facebook that connecting with people who care can really help those who are struggling to cope – whether offline or online. Today, more and more people only reveal what they feel online, in social network, “leaving a note” or sending a signal in the hope someone will understand and help them. Facebook is normally used to connect with friends and family, and the company is doing its best to evolve the support, resources and advice available to users in distress, along with their concerned friends and family members. For example, over 50% of the UK’s population is on Facebook, which makes it a powerful resource.
That tiny dongle plugged into your USB port and paired with your wireless keyboard or mouse isn’t as monogamous as it pretends to be. For millions of cheap peripherals, those innocent-looking radio receivers may be carrying on a sly, long distance relationship—letting an antenna-wielding intruder silently type malicious commands on your PC.
That’s a new warning from researchers at the Internet of things security firm Bastille, who released an advisory today that seven different companies’ wireless keyboards and mice are vulnerable to an exploit they’ve dubbed “mousejacking.” The attack—which affects a broad collection of devices sold by Logitech, Dell, Microsoft, HP, Amazon, Gigabyte and Lenovo—lets an interloper inject mouse movements or keystrokes at a rate of a thousand words per minute from an nearby antenna, even when the target device is designed to encrypt and authenticate its communications with a paired computer.
“With about fifteen lines of code, you can take over a computer more than a hundred yards away,” says Chris Rouland, Bastille’s CEO, who previously founded the hacking firm Endgame, a US government contractor. In their tests, Rouland and Bastille researcher Marc Newlin used a $12 Geeetech Crazyradio USB radio dongle attached to a laptop running their exploit code to pair with the victim devices. They tested as far as that hundred-yard range, though they found that the attack was more reliable with a more powerful Yagi antenna and believe it could likely be extended further. They also built a radio-enabled Nintendo controller capable of running their attack software, which they plan to show off at the RSA conference in San Francisco next week. Rouland points out that the exploit, which affects devices that use a little-studied proprietary radio protocol rather than Wi-Fi or Bluetooth, leaves even PCs that have been “airgapped”—isolated from the Internet—vulnerable if someone has plugged in a wireless keyboard dongle.
“We can compromise an airgapped network, going in through a different frequency protocol, directly to the USB port,” he says.
How It Works
Bastille’s mousejack attack doesn’t take advantage of one single vulnerability, but instead a collection of distinct problems in the firmware of wireless devices that use chips sold by the Norwegian firm Nordic Semiconductor. Nordic chips are capable of encryption. But unlike standard Bluetooth chips, the Norwegian firm’s cheap, low-power shortrange radio communications chips require that vendors write their own firmware to implement that encryption and secure the connection between computers and peripheral devices. The result, Bastille’s researchers say, is that many of the affected companies failed to take advantage of Nordic’s encryption option, allowing the dongles that receive those communications to accept keystrokes from another device using the same radio protocol. Most of the vulnerable keyboards did encrypt their communications, the researchers say, but didn’t properly authenticate communicating devices; they would still allow another rogue device to inject unencrypted keystrokes over the same connection. “It’s like having an expensive deadbolt and leaving it unlocked,” says Rouland.
I don’t think people even understand that there’s firmware in the dongle connected to their mouse. Chris Rouland
The Bastille researchers claim that “more than a billion” devices are affected in total. They back up that questionable number by pointing to a 2008 press release from Logitech touting the shipment of its billionth mouse, but couldn’t point to a more recent count or one that distinguishes between wireless and wired devices. Given the number of companies whose products Bastille successfully attacked, however, the count of vulnerable mice and keyboards is likely high, possibly in the millions; Rouland says that in Bastille’s tests, they were able to spot vulnerable wireless device dongles in most office buildings they targeted with their antennae. “Once you start looking for that little dongle, you’ll see it everywhere,” Rouland says.
Injecting keystrokes on a target computer, of course, isn’t in itself a full compromise of the machine. The hacker would only have the same privileges as the person using the computer and wouldn’t necessarily be able to type his or her passwords. Rouland argues that the attack could quickly be used to download malware and take full remote control of a PC. But the computer would have to already be unlocked, a caveat that would likely require the attacker to be able to see his or her target’s screen.
What Manufacturers (And You) Can Do
When WIRED reached out to the affected vendors, some downplayed the severity of the mousejacking attack. “The vulnerability would be complex to replicate and would require physical proximity to the target,” a Logitech spokesperson wrote in a statement. “It is therefore a difficult and unlikely path of attack.” Lenovo wrote in a statement that the severity of the attack is “low” and claims without explanation that its range would be only 10 meters. Bastille’s researchers, who had initially range-tested their attack only on Logitech devices, responded to Lenovo’s claim by trying the long-range mousejacking attack on Lenovo devices, too, and say they were able to inject keystrokes from 180 meters away.1
But Dell, Logitech, Lenovo and Microsoft all acknowledged the vulnerability; Logitech says that with Bastille’s help it’s developed a firmware update for affected devices. Some Dell wireless keyboards use the same firmware, and the company tells WIRED those keyboards receive the Logitech update, too. Lenovo’s statement, on the other hand, admits it can’t update its vulnerable devices and instead will offer to replace them for any customer who asks. Microsoft’s statement notes only that the company “will proactively update impacted devices as soon as possible.” Amazon, Gigabyte, and HP have yet to respond to WIRED’s request for comment.
Now someone has finally taken a look and found that virtually all of (these devices) are vulnerable. That’s pretty scary. Samy Kamkar
The Bastille researchers aren’t the first to warn of the dangers of wireless keyboards and mice. Early last year, independent hacker Samy Kamkar released the code and specs for KeySweeper, a fake USB charger that can both eavesdrop on and inject keystrokes from certain Microsoft wireless keyboards. Kamkar’s device worked only on unencrypted keyboards, however. Bastille researcher Marc Newlin painstakingly tested this mousejacking attack on many more brands and models of devices. “It’s a bit of a wakeup call when it comes to the implementations of these protocols that are entirely proprietary and that no one has looked at before,” says Kamkar of Bastille’s research. “Now someone has finally taken a look and found that virtually all of them are vulnerable. That’s pretty scary.”
Unlike Kamkar in the case of his KeySweeper attack, however, Bastille has opted not to release its code for fear that mousejacking could be used for malicious purposes against unpatched—or unpatchable—devices. “This vulnerability will be out for 10 years,” estimated Rouland. “When was the last time people updated their router’s firmware? I don’t think people even understand that there’s firmware in the dongle connected to their mouse.”
But even without a public exploit from Bastille, other hackers could soon reverse engineer the attack and release it publicly. That means you should update or replace any vulnerable devices you use ASAP. Or better, says Kamkar, you should switch to more fully tested Bluetooth mice and keyboards. Or the safest fix of all, he suggests: give up on wireless peripherals altogether. “If you can go wired,” says Kamkar, “Go wired.”
Google made some notable news yesterday—by doing nothing.
Company representatives did not appear on stage in a lecture hall at the corporate headquarters of chipmaker Qualcomm. Typically, this would be no big deal. Google isn’t in the habit of doing such things. But earlier this month, Bloomberg reported that Google would somehow make its presence felt at Qualcomm HQ, in front of Qualcomm stockholders, offering its “stamp of approval” for a new kind of Qualcomm chip. That would have been a big deal.
Qualcomm is the world’s largest smartphone chip maker, but this wasn’t about smartphone chips. This was about chips for computer servers, the machines that deliver all those services and emails and other data to your phone from across the Internet. And if we’re talking about Google—the world’s largest Internet company—server chips are no small thing. But Google didn’t show.
According to one person familiar with the matter who asked not to be identified because he wasn’t authorized to speak to the press, Google was indeed slated to deliver some sort of message, but pulled out. It’s a back-and-forth that shows how important the chip market is to Google, how important Google is to the world’s chipmakers, and how both Google and all those chipmakers are so carefully working behind the scenes to tip the scales in their favor. That includes the chip industry’s behemoth: Intel.
In delivering its many services, from Search to Gmail to Maps, to tens of millions of people across the globe, Google operates a network of data centers stretching from Oregon to Finland to Taiwan. Google engineers custom design the tens of thousands of servers that drive these computing hives. And Google buys all the chips for these servers directly from the companies making them. Right now, that means Google buys an enormous number of chips from Intel, the chipmaker that dwarfs all others. And we mean enormous. In late 2012, Intel bigwig Diane Bryant told us that Google bought more server chips than all but five companies on earth.
That’s remarkable when you consider that everyone else on that list actually sells servers, including Dell and HP. Google builds servers only for itself. According to Shane Rau, an analyst with research firm IDC, Google now accounts for 5 percent of all server chips sold worldwide. Over the course of a recent year-long period, he says, Google bought about 1.2 million chips. So, if it looks like Google may start buying chips from someone other than Intel, industry insiders sit up and pay attention. Now that Google is serious about cloud computing—inviting the world’s business to run all their software on its state-of-the-art infrastructure—its already massive slice of the chip market will only grow. As more companies move onto the Google cloud, they’ll buy fewer servers from the likes of HP and Dell. Intel wants to keep that massive Googly slice of the market—while newer rivals like Qualcomm and Applied Micro and Cavium are looking to take it away.
But in the end, Google may take things in yet another direction. It may design its own chips.
Keeping Options Open
Google certainly has an interest in seeing other chipmakers challenge Intel’s dominance. Today, according to IDC, Intel controls 99 percent of the server chip market. If Google can buy chips from more companies, prices are bound to drop. Simply by flirting with other players, Google can encourage Intel to keep prices down. But in addition to lower prices, Google may have an interest in server chips that are more like those that Qualcomm makes for smartphones—that is, chips that consume extremely small amounts of power.
You may think Google would need big beefy chips to run its sweeping Internet empire, but the trick to running so vast an operation is finding ways of breaking tasks into tiny pieces and spreading them across many modest pieces of hardware. That way, one failure doesn’t really matter. The others can pick up the slack. Plus, this distributed model is more efficient. If you run an empire as big as Google network, you must keep costs—meaning power usage—as low as possible.
That’s why Qualcomm is making server chips more like energy-sipping smartphone chips than the old-school energy-gulping variety. It knows Google is interested. It also knows that other Internet giants like Amazon and Facebook are interested. Facebook has made no secret of this. That’s why Qualcomm wanted Google at its event: to show the rest of the Internet that the big dog was interested.
A Chip to Call Its Own
This isn’t the first time Google has flirted with alternative chip architectures. And for years, rumors have swirled indicating that Google may end up designing its own chips. That isn’t beyond the realm of possibility. After all, Google designs so much other hardware, from servers to storage systems to networking gear. And unlike Intel, chipmaker ARM will license its basic processor designs to anyone so they can further customize them. Qualcomm makes custom ARM designs. Google could, too.
Just this week, the rumors started flying again. Google contributions to an open source software project indicated that it has built its own chip, and some people got very excited. As it turns out, this is merely a chip for network interface cards—cards that connect servers to a larger network. Even if Google did design such a chip, it’s most likely a small development. It’s “unlikely to be much in the way of revolutionary architecture,” says JR Rivers, who once helped design networking gear at Google and now runs a networking startup called Cumulus. These less sophisticated chips are quite different from a CPU, the chip that represents the brain of a server. But it shows how Google thinks about data center hardware: It’s always looking to gain any advantage.
At the moment, server chips based on the ARM design are still maturing. Rau, the IDC analyst, says they account for less than 1 percent of the market, and at this point, he suspects, companies like Google are buying these chips only to experiment with them. “I don’t think these chips are in major volume in big cloud companies—yet,” he says.
So, Google is still very much dependent upon Intel for its processors—as much as Intel is dependent upon Google. That means Google must carefully balance the situation. It wants the Qualcomms of the world to succeed, to build ARM chips it can use in bulk. But it doesn’t necessarily want to tick off Intel. Thus, this week’s no-show. Right now, it’s status quo. But the balance may shift.
Congress has voted to make permanent a federal law that prevents states or localities from taxing Internet access.
The US Senate accepted the measure as part of a larger trade bill, which passed today on a 75-20 vote. Since the House has already passed a similar measure, the bill now heads to President Barack Obama for his signature.
There’s long been general agreement in Congress that taxing access to the Internet is a bad idea and shouldn’t be allowed. But permanent consideration of the tax ban was held up by some lawmakers, including Sen. Dick Durbin (D-Ill.), who wanted it to be passed together with the Marketplace Fairness Act, or MFA.
The MFA mandates that Americans must pay sales tax on all online purchases. It, too, has garnered majorities on both sides of the aisle, and a version was passed by the Senate in 2013—yet, it still hasn’t become law. Durbin reportedly dropped his opposition to the access tax once Majority Leader Mitch McConnell (R-Ky.) promised a vote on a new MFA later this year.
Durbin’s home state of Illinois is one of seven states that currently do tax Internet access, because their taxes were grandfathered in when the original tax ban was passed in 1998. The permanent ban passed today will sunset those taxes by 2020, which Durbin claims will cost his state as much as $390 million annually.
Of all the ways tech companies are gunning to drive virtual reality, Google may be working on something no one else has done yet: make the experience self-sufficient.
The search giant is developing a stand-alone headset, which would be the first VR goggle that isn’t powered by a smartphone, laptop or gaming console, according to a report Thursday by The Wall Street Journal that cited unnamed sources.
The move would represent a novel approach as a scrum of tech companies, including Facebook, Samsung and HTC, jostle to bring the nascent technology to a mass audience. Facebook’s $599 Oculus headset, available starting next month, requires a PC, while the offerings from HTC and Sony are also tied to a PC or game console. Samsung’s Gear VR requires a Samsung Galaxy smartphone.
Google’s device will reportedly have its own screen, outward-facing cameras and a high-powered processor based on chips from Movidius, a San Mateo, California-based startup that specializes in machine vision.
A Google spokeswoman declined to comment. Movidius acknowledged it has a “business relationship” with Google but wouldn’t say anything else.
Once mostly the dream of video game makers, VR now has the tech industry taking a wider view of the experiences it can offer, from virtual trips to the Alps and walks through the world’s most famous museums to visits to digital doctor’s offices.
Google’s attempt at something new shows how seriously it’s taking virtual reality. Last month, the tech titan created its own VR division, putting vice president Clay Bavor at the helm. Bavor used to be in charge of Google apps as well, but the company cleared away all his other responsibilities to focus squarely on virtual reality.
It’s a far cry from Google’s first foray into VR two years ago with Cardboard, a no-frills, made-of-paper headset design that uses a smartphone as the screen. Cardboard, developed by Bavor and his team, isn’t sold by Google — it licenses the design to others who create the $20 to $25 headset. Google will also release an updated version of that product to be made of plastic, according to the Journal. The headset is also expected to include chips and sensors.
Glass, Tango and beyond
If the reports are true, it seems as though Google’s approach could be similar to what it did with Google Glass, in the sense that this will be a fully formed piece of standalone hardware. Glass, experimental eyewear with a high-tech flourish, could connect with a phone but also worked on its own over Wi-Fi, avoiding a need to be tethered to any other device.
It’s an unusual approach for VR at the moment, especially since it feels like both mobile and PC-based VR efforts are aiming for a level of industry consistency. Would a stand-alone unit run Android, Google’s mobile software? Or would it be more of a head-mounted display that interfaces with phones, tablets and computers wirelessly?
There are reasons for Google to pursue this type of product. Cardboard needed to work with several types of smartphones, which meant screen sizes, resolutions and processors were inconsistent. Having all those things built into the headset itself would let VR content makers all step up to the same standard. That could help the whole experience look and feel a lot better.
Facebook’s Oculus is intended to work with higher-resolution screens with greater pixel density than most of the phones that Google Cardboard supports, allowing for crisper, sharper images and more impressive 3D graphics.
An optimized device might also reduce lag, allowing for a smoother, more immersive viewing experience. In contrast, Cardboard didn’t have a head strap for a reason — it’s laggier, lower-powered feel means it’s not intended to be used for extended sessions.
A new device would also help Google adopt position tracking, something missing from mobile VR right now. Current phone-based VR headsets let you turn your head, but they can’t track movement across a room. Cameras on the outside of the headset, in the manner of what Google is reportedly building, could be depth-sensing ones like those used by Project Tango, another next-gen Google effort that tries to pack more sensors into phones and tablets. (Movidius also worked on Project Tango, which makes position tracking the likeliest reason for this specific hardware redesign.)
Cameras can help map out a room in 3D, help someone walk around in virtual reality without hitting a wall in real life (HTC’s Vive headset uses an external camera like this), or help boost the reality of VR. It’s a key step to developing next-level virtual reality applications, and the hardware to make that possible simply doesn’t exist on everyday phones. A specially-built headset would be a logical move.
A stand-alone device does raise questions about what sort of processor it will require or how much battery life it will have.