I Tried Google Maps’ Experimental Walking Directions of the Future

https://gizmodo.com/i-tried-google-maps-experimental-walking-directions-of-1833225629

Augmented reality is coming to Google Maps.
Image: Google

The perfect scenario for Google’s new Visual Position System, or VPS, is one you’ve probably been in already: Emerging from a subway station in a strange city. You’ve got your route planned out on your phone, but you don’t know which way is which, or the names of the roads—if you set off in the wrong direction initially, you can lose a lot of time retracing your steps.

Way back at last year’s Google I/O conference we caught a glimpse of this new augmented reality layer coming to Google Maps. It’s designed to overcome the limitations of GPS in heavily built-up areas and remove confusion when you’re following walking directions. It’s not publicly available yet, but we’ve been able to try it, and here’s what’s coming.

Meet the Visual Positioning System, or VPS.
Image: Google

VPS pulls its data from the buildings and landmarks around you rather than relying on satellite location alone. By combining VPS with Street View data and some machine learning, Google Maps can work out where you are (and where you should be headed) even if your GPS lock has you placed three blocks north.

It does have some quirks in it that you might not have spotted from Google’s brief demo last year. It’s also worth pointing out that it’s still labeled as an Alpha release—you should expect plenty of bugs at this stage, and plenty of improvements to come down the line.

If you’ve got access, the AR mode is built right into Google Maps.
Screenshot: Gizmodo

For access on Android or iOS at the moment, you need to be one of Google’s Local Guides at level 5 or above (that’s someone who leaves a lot of reviews and feedback on Google Maps). As yet, there’s no word on when it’s going to go live for everyone, but here’s how it works from our testing.

Fire up some walking directions as normal, and alongside the usual Start button you’ll also see a Start AR option—choose this to jump straight into augmented reality mode. You can also switch to AR later on your journey by tapping the little cube icon (bottom left) when you’re navigating normally.

The AR prompts are hard to miss as you wander around.
Screenshot: Gizmodo

Open up the AR interface, hold your phone in front of you, and Google Maps asks you to point your phone’s camera at the buildings opposite. This works separately to the GPS lock, and helps the app get its bearings when GPS might not be reliable.

Most of the time, the scene is recognized in a second or less, and you’re then into the AR mode proper. We did come across a few hiccups when buildings were particularly close (on narrow streets) or far away (on park greens) but by and large it was reliable enough—and remember this is still at the alpha stage, so we’re not even at beta yet.

The AR mode recognizes buildings and directs you accordingly.
Screenshot: Gizmodo

When Google Maps knows where it is, you can really start to make use of the augmented reality features. Directions appear as either a discreet blue dot at the side of the screen, a blue sign confirming you’re on the right road, or (when you get close to a turn) huge animated arrows pointing you down the next road you need to make a turn on.

This being AR, these giant floating pointers move around as you move yourself and the camera, and they work really well at helping you work out where you should be going. The AR removes all doubt about whether you’re facing the right way or whether you’re at your turn, and having the map also on screen underneath (in a smaller semi-circle) makes the directions even easier to follow.

You’re encouraged to put your phone down and can enable AR mode with a tilt.
Screenshot: Gizmodo

When you need to be going straight ahead, again this is clearly shown in the app, and street names are included so you can double-check them against signs. When you get close, you’ll see your destination marked with a red Google Maps pin.

This isn’t a mode you keep enabled and open all the time though—it’s something you activate every few minutes to get your bearings again. Tilt the phone down, and you go back to the normal view; tilt it back up, and the AR mode comes back into play. (You can adjust this behavior in the settings for Maps).

The AR mode is still in development and doesn’t always work.
Screenshot: Gizmodo

The app itself will warn you not to keep your phone up in front of you for too long—the warning appears after only a couple of seconds or a few steps of walking, and it’s really not for extended use. The idea is you stick to the traditional navigation view and turn to AR only when there’s some confusion about where you’re headed next.

In fact, it’s more for when you’re standing still rather than when you’re walking. Considering how unreliable GPS can be when you’re stuck between giant skyscrapers or other towering urban developments, we reckon this is something we’re going to be turning to a lot in the future.

There’s no sign of the fox from Google’s demo yet.
Image: Google

Think of it as an added bonus to walking navigation, rather than an alternative mode you’re going to spend a lot of time in: It’s something you’ll fire up at tricky intersections, or when you’re coming out of the subway, or when your phone’s GPS lock is getting seriously confused about where you are.

Unfortunately, there’s no sign of the fox yet—the digital creature that appeared in Google’s demo at I/O 2018. Google must still be thinking about how best to guide people using the new technology, but it’s already impressive, and something you should switch on if you find it pushed out to your phone.

via Gizmodo https://gizmodo.com

March 23, 2019 at 08:06AM

MIT’s AI can train neural networks faster than ever before

https://www.engadget.com/2019/03/22/mit-ai-automated-neural-network-design/

In an effort "to democratize AI," researchers at MIT have found a way to use artificial intelligence to train machine-learning systems much more efficiently. Their hope is that the new time- and cost-saving algorithm will allow resource-strapped researchers and companies to automate neural network design. In other words, by bringing the time and cost down, they could make this AI technique more accessible.

Today, AI can design machine learning systems known as neural networks in a process called neural architecture search (NAS). But this technique requires a considerable amount of resources like time, processing power and money. Even for Google, producing a single convolution neural network — often used for image classification — takes 48,000 GPU hours. Now, MIT researchers have developed a NAS algorithm that automatically learns a convolution neural network in a fraction of the time — just 200 GPU hours.

Speeding up the process in which AI designs neural networks could enable more people to use and experiment with NAS, and that could advance the adoption of AI. While this is certainly not uncomplicated, it could be a step toward putting AI and machine learning in the hands of more people and companies, freeing it from the towers of tech giants.

Source: MIT News

via Engadget http://www.engadget.com

March 22, 2019 at 02:03PM

AT&T’s “5G E” is actually slower than Verizon and T-Mobile 4G, study finds

https://arstechnica.com/?p=1478303

Screenshot from an AT&T commercial showing text that reads,
Enlarge /

Screenshot from an AT&T commercial.

AT&T’s “5G E” service is slightly slower than Verizon’s and T-Mobile’s advanced 4G LTE networks, a study by OpenSignal has found.

As Ars readers know, AT&T renamed a large portion of its 4G network, calling it “5G E,” for “5G Evolution.” If you see a 5G E indicator on an AT&T phone, that means you’re connected to a portion of AT&T’s 4G LTE network that supports standard LTE-Advanced features such as 256 QAM, 4×4 MIMO, and three-way carrier aggregation. All four major carriers have rolled out LTE-Advanced. But while Verizon, Sprint, and T-Mobile accurately call it 4G, AT&T calls it 5G E.

Sprint sued AT&T, alleging that AT&T is gaining an unfair advantage by making false and misleading claims to consumers.

AT&T’s network name change may well trick consumers into thinking they’re getting better service than a 4G operator, but they aren’t. We already knew that 5G E has no technological advantage over LTE-Advanced, because they are the same thing with different names. But actual speed tests could reinforce that point.

Such data now comes from an OpenSignal report that is scheduled to go live at this link today at 9am ET. After comparing user-initiated speed tests from more than 1 million devices, OpenSignal found that AT&T’s “5G E” phones get average speeds of 28.8Mbps, which is less than T-Mobile’s 29.4Mbps and Verizon’s 29.9Mbps but higher than Sprint’s 20.4Mbps. The AT&T average of 18.2Mbps on non-advanced LTE was also slightly behind T-Mobile and Verizon but ahead of Sprint:

For each network, this shows the average speed of LTE-Advanced ("5G E") phones versus the average speed of other 4G phones.

For each network, this shows the average speed of LTE-Advanced (“5G E”) phones versus the average speed of other 4G phones.

OpenSignal

In other words, AT&T’s 5G E speeds are typical LTE-Advanced speeds. It is a real upgrade over the older portions of AT&T’s 4G network, but that’s true for all four carriers.

“Our analysis shows that AT&T customers with 5G E in their status bars are receiving up to a 60 percent boost in speeds over AT&T customers without it,” OpenSignal CEO Brendan Gill said in a statement to Ars. “Of course, our analysis also shows that the same is true for the equivalent Verizon and T-Mobile customers even though they don’t see a ‘5G E’ label on their device. Bottom line, if one service is offering a meaningful boost over another, it probably should be labeled differently, just not with a name that confuses customers.”

Crowdsourced testing

OpenSignal data is based on crowdsourced speed tests that can be performed by anyone using OpenSignal’s apps for iPhone and Android. OpenSignal told us today’s report is based on tests on 1,057,522 devices nationwide, across all four carriers, between January 28 and February 26.

The tests don’t automatically distinguish between LTE-Advanced and regular 4G LTE networks. But OpenSignal is able to distinguish between phone models, and it compared the phones that AT&T says are 5G E-capable to those that aren’t. The resulting data thus compares LTE-Advanced phones to non-LTE-Advanced phones, regardless of whether the tests were conducted in a location where the network supports LTE-Advanced features.

When contacted by Ars, AT&T argued that this limitation invalidates the results.

“OpenSignal’s note reveals their methodology is flawed,” AT&T told Ars. “Speed-test data purporting to show the ‘real-world experience of 5G Evolution’ without verifying the capable devices were tested in a 5G Evolution coverage area as shown by the indicator does not accurately represent the 5G Evolution user experience.”

But all four carriers were measured the same way, and the resulting data may reflect the breadth of each carrier’s LTE-Advanced deployment. Carriers that have deployed more LTE-Advanced coverage across the US would likely get higher speed-test results than carriers that haven’t upgraded their networks to the same extent.

We don’t know exactly how much LTE-Advanced coverage each carrier has. But AT&T has repeatedly boasted about the size of its 5G E/LTE-Advanced footprint, saying it covers more than 400 markets across the United States. If AT&T’s 5G E was significantly faster or more widespread than other carriers’ LTE-Advanced service, the OpenSignal tests probably would have found a big speed difference in AT&T’s favor. Instead, the results showed higher average speeds for Verizon and T-Mobile.

AT&T offered no further comment.

The new results are consistent with OpenSignal’s previous testing. OpenSignal’s overall data on consumer 4G experiences, released in January, showed that AT&T lagged behind Verizon and T-Mobile in 4G availability, video quality, and both download and upload speeds.

OpenSignal’s testing doesn’t yet tell us what speeds consumers can expect when they get real 5G. But the data shows “the extent to which LTE, or 4G, networks have improved since LTE’s original launch,” OpenSignal wrote in today’s report. Deployment of LTE-Advanced technology has created “a much faster experience than the initial version of 4G that was launched back in 2009-2011” but no real advantage for AT&T over its top rivals.

If you’ve been following the AT&T 5G E story, the OpenSignal results won’t surprise you. But if you know anyone who’s confused by AT&T’s misleading marketing, now you’ve got some data to help clear things up.

via Ars Technica https://arstechnica.com

March 22, 2019 at 08:03AM

FCC has to pay journalist $43,000 after hiding net neutrality records

https://arstechnica.com/?p=1478621

Shredded documents with a magnifying glass and the words,

The Federal Communications Commission has settled a case over its refusal to comply with a public records request, agreeing to pay $43,000 to a journalist who sued the commission.

Freelance writer Jason Prechtel filed a Freedom of Information Act (FoIA) request with the FCC in mid-2017, asking for data that would identify who made bulk comment uploads in the proceeding that led to the repeal of net neutrality rules. Prechtel was trying to research comments that were falsely attributed to people without their knowledge.

The FCC didn’t comply with the request and allegedly didn’t even approve or deny the FoIA request within the legally allotted timeframe, so Prechtel sued the commission in September 2017. One year later, a US District Court judge presiding over the case ordered the FCC to stop withholding certain records sought by Prechtel, although the ruling didn’t give Prechtel everything he asked for.

A settlement agreement filed in court this week says the FCC agreed to pay Prechtel $43,000 to cover his attorneys’ fees and court costs. Chairman Ajit Pai’s FCC did not admit any wrongdoing, but the settlement has resulted in the case being closed.

Comment fraud under investigation

As Gizmodo noted in a story about the settlement yesterday, “The data Prechtel ultimately obtained through the case formed the basis of a Gizmodo report last month—which he coauthored—that revealed how investigators had linked various entities, including a prominent Washington, DC, publication, to potentially millions of fraudulent comments submitted during the 2017 net neutrality rollback.”

Those investigators were from the New York attorney general’s office, which also faced FCC stonewalling when it sought data including the same comment records obtained by Prechtel. The AG’s office eventually determined that up to 9.5 million comments were submitted using stolen identities, and it’s still investigating. The FBI is also reportedly investigating the use of stolen identities in public comments on the FCC’s repeal of net neutrality rules.

The Gizmodo/Prechtel story last month said that API logs obtained through the FoIA lawsuit provided details on each time an organization submitted comments using the FCC’s API system. “What’s more, [the logs] include the IP addresses of the uploaders themselves, as well as timestamps that record, down to the millisecond, precisely when floods of comments came pouring in from any given source,” the article said.

CQ Roll Call’s advocacy arm was among those using the system to submit comments supporting repeal of net neutrality rules. CQ Roll Call was working on behalf of a client, which the Gizmodo/Prechtel story says may have been a conservative nonprofit called the Center for Individual Freedom (CFIF).

CFIF arranged for the upload of anti-net neutrality comments that said, “The unprecedented regulatory power the Obama Administration imposed on the Internet is smothering innovation, damaging the American economy and obstructing job creation.” CFIF has denied submitting comments under people’s names without their knowledge.

Net neutrality advocacy groups such as Fight for the Future also used the FCC API system to demonstrate support for keeping net neutrality rules.

In a separate case that’s still ongoing, the FCC refused a New York Times request for records that the Times believes might shed light on Russian interference in the net neutrality repeal proceeding. The FCC last week asked a US District Court judge for a summary judgment in that case.

via Ars Technica https://arstechnica.com

March 22, 2019 at 11:02AM

The Gamevice Has Changed the Way I Play (PC) Games

https://theinventory.com/the-gamevice-has-changed-the-way-i-play-pc-games-1833390456

The Gamevice comes in many varieties for different phones and tablets, and the phone versions fold up nicely for portability.
Photo: Whitson Gordon
Best Gaming DealsThe best deals on games, consoles, and gaming accessories from around the web, updated daily.   

I’m a PC gamer through and through, but I rarely game at my PC anymore. Instead, I play PC games on my iPad, with the help of a brilliant controller called the Gamevice.

That’s right, I said PC games. Despite being designed for mobile gaming, the App Store and Google Play ecosystem is in a tough spot. Both are flooded with free-to-play nonsense and half-assed spinoffs of console games, with quality premium titles dwindling in popularity. But the utility of these controllers goes far, far beyond App Store games and Fortnite, and it’s why the Gamevice is one of my favorite gaming accessories around.

The Gamevice Basically Turns Your Phone or Tablet Into a Nintendo Switch

A vice-like grip allows you to play anywhere, without having to prop up your phone or tablet.
Photo: Whitson Gordon

There are plenty of mobile-focused gaming controllers out there, from the $50 SteelSeries Nimbus to the more Xbox-like PXN Speedy which allows you to clip in your phone. But these are bulky, and in many cases—like the Nimbus—they assume you’re going to prop up your device on a table.

I’ve never found that to be a particularly easy or enjoyable experience, which is why the Gamevice is so alluring. Rather than making yet-another-standalone-gamepad, Gamevice created a two-piece controller held together by a rubber strip, allowing you to “dock” it on the sides of your phone or tablet—much like the Nintendo Switch. They have versions for the iPhone, Galaxy, and Pixel line of phones, which fold up neatly and fit right in your bag, as well as versions for different iPad models that are a bit less portable, but give you a great big screen for your games. They’re more expensive than typical gamepads, ranging from $60 for the phone models to $100 for the iPad models, but boy are they worth it.

All Gamevice models plug into the USB or Lightning port on the bottom of your device, ensuring a reliable hardwire connection instead of Bluetooth, and negating the need to charge your phone and your gamepad. Each Gamevice has a USB or Lightning pass-through port for charging your phone, as well as a traditional headphone jack—huzzah! Each model is a bit different, and has its own benefits and quirks—the iPhone version has a little slider to ensure your phone fits in perfectly, while the Pixel version stretches to auto-size itself to your phone. The Pixel version also uses a more squared-off design with black-and-white buttons that are a bit louder and clickier than the already-kind-of-loud iOS versions, plus slightly different shaped analog sticks and an extra menu button. The button noise is probably my only real gripe with the Gamevice, as it’s a bit annoying to other people in the room when I’m playing, though the tactile feedback is nice.

The Pixel version auto-sizes to your device, whether you have the regular or XL phone.
Photo: Whitson Gordon

Despite their minor differences, though, all of the models I’ve tested have worked quite reliably once your phone or tablet recognizes them. My iPad Gamevice sometimes requires reconnecting when an app doesn’t register its presence, but I’m pretty sure this is iOS’ fault and not Gamevice’s. Ultimately, the vice design alone makes it the best game controller for phones and tablets.

It Isn’t Just for Mobile Games: Play AAA Titles on Your Phone and Tablet, Too

That’s all fine and dandy, but if you aren’t a fan of mobile games, you can still make use of the Gamevice.

Here’s the real reason I use this thing: I’m tired of playing games at my computer. I sit in a desk all day, and it’s really hard to motivate myself to sit at a desk all evening too, even for fun. I want to melt into the couch like the T-1000 after a long hard day and get sucked into another world. I could get a console, but then I’d have to fight my wife for the TV—a battle which I’d surely lose every time. I cycled more iOS games into my backlog, and some are quite decent—there’s a good crop of indie titles and older console ports on mobile, not to mention a few good AAA options like République (despite its minor flaws) and Fortnite (which just received gamepad support for its mobile version). You can see all the Gamevice-compatible games on Gamevice Live, but chances are, you’ll eventually be left wanting more like I did.

With apps like Moonlight, Steam Link, Shadow, and others, you can play AAA PC games on your iPad—like Rise of the Tomb Raider, shown here.
Photo: Whitson Gordon

That’s where Moonlight comes in. This open-source app (for iOS, Android, Chrome, and other platforms) reverse engineers NVIDIA’s GameStream tech, allowing you to stream a game from your PC to your phone, tablet, or laptop—the gaming PC does all the heavy graphical work, but you can play it on any screen in the house, as long as your gaming PC has an NVIDIA card. If you don’t have an NVIDIA card, Steam Link is a slightly less mature app that does something similar on Android (Apple has rejected its iOS version for the time being). Or, if you don’t have a gaming PC at all, you could use a service like Shadow to stream games from a cloud server, or the newly-released PS4 Remote Play to stream games from a PlayStation 4.

In other words: you aren’t limited to mobile games on your mobile device. With the right setup, you can stream just about any AAA console or PC title to your tablet, hook up a controller, and turn it into the ultimate portable couch gaming machine.

Since discovering this, the Gamevice has become the saving grace of my gaming life. I still have to play some games at my PC—I’m absolutely terrible at first-person shooters unless I have a mouse and keyboard—but controller-friendly games like Rocket League, The Witcher series, the Middle-Earth series, and others feel right at home with a controller on my iPad. I can plug in the Gamevice, set up shop on the couch or in bed, and play to my heart’s content without fighting for the TV. And after getting used to that, it’s really, really hard to go back to sitting upright at my desk.


via Kotaku https://kotaku.com

March 21, 2019 at 06:34PM

Nintendo Switch to Get Xbox Live Enabled Games Over the Year

https://www.legitreviews.com/nintendo-switch-to-get-xbox-live-enabled-games-over-the-year_211108

Posted by

Shane McGlaun |

Thu, Mar 21, 2019 – 9:02 AM

Nintendo has been doing very well with its Switch game consoles. The hybrid portable and home console has proven very popular with significant hardware and software sales. While the Switch has its own online gaming service, it pales in comparison to the size of Microsoft’s Xbox Live service.

Microsoft has been a bit coy with whether or not Xbox Live will come to the Switch as the online service rolls out to other devices. At one point Microsoft Cloud Gaming head Kareem Choudhry said that there were no specific Xbox Live announcements fo the Switch.

At GDC 2019 going on in San Francisco, Microsoft again touched on Xbox Live for the Nintendo console reports Nintendolife. While there are no announcements now, Microsoft stated at the conference that players could “expect more games like Cuphead to come to Switch with Live enabled…”

The timeframe for games with Live-enabled on the Switch is over the next year. It’s a good thing for Xbox Live fans to get the service on other platforms, but it is rather surprising to see this happen.

via Legit Reviews Hardware Articles https://ift.tt/2Y6Fy3O

March 21, 2019 at 09:05AM