Highlight

Những điều thú vị khi dùng Trí tuệ nhân tạo của Viettel

Những người dùng Internet tại Việt Nam thường lấy “chị Google” ra để… giải trí. Khi “chị” đọc văn bản hay chỉ đường cho người tham gia gi...

Tuesday, September 20, 2016

Google acquires API.AI, a company helping developers build bots that aren’t awful to talk to

api

Google has just disclosed that it has snatched up the team behind API.AI. API.AI provides tools to developers to help them build conversational, Siri-esque bots.
As humans, we’re pretty good at communication. If someone says “the girl saw a man with the binoculars,” we can generally use contextual clues to figure out if they meant that the girl saw the man by using binoculars, or saw a man who was carrying binoculars.
Teaching a robot to do the same is a bigger challenge. Add in ambiguity (by “get me a lift” do you mean any old car, or a Lyft, specifically?) and there’s an endless number of ways to say pretty much any one thing… and, well, the challenge becomes huge.
API.AI helps developers who are building bots tackle this by providing them with tools to keep them from endlessly reinventing the wheel. Their APIs handle things like speech recognition, intent recognition and context management, and allows devs to provide domain-specific knowledge (like that “deep dish” and “Chicago-style” can probably mean the same thing to your pizza delivery bot) that might be unique to your bot’s needs.
API.AI currently plays friendly with 15 languages/dialects, including English, Chinese, French, German and Spanish.
According to a running counter on its homepage, API.AI has processed a little over 3 billion API requests to date. Meanwhile, Google says over 60,000 developers have built stuff with API.AI’s toolset.
The price and terms of the acquisition have not been disclosed yet, but API.AI had raisedaround $8.6 million to date according to Crunchbase.

Etsy buys Blackbird Technologies to bring AI to its search



t’s not just gigantic search engines that are looking to acquire tech and talent around emerging areas like speech recognition. Even design and craft marketplaces can use a little machine learning and artificial intelligence to make their wheels turn a little better.
Today, the popular handmade-goods site Etsy announced it has acquired a startup calledBlackbird Technologies, which developed algorithms for natural language processing, image recognition and analytics — similar to those used by Amazon and Google for product and other searches — and then “democratized” them to be used by any company of any size. At Etsy, the tech will be used to improve its own search features.
Financial terms of the deal have not been disclosed, but we’re asking and will update if we learn more. The company’s employees — it appears there were around 10 — will all be joining Etsy. “The team possesses a deep expertise in Artificial Intelligence, search, and distributed systems and has direct experience working in these areas for some of the largest technology companies in the world,” Etsy said in a statement.
Little is known about Blackbird — no listings on Crunchbase or AngelList, a fairly unspecific website for Blackbird itself — and even Etsy’s announcement of the deal is light on company details. According to LinkedIn, though, the two co-founders are Nikhil Raghavan(CEO) and Manju Rajashekhar (CTO), who collectively had worked on similar technologies across places like Google, Twitter, Microsoft, Oracle and Yahoo.
Because Etsy is going to be using Blackbird’s tech for its own platform, it seems that Blackbird’s current business plans, which deployed its tech by way of an API, will shut down. Customers had included Nasty Gal, ThredUP, and Tophatter.
Etsy has had a fairly strong year so far, beating analysts’ expectations in two consecutive quarters on the back of growing sales and customer base.
But at the same time, it’s facing some heat in the form of other competitors hot on Etsy’s crafty tail with their own marketplaces. The biggest of these is Amazon (perhaps every e-commerce company’s worst nightmare) with its own Handmade storefront that will vie for the same producers, the same customers, but throwing in some attractive Prime shipping (and one-stop shopping) into the mix.
This is why buying Blackbird is a smart move. It will essentially help Etsy improve its search and discovery to point shoppers to items better matched to their tastes, and keep them from abandoning their shopping carts, a common pain point for any e-commerce enterprise.
“Buyers come to Etsy.com for items they can’t find anywhere else and our goal is to help them discover exactly what they want among our 40 million unique listings. Leveraging Blackbird’s technology, we believe we can enhance the buyer experience by making search quicker and easier and by surfacing even more relevant, tailored product recommendations,” said Chad Dickerson, Etsy, Inc. CEO, in a statement. “Our team has already made substantial enhancements to the search & discovery process on Etsy, especially through features like Exploratory Search, and we are excited for Blackbird’s world-class team and technology to accelerate our progress in this key area.”
Longer term, it seems that Etsy will look to use some of Blackbird’s tech in other areas beyond search, too. “In addition, longer-term, we believe there may be opportunities to deploy Blackbird’s Artificial Intelligence technology in areas beyond search that will help strengthen our markets and seller services platform,” the company said.
This is Etsy’s sixth acquisition, but only its first since going public in 2015. Interestingly, it looks like the first Etsy has made that points directly to its technology around its product (Adtuitive, Etsy’s first buy, was a move to bring in some adtech to the business).

AI can make your money work for you



The way we thought of apps back in 2008 is how we’ll view artificial intelligence five years from now; 2016 will look like the AI stone age.
Although apps have created completely new behaviors, especially in the fintech space with mobile banking and shared payments, the app economy, unfortunately, seems to be reaching a peak, and monetizing an app has become an uphill battle.
Apps will be the technology of the past, ushering in a new era for AI. Bots today are merely app replacements or novelties, and the experience still feels like talking to a robot. But with bots eventually becoming smarter than today’s apps, they’ll inevitably solve the problems that apps cannot fix in entirely new ways that haven’t even been realized yet.
Let’s take a look ahead. In the future, for example, bots, and AI in general, will help you make money off your checking account.
Did you know that extra cash in your checking account is a missed opportunity? Every day, it loses value to inflation. To generate better returns, you could keep the bare minimum in your checking account and invest the rest. However, unexpected expenses can drain your account suddenly. Without extra cushioning in your checking account, you risk getting slapped with bank fees or credit card debt that quickly cancel out any gains from your investments.
AI will enable us to increase our wealth while decreasing our anxiety.
It feels like you can’t win. Either you’re missing out on capital gains, or you’re playing limbo with your account balance. AI will make this struggle a thing of the past.
Advances in AI will create a robo-accountant that knows your spending better than you do. By analyzing your purchase history, it will constantly move money between your checking, savings, investments and credit cards. This way, your checking account’s balance is always in the narrow “sweet spot:” high enough to avoid fees, but not so high that you miss out on investment yield.
Right now, finding that sweet spot is time-consuming and anxiety-inducing. In time, the robo-accountant will know when you’re likely to splurge. It will know when your car will need a repair, when your electric bill will spike. It will know when you’re actually better off carrying a balance on your credit card than paying your bank’s minimum-balance fee.
A fee-minimizer/yield-optimizer could exist without AI, but it wouldn’t work that well. AI incorporates the ability to make complex judgments based on your history of spending habits, the fees from your various financial institutions and countless other factors. This is the crucial link that makes the product useful: It plans ahead so you don’t have to.
Your robo-accountant will take everything into consideration to make sure you get the most yield from your money while minimizing the fees you pay across the board. AI will enable us to increase our wealth while decreasing our anxiety. Given that more than 60 percent of Americans worry about missing their retirement goals, this is no small feat.
By Nathan Richardson - is the founder and CEO ofTradeIt.

Globalization, Robots, and the Future of Work: An Interview with Jeffrey Joerres

en Jeffrey Joerres first joined Manpower, in 1993, the labor market was relatively stable and the company was still largely focused on traditional office, clerical, and industrial staffing. But since then, the employment landscape has been dramatically reshaped by globalization and rapid advances in technology. Joerres, who led ManpowerGroup for 15 years before stepping down in 2015, responded to the shifts in kind, expanding the company’s international operations and moving into the increasingly competitive market for IT, finance, and engineering professionals. Joerres, now 56 and a private investor based in Milwaukee, talked with HBR’s editor, Amy Bernstein, about the transformation of work and how to manage it.
HBR: What trends do you see shaping the workforce?
Joerres: The obvious one—the one that has shaken up the market most—is globalization. In the 1990s, if you didn’t have a China strategy, you were missing out and putting your company at a competitive disadvantage. These days, companies need a global strategy for finding a highly skilled, cost-effective labor force.
Lately we’ve seen the emergence of micromarket analysis that reveals geolocated pools of skills. So companies are tapping specific areas for specific skills. They may put their call center in Manila and a transaction processing center in Bratislava. But after the initial move to take advantage of available skills and labor arbitrage, the location matures quickly and those benefits dry up.
Let’s say you find a previously undiscovered workforce of multilingual IT professionals in Bratislava, and you set up a processing center. Pretty soon your competitors are going to follow you there, deplete the skills pool, and drive up wages. In other words, the greenfield you originally had all to yourself has become a battlefield. And now it’s not just the decreasing value of labor arb, it’s also a shortage of the most competitive skills. You’re once again fighting for talent, you’re investing more in training, and you’re back to the micromarket mining of these labor skills.
How is micromarket mining different from traditional skills sourcing?
Companies are doing more “micro footprinting,” and that takes a nomadic mentality: You’re ready to pick up and move when required. Large footprinting, on the other hand, means you’re committed to a community for better or worse. More and more, companies will need to take a dual approach, establishing large locations and more-temporary, smaller operations at the same time.
The location of talent pools isn’t constant either. Now we’re seeing in-demand skills pop up in different areas of a country or the world. For a few years the best place to find IT developers was Kraków; then it shifted to Kiev, because Kraków became saturated. It’s kind of like Whac-A-Mole: You’re confronting this perpetual, fast-moving skills dodge, and it’s only going to get worse.
Darren Hauck
PHOTOGRAPH BY DARREN HAUCK
At this point most of the greenfields have been developed; very few remain. Clearly Central Africa is the next greenfield for skills, but you have to have a lot of courage to get in there right now. As soon as that region has matured enough from a labor law perspective, and the problems with bribery and the black market have been addressed, it too will get soaked up. We are not going back to the days of finding a little gold vein of labor of your very own.

Robots and Jobs

Are robots really as much of a threat as some people say?
Artificial intelligence and robotics are affecting the labor market, but they’re not yet in broad use. As soon as you can get a robot for $5,000 instead of $100,000, as soon as you can get AI with better voice recognition, and as soon as you can get full contextual AI that can anticipate and answer questions without human intervention—that’s going to throw the labor markets into a tizzy.
How’s that going to play out?
The conventional wisdom is that increased productivity helps labor markets—it creates a temporary disconnect while workers scramble to acquire new skills to tackle new jobs, but they catch up relatively quickly. In today’s context I don’t buy that. The disconnect is happening a lot faster than in the past, and as a result there are more displaced and discouraged workers than ever before.
In many ways, what we have now in the U.S. is similar to the early 19th century, when the Luddites first worried that machines were going to steal their jobs. We must deal with the reality that when full-scale robotics and AI arrive in a broad-based, affordable, easily justifiable way, we’ll see enormous waves of workers put out of work and ill prepared to take on very different jobs. This is going to create challenges that our institutions are not ready for.

Building and Leading Your Future Workforce

How can organizations ensure that they have the right skills for the future when they don’t even know what they’re going to need?
Most companies have to operate in a legacy world, meaning that they have to be able to keep doing the nuts-and-bolts work at the core of their business. But they also have to be ready to compete in a fast-changing environment, one that’s really hard to predict. We know that the skills needed in the future will be durable and broad—like problem solving and the ability to work on fluid teams—but they’re hard to put your finger on. If I can develop that sort of agility in my organization to tackle the leading-edge stuff, I can then bring on the more discrete, recognizable skills as they’re required. But no one knows what all those specific skills are going to be.
Investors don’t ask for your workforce strategy, they ask for your business strategy.
Nevertheless, companies routinely fail to put in place a workforce strategy that supports their business strategy. Most have done a pretty good job of looking out three years, maybe five, but they don’t then marry that up with the skills they’re going to need. They don’t deal with how their strategy is going to change their sales force, or their engineering core, or their logistics management. This is partly because the investor community doesn’t ask for your workforce strategy; it only asks for your business strategy. The really good companies, particularly the large ones that have been burned—the Accentures and IBMs of the world—are pretty good at it, because they’ve learned to live in multiple worlds and continually look for ways to better serve their customers. The customers themselves are often less good at it.
What’s your advice to companies that want to develop a workforce strategy?
For most companies today, business takes place on multiple fronts at multiple speeds. It’s mandatory to put in place multiple work models and truly practice them. I’m talking about putting in crowdsourcing as one element, distant manufacturing or technology or transaction processing as another, contract-temporary-moving-to-full-time as yet another. My point is that you don’t know what you’re going to need three to five years from now, but if you have your skills sharpened in multiple work models, you can turn the volume up in one and down in another, minimizing the latency time to get what you need. And that’s going to be the difference between success and failure.
In many ways, the labor market is like a product life cycle—in fact, the two are now locked together in many instances. Think of mobile phones: They’re a high-demand, low-margin business, and their product life cycles are really short. That has worked its way down into the labor market. You don’t have a couple of years to develop the workforce you need. If you wait that long, you’re going to have your head in your hand.
Keeping up with workforce shifts isn’t easy, particularly for big, traditional companies.
The problem is that they’re trapped by their own history. So many times I see big companies try to move more nimbly, but their stumbling blocks are their own culture and middle managers who are not fully committed to multiple work models and see them as a threat to their span of control.
Clearly you can’t just blow your culture up. You have to figure out how to modify it so that people are able to learn and adapt. It’s doable, but you have to continually break down behaviors that once worked but now get in the way. Command-and-control behaviors, for example, don’t foster agility. Fifteen years ago, managers held the knowledge; now the systems do. It’s hard to make that shift.
How do we train people to manage this workforce that you’re describing?
As we look to the future, role modeling of behaviors is going to be more important than training. You’ll still have some form of leadership training, but it’ll evolve. When I went through my first training class, the whole first week was on performance appraisals and salary administration—you know, the rule book. Well, managers don’t need that now.
It really comes down to this: You need to be vigilant and make sure that your organization is intolerant of the behaviors that dampen flexibility and agility and learning and adaptation. As a leader, you’re a parent in the best sense. You can’t just say, “Here are my rules—now follow them.” You have to role model all the time, you have to demonstrate accountability all the time—it’s never-ending. This kind of leadership is exhausting work.
But the rewards are enduring competitive advantage, right?
Absolutely. The rewards are great, but they’re not binary. In other words, a switch doesn’t go on and you suddenly get all the benefits. The flywheel just goes faster, and you need to go faster too. You can never be satisfied, and that’s where you get frustration. When are we done with this efficiency thing? The answer is never.
People are still saying we’re in recovery from the last recession. I get that from an economic viewpoint, but that’s not good language inside a company. Better to say, “This is it, guys. It’s not going to get any better, so let’s enjoy what we’re doing here. Let’s have some fun. Let’s win at this.” Because that is the secret sauce. It may mean that the flywheel has to turn faster, so a company must parse work differently to keep up. Parsing of work is not new—think of outsourcing call centers, tagging photos, or using contractors. However, dynamic parsing is new and will become mandatory. The ability to rapidly shift the location of the work according to skill availability and criticality will be the competitive advantage.
Our systems are trying to fit the way labor markets worked in the past.
This makes the job of communicating strategy and motivating workers much more difficult and much more critical. Because there’s an emotional toll to all this: Employees can easily feel as if they’re getting jerked around, especially if the communication from management is not consistent and authentic.

Institutions Must Catch Up

You’re describing a working environment that’s very different from the one many of us first entered. Is the world ready for that?
Our institutions are inadequate. Look at unemployment compensation, welfare, Social Security—these were all put in place in the middle of the past century. And they were based on certain assumptions: that when you lost your job, you would go through a process and on the other side find a job that you’d then have for a long time. Today that’s not going to happen. Look at the gig economy, look at parsed work—all these models just allow us to move faster. My dad had a second job. He went to a gas station after he got home from his first job, and he ate his dinner in between. Well, second jobs look different now. Uber is a second job. So what do you do when someone is collecting unemployment and takes a job with Uber to moonlight while he’s in training for a new full-time job? Should he lose food stamps or health care because he’s earning a little extra money to get by? Our systems look broken because they’re trying to fit things into the way the labor markets worked in the past.
The same goes for broad-access universities. They’re built on the old labor models. They’re not turning out graduates with the skills companies need. So we have to refashion these institutions that are so important to our society.
What’s your solution?
I think we need an iterative model. Why does it have to be all in or all out? Why can’t someone be on partial welfare? Or on partial unemployment compensation? If a worker loses her job that paid $50,000 a year but can only find a new job for $40,000, her unemployment compensation goes away, but she has lost $10,000. Why don’t we make up the difference for her for another six months because she had what it took to go out and find a job? Some people might see that as a giveaway. It’s not. It’s a small price to pay to encourage that worker to get back into the market. I’d rather pay someone to be in the market than out of the market.
By Amy Bernstein - is the editor of Harvard Business Review.