When Computers Get a Right Brain

Watson is kind of a left brain decision engine. It is very strong numerically, but it is kind of an extreme version of Sheldon on The Big Bang Theory. It isn't very empathetic, and it isn't intuitive. Neural networking and cognitive computing will close the gaps, resulting in machines that can do more than suggest the top choices in a decision. They will actually make the decision.

 

IBM is massively increasing its commitment to Watson, which was evident at last week's IBM Think Forum. IBM recognizes that if it can get a product that thinks first, then it can again take absolute control over the technology market, and every other company will be chasing it again.

It is making some impressive headway, and I agree that the firm that succeeds at this will not only massively change the technology market, but also change the world. The very real outcome is that many of us will find our skills have become redundant -- and when I say "many of us," I include analysts, because we are obviously at risk.
I'll close with my product of the week: Gorilla Glass, which is what's keeping us from having far more broken phone screens and likely helping Apple to avoid having to recall the iPhone 6 (though I still think it should).

The Wonder of Watson

Watson is a fascinating system. I'm referring to this class of systems as "decision engines," because they aim to help people make better decisions. What makes them different from any other tool is that they learn about you, rather than you having to learn as much about them.
Granted, with these early versions, it is kind of a shared experience -- but prior to systems like Watson, all of the effort to learn how to interface was on the user side. With Watson, the end goal is to shift virtually all of this burden to the system.
What this means is that in the near-term future, when you need something from Watson, you'll be able to sit down and immediately be productive rather than spend the weeks to months it typically now takes to learn a new system. Because Watson learns from you, the longer you work with it, the better it becomes.
As more and more Watson computers are deployed, they learn from each other creating sort of a digital Gestalt providing a massive acceleration in the learning process.

Near-Term Expectations

There were a number of examples at the Think Forum of working systems and prototypes that showcased what Watson can do today. For instance, a hospital treating a cancer patient can look at the detailed information captured on the patient and then determine the highest-probability path to wellness. Watson already has information on obscure treatments and illnesses, which it can apply to the solution.
The end result is a customized program that is developed from information gathered all over the world -- both preventing mistakes that already have been made and identifying little known successes to reach an optimal solution.
Watson could provide travelers with an experience similar to what travel agents once provided. For instance, if your plane were delayed or grounded, it would automatically find and recommend an alternate route while you were in transit, which you could execute with one Amazon-like click. When called upon to help plan a trip, Watson could apply what it knows about you and about the various airlines and hotels to create the itinerary most likely to make it trip the best you've ever had.
For help desk issues, it would take customers through custom decision trees defined by the system's ability to diagnose the problem remotely and gauge the technical capability of the caller. That would enable it to provide the fastest resolution for the lowest cost, optimizing on a blend of customer satisfaction and cost containment.
It automatically would identify and report bad trends to decision makers, along with optimized recommendations, both to resolve the core issues in a timely way and optimally reduce the related costs.
Company after company testified at the Think Forum that they were seeing massive increases in customer satisfaction and getting a far better handle on costs from Watson-type technology, which is still in its infancy.

Jobs at Risk

Now with the next range of technology advancements, there are a ton of jobs at risk -- and not the obvious ones, either. Specialists are pretty safe for a while, but folks who provide general services -- like accountants, tellers, bankers, stock traders and, well, analysts -- are kind of screwed in the long term.
Fortunately, this won't happen overnight, and in the near term there will be some great jobs training systems like Watson but, once trained, these systems will be able to train each other almost instantly.
It will be even longer before specialists are at risk, because the costs of training a system like Watson to handle a specialty will likely exceed the benefits for some time. If the specialty is unique, that time could be indefinite.

Wrapping Up: The Future

The Think Forum wrapped up with a look into the future, and it was a fascinating discussion. Watson is kind of a left brain decision engine. It is very strong numerically, but it is kind of an extreme version of Sheldon on The Big Bang Theory. It isn't very empathetic, and it isn't intuitive.
Efforts like neural networking and cognitive computing will close the gaps, resulting in machines that can do more than suggest the top choices in a decision. They will actually make the decision, moving to the next stage and doing the whole job. Fortunately, this right brain aspect is still years off, and until it is ready, people will fill that role.
Two clear job growth areas, for now and in the near term, will be trainers for systems like Watson -- people who understand both a particular subject area and how best to present it to Watson. They will be Watson's right brain partners.

Product of the Week: Gorilla Glass

I was on Fox Business News this week covering the pain of a broken smartphone screen. Toward the end of the show, we tried to break a phone on air, but the phones bounced and didn't break. That was largely because the phones used Gorilla Glass, which is broadly used on smartphones today. It isn't invulnerable -- phones that use it still break -- but they break far less frequently, and as we so humorously demonstrated, it often takes a real boneheaded move to break one.
There was a lot of speculation that the iPhone 6 was going to use sapphire glass, and that would have ended badly. That's because for some screwy reason -- likely to save money -- Tim Cook allowed the iPhone 6 to have an aluminum frame rather than the steel or magnesium alloy frame that other large phones use, and that makes the screen the stiffest part of the phone.
Gorilla Glass bends a little, which is why you have a lot of bent iPhone 6s that don't have cracked screens as well.

Corning's Gorilla Glass

Sapphire doesn't bend at all, and though you'd likely have fewer bent phones, you'd have a far higher number of even more expensive cracked screens.
Much like jewels, if you do get a nick in the screen with sapphire, that becomes a flaw -- and the screen will break on that flaw. So sapphire would have cost more and not held up well at all on a large phone.
Now, if phone makers would just work a little harder on protecting the screens, we likely could eliminate most of the breaks that occur when the phone lands on its corner on something as unforgiving as cement or tile. However, Corning's Gorilla Glass is why far more of us don't have smashed phones, and it turned my spot on Fox Business into a comedy sketch, so it is my product of the week.

Hacker Gives Google a Hand With Chrome-Android Compatibility

A hacker working independently appears to have topped Google's own efforts to extend compatibility between its Chrome OS and Android. Google last week unveiled four Android apps for Chrome books based on its ARC API. Vlad Filippov, aka "Vladikoff," went Google several giant steps better, publishing a runtime dubbed "ARChon" that allows users to run any number of Android APKs in the Chrome OS.

An unplanned convergence of Android apps and the Chrome OS may be setting the stage for a wide-open cross-platform architecture that combines Android and Chrome.

Freelance programmer Vlad Filippov, aka "Vladikoff," discovered a way around Google's limitations on its ARC, or App Runtime for Chrome, which is essentially a Chrome extension application programming interface.
Filippov hacked his way into the Google process to allow it to run on the Chrome OS and other platforms that run the Chrome Web browser. He dubbed that modification "ARChon."
He worked independently on most of the hack he developed, driven by a desire to go beyond the initial Google Android runtime program.
"I feel like this quick hack helps us get a small glimpse at the future of Android as a truly cross-platform application platform," Filippov told LinuxInsider.

Hack Overview

The ARChon runtime allows users to run any number of Android APKs in the Chrome OS or other Chrome-browser supporting platforms. Android's installable programs and middleware are distributed in an APK file format that Google developed.
ARChon creates the modifications so they run using Chromeos-apk on Chrome OS and across any desktop platform that supports Chrome. A Google project already was developing a method to use ARC to run certain Android applications on the Google Chrome OS. Google last week released four Android apps for Chromebooks based on the ARC API: Vine, Evernote, Duolingo and Sight Words.

 

Future Impact

"This is a huge development for Android. You do not want a Tower of Babel. Users just want things to work," Laura DiDio, principal at ITIC, told LinuxInsider.
The end game is that you want people to be able to use these APKs. In no way should Google try to rein in a rogue developer, she said, but Google's intellectual property should count for something.
Generally, application platform owners like Google want to control their ecosystem to some level. This is important for a variety of reasons, Al Hilwa, program director for software development research at IDC, told LinuxInsider.
"For Google, Android is monetized by its ability to take users to its back-end services such as Google Search, Maps, etc. Indeed, Google has been shifting the focus of Android to be something more about its back-end services than just open source code to run devices," he said.

Filippov envisions a future when people can enjoy their favorite apps on any device, synchronized and scaled to fit the appropriate screen.
This is a unique opportunity for tablet applications to shine, he said.
"There are not enough good tablet applications for Android, and this might be the key to create experiences that are enjoyable on desktop and tablet form factors," Filippov added.

Google has full control of the Chrome browser and the Chrome OS ecosystems, so it certainly is possible that it will include some sort of a validation component in the future to limit this feature to Chrome OS.
"I really hope they will not lock down this approach in the future versions of the runtime, because it creates a great opportunity for people to experience, develop and test Android apps on their favorite desktop platform," said Filippov.
It's his hope that most of these components will stay open source as part of the Chromium and Android Open Source projects.
"The threat of ARChon is somewhat limited," noted IDC's Hilwa, "in that increasingly more and more Android apps have to effectively leverage Google Services. Additionally, it sounds like it will be a long time before it is possible to run the entire Google Now portfolio on any platform through this technology in a reliable manner." 

VideoSelfie

      A soon-to-launch mobile application called VideoSelfie, whose founders participated in the 500 Startups accelerator last year, has raised $1.2 million for its forthcoming mobile application that lets you edit your videos (of yourself, duh) in real-time. You can add filters, decorations and GIFs, hit record, then share, all at once.
The new app is a pivot from the company’s previous product called Unda, still live here on the App Store, which offered similar technology but was focused more on video messaging.
“Video messaging became a super saturated space and we knew we had to differentiate ourselves, so we started investing a lot of time developing our own technologies for motion tracking and face tracking to enable users fast ‘decoration’ of their video messages,” explains co-founder and CEO Oscar Noriega. “Now we want to focus on recording, decorating, editing and sharing, although we keep some basic messaging and community functions,” he says.
The company is also killing off its Unda Android application and will focus on VideoSelfie on iOS for now. (The app is still in Apple’s review queue at present.)
The main technology the company has built involves face tracking and motion tracking, which makes graphics added to videos interactive by detecting the movement of the camera and the movement of your head. Users tap the GIF button, pick a graphic then configure basic settings like enabling face tracking, making the graphic full size, or changing the animation speed. Then you pick a video filter and start recording, optionally adding music. The app ships with its own standard transparent GIFs, but it supports transparent PNGs if you want to add your own.
The end result are silly, short video clips where you can apply things like kitten whiskers and ears to your head, cover your face with an animated emoji, have some dancing cat GIFs hanging out next to you, and a lot more. You can then share these clips however you choose, including with the built-in community on the app. It’s definitely not serious stuff, but it is kind of fun.

The 7-person company closed on its seed round of $900K from mainly Japanese investors, which makes sense as the team is based in Tokyo and the app has a Japanese vibe to it with GIF categories like Kawaii and Manga, for example. Investors included East Ventures, Klab Ventures, Cyberagent Ventures, and a follow on from the 500 Startups fund. Angel in the round included Takeo Matsuda (the ex-CFO of Groupon Japan) and Masi Oka (the hollywood actor from the TV show Heroes andHawaii Five-0.) The round brings its total raise to date to $1.2 million.
As for why users would try out a dedicated video selfie app instead of just using popular apps for video they already have installed on their phones, Noriega explains: “Instagram and many other apps were never intended to be deeper video editing tools or real-time composition tools.”
The company says its value is in its real-time editing and mixing tools, face-tracking tech and real-time filters. It plans to generate revenue by way of selling additional filters and effects as in-app purchases, but one could imagine that it could also license its technology to bigger companies as another avenue for monetization.

File Sharing Between Android and iOS

            As handy as services like AirDrop or Android Beam may be for shuffling content between nearby devices, they're platform-exclusive. That's not much help if you want to share photos from your Android phone to an iPhone, or vice versa. However, Google may soon overcome that barrier. Android Police,Techaeris and GigaOM all have evidence of Copresence, a service that would let Android and iOS devices swap content over WiFi. Reportedly, it uses location data (including Bluetooth) to set up the connection; after that, you can send directions, photos and other info without having to either bump devices or rely on cloud storage options like Google Drive or Dropbox. The technique shouldn't require a Google account, either.

            It the details are accurate, Copresence should be available within a matter of weeks. Having said this, you shouldn't expect a file sharing Utopia. The feature is most likely to be limited to Google apps at first, and it's probably not going to be a system-wide feature on at least iOS -- you may have to be satisfied with exchanging files through a handful of programs. Still, that's better than the walled-off sharing you likely deal with today...

Source: www.engadget.com