A nerdy father of two, a husband of a beautiful and understanding wife, an engineer who loves anime and tinkering with PCs and games, but most of all, loves God.
Until now, Halo Neuroscience’s signature brain-bending headphones were only available to very specific groups: college-level athletes, pro athletes and the military. They helped build up hype for the eventual public release. That all changes today, however: you can buy your own set of Halo Sport headphones for $699. As before, that high price stems from the Sports’ "neuropriming." The over-ears send electrical currents to your brain that, at least in theory, make it extra-receptive to training. You won’t be inherently faster or stronger, but you might hit your goals sooner than you would otherwise.
How well do they work, though? It’s hard to quantify, since there are other factors that go into an athlete’s success beyond their receptiveness to training: the quality of that training, the athlete’s other traits and, of course, the competition. Oakland Raiders cornerback TJ Carrie is having a good year so far after wearing the Halo Sport, but the Olympians who used them? Their results are… mixed. Natasha Hastings helped win gold in the Rio Olympics’ 400m women’s relay. Mike Rodgers’ sprint relay team was disqualified in its final race, though, and Michael Tinsley was knocked out in the first round of the 400m hurdles.
This doesn’t mean that the brain-altering technology is ineffective. However, it does suggest that you should temper your expectations. No matter how well the Halo Sport works, you aren’t guaranteed fame and glory — you may perform better than you would otherwise, but you probably won’t win championships or smash records unless you’re already talented.
Google announced a big update to its Translate service today that involves Neural Machine Translation. That sounds like some future AI, machine learning science, secretive robotic madness, and it may very well be, but it also means Google Translate is going to get a whole lot better everywhere you use it.Â
Google explains Neural Machine Translation as a system that “translates whole sentences at a time, rather than just piece by piece.†In other words, Google Translate will use this technology to “help it figure out the most relevant translation, which it then rearranges and adjusts to be more like a human speaking with proper grammar.†Instead of getting a translation that is almost as impossible to read as the language you didn’t needed to translate, the translation should return in a much more natural format for you.
That make sense? Take a look at the image above and I’m guessing it will.
Google is calling this improvement the single biggest leap that Translate has “seen in the last ten years combined.†Yep, it’s a big one.
To start, Google Translate will Neural Machine Translation to work to and from English and French, German, Spanish, Portuguese, Chinese, Japanese, Korean, and Turkish. Those languages make up for more than 35% of Translate queries, so we’re off to a good start. Google did mention that the goal, as you probably guessed, is to make this magic happen for all 103 supported languages.
Neural Machine Translation will be available in in the Translate app and website.
The perils of leaving computers unattended just got worse, thanks to a newly released exploit tool that takes only 30 seconds to install a privacy-invading backdoor, even when the machine is locked with a strong password.
PoisonTap, as the tool has been dubbed, runs freely available software on a $5 Raspberry Pi Zero device. Once the payment card-sized computer is plugged into a computer’s USB slot, it intercepts all unencrypted Web traffic, including any authentication cookies used to log in to private accounts. PoisonTap then sends that data to a server under the attacker’s control. The hack also installs a backdoor that makes the owner’s Web browser and local network remotely controllable by the attacker.
The primary motivation is to demonstrate that even on a password-protected computer running off of a WPA2 Wi-Fi, your system and network can still be attacked quickly and easily. Existing non-HTTPS website credentials can be stolen, and, in fact, cookies from HTTPS sites that did not properly set the ‘secure’ flag on the cookie can also be siphoned.
Unsecured home or office routers are similarly at risk. Kamkar has published the PoisonTap source code and additional technical details here and has also released the following video demonstration:
PoisonTap – exploiting locked machines w/Raspberry Pi Zero
Once the device is inserted in a locked Mac or PC (Kamkar said he hasn’t tested PoisonTap on a Linux machine), it surreptitiously poisons the browser cache with malicious code that lives on well after the tool is removed. That makes the hack ideal for infecting computers while they are only briefly unattended. Here’s how it works.
Once the PoisonTap software is installed, the Raspberry Pi device becomes a miniature Linux computer that presents itself as an Ethernet network. Like a router, it’s responsible for allocating IP addresses for the local network through the dynamic host configuration protocol. In the process, the device becomes the gateway for sending and receiving traffic flowing over the local network. In this sense, PoisonTap is similar to a USB exploit tool demonstrated in September that stole login credentials from locked PCs and Macs.
Through a clever hack, however, PoisonTap is able to become the gateway for all Internet traffic as well. It does this by defining the local network to include the entire IPv4 address space. With that, the device has the ability to monitor and control all unencrypted traffic the locked computer sends or receives over its network connection.
PoisonTap then searches the locked computer for a Web browser running in the background with an open page. When it finds one, the device injects HTML iframe tags into the page that connect to the top 1 million sites ranked by Alexa. Because PoisonTap masquerades as the HTTP server for each site, the hack is able to receive, store, and upload any non-encrypted authentication cookies the computer uses to log in to any of those sites.
Given its highly privileged man-in-the-middle position, PoisonTap can also install backdoors that make both the Web browser and connected router remotely accessible to the attacker. To expose the browser, the hack leaves a combination of HTML and JavaScript in the browser cache that produces a persistent WebSocket. PoisonTap uses what’s known as a DNS rebinding attack to give remote access to a router.
That means attackers can use PoisonTap to remotely access a browser as it connects to a website or to gain administrative control over the connected router. Attackers still must overcome any password protections safeguarding an exposed router. But given the large number of unpatched authentication bypass vulnerabilities or default credentials that are never changed, such protections often don’t pose much of an obstacle.
PoisonTap challenges a tradition that can be found in almost any home or office—the age-old practice of briefly leaving a locked computer unattended. And for that reason, the ease and thoroughness of the hack may be understandably unsettling for some people. Still, several safeguards can significantly lower the threat posed by the hack. The first is to, whenever possible, use sites that are protected by HTTPS encryption and the transmission of secure cookies to prevent log-in credentials from being intercepted. A measure known as HTTP Strict Transport Security is better still, because it prevents attack techniques that attempt to downgrade HTTPS connections to unsecured HTTP.
As a result, neither Google nor Facebook pages can be triggered by computers infected by PoisonTap. Sadly, multi-factor authentication isn’t likely to provide much protection because it generally isn’t triggered by credentials provided in authentication cookies.
End users, meanwhile, should at a minimum close their browsers before locking their computer or, if they’re on a Mac, be sure to enable FileVault2 and put their machine to sleep before walking away, since browsers are unable to make requests in such cases. Regularly flushing browser caches is also a sound, albeit imperfect, measure. For the truly paranoid, it may make more sense to simply bring laptops along or to turn off machines altogether.
Eating healthy will do more for your immune system than megadoses of supplements.
Gillian Blease/Ikon Images/Getty Images
hide caption
toggle caption
Gillian Blease/Ikon Images/Getty Images
Eating healthy will do more for your immune system than megadoses of supplements.
Gillian Blease/Ikon Images/Getty Images
Flu season is upon us, which means it’s time for the wave of advertisements promoting $8 juices or even more expensive supplements to “boost your immunity” or “support immune function.”
But those are marketing terms, not scientific ones. And there’s no proof that those products are going to keep you from getting sick.
When you’re exposed to a virus like the influenza virus, a number of factors determine if you actually get sick, and if so, how severely. One is pre-existing immunity, either from being previously exposed to a similar strain or through a vaccine, says Gregory Poland, a spokesman for the Infectious Diseases Society of America and a professor of medicine at the Mayo Clinic. Just last week, a paper published in Science reported that the flu strain you were first exposed to can affect your protection against new strains that jump from animals to humans.
Your immune status also matters; people who have untreated HIV or have recently received a bone marrow transplant, for example, cannot fight off infections like healthier people can.
Age, too, is a factor, with the very young and the very old suffering worse bouts of the flu.
And yes, what you eat does matter. “We know for a healthy immune system you need a healthy diet,” says Joan Salge Blake, a registered dietitian nutritionist and clinical associate professor at Boston University. You need protein as well as micronutrients including vitamins C, A, and E and zinc, she says.
The ideal way to get those nutrients, however, is to eat a healthful, varied diet, including sufficient protein and a variety of fruits and vegetables, says Poland. If you’re already doing that, it’s unlikely that you have major nutritional deficiencies. One exception is vitamin D, which is necessary for bone health and can be hard to get from food alone, though there’s not a consensus on the cutoff for a vitamin D deficiency.
Even those who aren’t eating the most healthful diet (i.e. most of us) are likely getting a lot of nutrients through fortified packaged foods like cereal. Nutrient deficiency does happen, but it’s relatively uncommon in the U.S. According to the Centers for Disease Control, less than 10 percent of the population is deficient in micronutrients, though Poland says certain groups of people are at risk, including vegans who are not careful about their food choices and older people who eat scant, unvaried diets.
“If you are malnourished, your immune system is going to suffer,” says Salge Blake.
If you are not nutrient deficient or malnourished, though, taking megadoses of vitamins are not going to super-charge your immune system or prevent you from catching the flu or other respiratory viruses. Vitamin C, often touted as a way to stay healthy in the winter, doesn’t seem to reduce the incidence of colds, though there is some evidence it may cut their duration and it might be helpful for people who experience short periods of heavy physical activity, according to a 2013 Cochrane review.
Juices sound attractive; after all, they are made from real foods. But Salge Blake says the best way to get the nutrients supplied by fruits and vegetables is to actually eat the fruits and vegetables themselves. That way you get the fiber, which slows the absorption of natural sugars and carries its own health benefits.
Drinking juices also makes it easy to consume too many calories, and obesity suppresses immune function.
What can the average person do who wants to make sure their immune system is as healthy as possible? In addition to a healthful diet and sufficient sleep, Poland recommends exercise, staying up to date on flu and pertussis vaccinations, staying away from people who are obviously sick, and washing your hands.
Katherine Hobson is a freelance health and science writer based in Brooklyn, N.Y. She’s on Twitter: @katherinehobson.
from NPR Topics: News http://ift.tt/2ghJuwJ
via IFTTT
On the surface, Google Photos has a simple mission: to store all your pictures. Specifically, Google says it wants the service to be a home for all of your photos, and today that mission expanded to encompass the old photos you took on a point-and-shoot back in the ’90s. The company just released a new app called PhotoScan for iOS and Android, and it promises to make preserving the memories in your old printed photos much easier. Additionally, while Google was at it, it also issued several updates to its core Photos app.
PhotoScan is definitely the star of the show, though. According to engineers from Google who showed the app to the press earlier today, PhotoScan improves on the old "photo of a photo" technique that many now use to quickly get a digital copy of old prints. It’s also a lot cheaper than sending pictures out to be scanned by a professional, not to mention faster and more convenient than using a flatbed scanner.
When you open up the PhotoScan app, you’re prompted to line up your picture within a border. Once you have the picture aligned, pressing the scan button will activate your phone’s flash and start the process of getting a high-quality representation of the photo. Four white circles will appear in four different quadrants of the image. You’ll be prompted to move your phone over each dot until it turns blue; once all four dots are scanned, the app pulls together the final image.
When moving the phone to scan each dot, the app is taking multiple images of the picture from different angles to effectively eliminate light glare — something Google cited as the biggest culprit that ruins digital pictures of photo prints. In practice, in Google’s tightly controlled demo setup, it worked perfectly. It was easy to see how the lights in the room cast glare on the photo print and equally obvious how the app managed to eliminate it in the final scan. It’s a bit of an abstract process to describe, but it worked as promised. We’ll need to test it further outside of Google’s own testbed, but the early results are definitely encouraging.
The app also lets you adjust the crop to remove any hint of the background surface peeking into the photo, but it’s otherwise a pretty minimal experience. Once you’re done scanning, the app prompts you to save your scans. They’re saved directly to your phone’s storage; you can then upload them to Google Photos or the backup service of your choice. Google specifically said that it wanted this app to exist outside of Google Photos so that people could scan images and use whatever service they want to back them up.
Beyond PhotoScan are some noteworthy additions to the proper Google Photos app. The biggest change here is that there are a host of new photo-editing options on board. The Google+ app actually used to have a pretty robust set of editing options, but when Photos was liberated as a standalone app, the editing features were significantly culled down.
As of today, Google Photos for both iOS and Android now has a entirely redesigned set of editing tools and filters. The "auto enhance" feature, which tweaks brightness, contrast, saturation and other characteristics of your photo has been improved thanks to the machine learning technology that is at the core of nearly all of Google’s products. It can look at a photo and recognize what a photo editor might do to try and improve the image. Auto Enhance has long been a solid feature, so seeing it continue to get smarter is definitely a good thing.
If you want to make further adjustments, the simple "light," "color" and "pop" sliders that were in the previous Google Photos app have been greatly expanded. Now, you can tap a triangle next to "light" or "color" to see a view with a host of more granular editing tools like exposure, contrast highlights, saturation, warmth and so on. Those tools aren’t right in your face, so people who don’t want to dive in can still make adjustments — but those who really want to go deep on editing their pictures will surely appreciate the option. I used to be a big fan of the Google+ photo editing tools so seeing these features come back is very welcome.
Google called out two of those adjustments in particular as things that only it can do with its vast store of photographic information. A new slider called "deep blue" saturates blues in an image like the sky or water to make them more vibrant, and it knows to specifically target those hues while leaving others unchanged. There’s also a skin tone filter that can adjust saturation specifically on a subject’s skin without altering the rest of the image. Other editing programs have similar filters, but Google says that this one is particularly accurate because of the millions of photos it has analyzed — it just has a better sense of what is skin is, compared with other editors.
Lastly, Google added 12 new filters (of course it did) that take advantage of machine learning to be a little smarter than the standard option. Rather than always slapping a default set of adjustments on a picture, Google Photos will make subtle improvements to the image first; it sounds like a combination of auto enhance as well as a filter. But those enhancements will be optimized to work well with the filter you’re adding. It sounds nice, and the filters looked good on the images Google was showing off, but we’ll need to spend some time playing around with it to see if they’re really any better than what Instagram already offers.
Editing is the main addition to Google Photos, but there are a few other improvements here as well. If you’re invited to a shared album, the app will prompt you with suggestions from your own photos to add. It’s another place where Google’s machine learning comes into play. And the movie maker, which can automatically select related photos and set them to a soundtrack, will gain some new event-focused options in the coming months.
The first of those is "lullaby," a video made by specifically looking for sleeping baby photos and combining that with a peaceful soundtrack. Specifically, Google’s servers can find groups of pictures of the same child and look for ones when it is sleeping and pull those all together. Google said it went for the sleeping baby specifically as a way to put together some calming memories for parents, who always seem to cherish those fleeting moments when their new baby is asleep.
For those of us without kids, Google has a new Christmas memories movie that’ll look at your pictures over the years and pull together ones with Christmas "markers" — things like trees, Santa hats, presents and so forth. In April next year, Google will also start auto-creating movies of pets as well as outdoor moments (timed to launch around Earth day).
All told, this is probably the biggest update to Google’s photo products since it launched in mid-2015. There are plenty of other services that offer near-unlimited photo backups, but Google’s machine learning based on all the data in its systems is second to none. Yes, that requires Google to analyze everything you put into it, but that’s been the case for years now. If you’re comfortable giving Google access to your data, these new photos updates are definitely worth checking out. And if you want to try PhotoScan but are worried about your privacy, you don’t even need to upload your pictures to Google. The new PhotoScan app and updated Google Photos should be available in the app store and on Google Play now.
The BLU R1 HD is one of the devices that was backdoored by a Chinese software provider.
BLU Prod
Security firm Kryptowire has uncovered a backdoor in the firmware installed on low-cost Android phones, including phones from BLU Products sold online through Amazon and Best Buy. The backdoor software, initially discovered on the BLU R1 HD, sent massive amounts of personal data about the phones and their users’ activities back to servers in China that are owned by a firmware update software provider. The data included phone number, location data, the content of text messages, calls made, and applications installed and used.
The company, Shanghai AdUps Technologies, had apparently designed the backdoor to help Chinese phone manufacturers and carriers track the behavior of their customers for advertising purposes. AdUps claims its software runs updates for more than 700 million devices worldwide, including smartphones, tablets, and automobile entertainment systems. It is installed on smartphones from Huawei and ZTE sold in China.
A lawyer for the company told The New York Times that the data was not being collected for the Chinese government, stating, “This is a private company that made a mistake.â€
The backdoor was part of the commercial Firmware Over The Air (FOTA) update software installed on BLU Android devices provided as a service to BLU by AdUps. In a report on the finding, a Kryptowire spokesperson said:
These devices actively transmitted user and device information including the full-body of text messages, contact lists, call history with full telephone numbers, unique device identifiers including the International Mobile Subscriber Identity (IMSI) and the International Mobile Equipment Identity (IMEI). The firmware could target specific users and text messages matching remotely defined keywords. The firmware also collected and transmitted information about the use of applications installed on the monitored device, bypassed the Android permission model, executed remote commands with escalated (system) privileges, and was able to remotely reprogram the devices… The firmware that shipped with the mobile devices and subsequent updates allowed for the remote installation of applications without the users’ consent and, in some versions of the software, the transmission of fine-grained device location information.
The transmissions were discovered by Kryptowire in lab testing. The company immediately notified Google, BLU, AdUps, and Amazon—which is the exclusive retailer of the BLU R1 HD—of its findings.
The user data was sent in JavaScript Object Notation (JSON) format to a number of servers, all with the hostname bigdata: bigdata.adups.com, bigdata.adsunflower.com, bigdata.adfuture.cn, and bigdata.advmob.cn. The data collection and transmission capability is spread across different applications and files. Text message data (encrypted with DES, which Kryptowire researchers were able to recover the key for) and call log information were sent back every 72 hours. Other data, including location data and app use, was sent every 24 hours.
A BLU spokesperson told Ars that the software backdoor affected a “limited number of BLU devices†and that the “affected application has since been self-updated and the functionality verified to be no longer collecting or sending this information.†According to The New York Times report, BLU reported about 120,000 devices were affected and patched.
