All posts in “Microsoft”

New Xbox Wireless Adapter for Windows 10 is actually reasonably sized


Microsoft has redesigned the USB dongle used to connect your Xbox One controller to your Windows 10 laptop, and thank goodness for that. The formerly hulking black slab is now a much more modestly sized black slab, which is roughly the size of most flash drives, and won’t stick out like a sore, incredibly huge thumb on your notebook or PC.

The new adapter is $24.99 and ships starting on August 8 in the U.S., and is a full 66 percent smaller than the original, which I’m staring at right now in my office with a mixture of disgust and disdain. It also can connect to up to eight controllers at once, in case you’re a PC gamer with a deep love of local multiplayer.

There’s a little cap, too, and an indicator light to let you know when you’re in pairing mode and when you’re connected. It can also beam wireless stereo sound to your controller if you’re using a plugged in headset.

Sony released its own PC wireless adapter for its PS4 DualShock controllers not too long ago, and that was also a much more svelte device than Microsoft’s adapter, so it’s nice to see the Xbox maker get with the times.

Snap joins rivals Facebook and YouTube to fight terrorism


Snap Inc has joined the Global Internet Forum to Counter Terrorism, which sees consumer internet companies cooperating to stop the spread of terrorism and extremism online. Facebook, Google and YouTube, Microsoft and Twitter formed the GIFCT last month, and tomorrow it will host its first workshop with fellow tech companies plus government and non-governmental organizations.

The GIFCT started as an extension of the shared industry hash database that allows tech companies to share the digital fingerprints of extremist and terrorist content, such as photos and videos, so that once one identifies a piece of prohibited content, all the others can also block its upload. It’s almost like a vaccine program, where one company beats an infection, then shares how to produce antibodies with the rest of the group.

In identical blog posts published by Facebook, YouTube, Twitter and Microsoft, the GIFCT wrote “Our mission is to substantially disrupt terrorists’ ability to use the Internet in furthering their causes, while also respecting human rights.”

The first GIFCT workshop, held in San Francisco on August 1st, will host the United Kingdom Home Secretary Rt Hon Amber Rudd MP and United States Acting Secretary of Homeland Security Elaine Duke, plus representatives of the European Union, United Nations, Australia and Canada. The event’s goal is to formalize how the tech giants can collaborate with smaller companies, and what those companies would need as far as support to get involved.

In the coming months, the group’s goals include adding three more tech companies to the hash sharing program beyond new members Snap and JustPaste.it, get 50 companies to share their best practices for countering extremism through the Tech Against Terrorism project and plan four knowledge-sharing workshops.

Improving automated moderation and deletion of terrorist content is critical to preventing it from slipping through the cracks. While internet giants like Facebook typically employ thousands of contractors to sift through reported content, they often have to work extraordinarily fast through endless queues of disturbing imagery than can leave them emotionally damaged. Using shared hash database and best practices could relieve humans of some of this tough work while potentially improving the speed and accuracy with which terrorist propaganda is removed.

It’s good to see Facebook and Snap putting aside their differences for a good cause. While Snap is notorious for its secrecy, and Facebook for its copying of competitors, the GIFCT sees them openly sharing data and strategies to limit the spread of terrorist propaganda online. There is plenty of nuance to determining where free speech ends and inciting violence begins, so cooperation could improve all the member companies’ processes.

Beyond banishing content purposefully shared by terrorists, there remains the question of how algorithmically sorted content feeds like Facebook and Twitter handle the non-stop flood of news about terrorist attacks. Humans are evolutionarily disposed to seek information about danger. But when we immerse ourselves into the tragic details of any terrorist attack around the world, we can start to perceive these attacks as more frequent and dangerous than they truly are.

As former Google design ethicist Tristan Harris discusses, social networks know that we’re drawn to content that makes us outraged. As the GIFCT evolves, it would be good to see it research how news and commentary about terrorism should best be handled by curation algorithms to permit free speech, unbiased distribution of information and discussion without exploiting tragedy for engagement.

Microsoft’s Imagine Cup crowns its 15th winner, the X.GLU smart glucose meter for kids


Microsoft’s annual Imagine Cup, a global competition of high-tech student projects, reached its culmination this week accompanied by the usual pomp at the company’s headquarters in Redmond. The winner of the $100,000 grand prize and 365 days of bragging rights is X.GLU, a Czech team that created a custom smart glucose meter for children with diabetes.

Hundreds of teams competed worldwide, and 54 were selected as finalists, presenting their projects in Redmond. That was further reduced to a final 4, who presented to Macklemore, of all people, and a panel of expert judges. (Strangely, the finals ceremony wasn’t held, as it was last year, at the high school he and I both attended.)

The device is thinner than any glucose meter available now, has no buttons or interface to fiddle with, and can be customized like a smartphone with colored or printed cases. It’s controlled entirely from a smartphone app, and powered by an NFC connection — the disposable blood analysis strip doesn’t actually need any power, so only a little is used to transmit the small amount of data it produces.

Marek Novak, the 22-year-old hardware designer for the team, made it all using off-the-shelf parts to keep prices down. The other team members, Tomas Pikous and Barbora Suchanova, put together the cloud-based infrastructure that allows caregivers or emergency contacts to see the kid’s glucose levels or get alerts if, say, the kid doesn’t test on time.

In addition to the $100,000, the team gets a one on one mentoring session with CEO Satya Nadella, $125,000 in credit on Azure services and tickets to next year’s Build conference. Congrats to X.GLU for an innovative project taking on a serious problem.

Second place went to Oculogx, which created a warehouse navigation app for HoloLens — sounds kind of dry, but it’s pretty neat. It overlays directions and highlights items for human “pickers,” increasing their efficiency and reducing errors. One wonders how long such a technology would be applicable, however, considering the advance of robots in the warehouse labor space.

Third was ResCue, which equips drones with computer vision and navigation modules that let them be easily deployed to disaster areas, where they can perform zone searches for people who need help. And fourth was NeuroGate, which uses Kinect to do gait analysis, for diagnosis or monitoring of neurodegenerative diseases.

The Imagine Cup is always a pleasure to cover, since the young creators involved haven’t yet developed that thick shell of cynicism and pragmatism that plagues so many startups. You can read more about the (many) other teams that competed over at the competition’s microsite.

Featured Image: Microsoft

Microsoft’s second-generation HoloLens will include a dedicated AI coprocessor


Microsoft has revealed that it is preparing to equip the second version of its HoloLens VR headset with its own AI processing capabilities to enable more features and services.

The firm said that it is customizing the current processor within the headset with an AI coprocessor that will allow HoloLens to analyze data without needing to use the cloud. That means faster processing times, and more mobility for the device since it won’t always need to be online.

Microsoft is designing the silicon itself because it believes that is the only way to unlock future uses and services for both augmented reality and mixed reality. Speaking at Microsoft’s Build event in May, CEO Satya Nadella gave a number of examples for future AI uses which include industrial working scenarios. Faster processing and greater mobility, just two benefits of a custom AI chip, could be critical factors in those scenarios.

“This is the kind of thinking you need if you’re going to develop mixed reality devices that are themselves intelligent. Mixed reality and artificial intelligence represent the future of computing, and we’re excited to be advancing this frontier,” Microsoft said of the chipset plans.

Microsoft isn’t the only one creating its own chips. Google built its own AI processor to power bots while Apple is said to be secretly developing a dedicated chip for the iPhone. The other approach to unlocking future potential is to create lightweight neural networks which remove the strain from device processors as Google and Facebook are doing.

Microsoft’s Path Guide is an unconventional approach to indoor navigation


Indoor mapping is one of those stubborn problems that seems to only have solutions that involve a great deal of money and infrastructure: beacons, lasers, emitters, scanners… who wants to install those in malls and office complexes across the world? Microsoft Research has produced an indoor navigation app that keeps things simple, utilizing existing sensors and the fact that people already go most of the places you want to go.

Path Guide is an app currently available only on Android that focuses on the navigation part of indoor mapping. It makes no attempt to visualize the entire volume in which the user is traveling, but trusts other users to create “traces” to and from static locations.

It uses the sensors already in the phone, from accelerometers (to count steps) to magnetometers (to sense the general area where one is starting). No need for GPS, wireless beacons or anything like that.

For example, one trace might take you from the front entrance of the office building to a certain meeting room on the 12th floor. Someone would open the app while in the lobby, start the trace then walk normally to the room. The next person would follow that trace, which has been processed and rendered into basic navigational stages: “in 20 steps, take a right,” that sort of thing.

The trace creator can augment the path with notes, images and voice recordings in case there are PINs, secret doors or spoken passwords. Traces can be capsulized and sent in emails as well, for people who don’t have the app. It’s not quite augmented reality, but I’m using that tag anyway.

Naturally this will be helpful for people trying to find the Orange Julius at an unfamiliar mall, but the benefit is more tangible for people with impaired vision. Directions at this granularity are hard to come by and this would be hugely helpful for a blind person navigating a location they’ve never been. Same for getting around when you don’t speak the language.

The project has been in the works for nearly two years, led by Yuanchao Shu, Börje Karlsson, Yiyong Lin and Thomas Moscibroda. It’s still a prototype, and even Microsoft’s own post says it “admittedly still has rough edges,” but maybe you can help with that — feedback is appreciated.