Daily Tech Snippet: Thursday, May 12

• What happened when a professor built a chatbot to be his teaching assistant: To help with his class this spring, a Georgia Tech professor hired Jill Watson, a teaching assistant unlike any other in the world. Throughout the semester, she answered questions online for students, relieving the professor’s overworked teaching staff. But, in fact, Jill Watson was an artificial intelligence bot. Ashok Goel, a computer science professor, did not reveal Watson’s true identity to students until after they’d turned in their final exams. Students were amazed. “I feel like I am part of history because of Jill and this class!” wrote one in the class’s online forum. “Just when I wanted to nominate Jill Watson as an outstanding TA in the CIOS survey!” said another. Now Goel is forming a business to bring the chatbot to the wider world of education. While he doesn’t foresee the chatbot replacing teaching assistants or professors, he expects the chatbot’s question-answering abilities to be an invaluable asset for massive online open courses, where students often drop out and generally don’t receive the chance to engage with a human instructor. With more human-like interaction, Goel expects online learning could become more appealing to students and lead to better educational outcomes. At the start of this semester Goel provided his students with a list of nine teaching assistants, including Jill, the automated question and answering service Goel developed with the help of some of his students and IBM. Goel and his teaching assistants receive more than 10,000 questions a semester from students on the course’s online forum. Sometimes the same questions are asked again and again. Last spring he began to wonder if he could automate the burden of answering so many repetitive questions. As Goel looked for a technology that could help, he settled on IBM Watson, which he had used for several other projects. Watson, an artificial intelligence system, was designed to answer questions, so it seemed like a strong fit. To train the system to answer questions correctly, Goel fed it forum posts from the class’s previous semesters. This gave Jill an extensive background in common questions and how they should be answered. Goel tested the system privately for months, having his teaching assistants examine whether Jill’s answers were correct. Initially the system struggled with similar questions such as “Where can I find assignment two?” and “When is assignment two due?” Goel tweaked the software, adding more layers of decision-making to it. Eventually Jill reached the point where its answers were good enough. The system is only allowed to answer questions if it calculates that it is 97 percent or more confident in its answer. Goel found that was the threshold at which he could guarantee the system was accurate. There are many questions Jill can’t handle. Those questions were reserved for human teaching assistants. Goel plans to use Jill again in a class this fall, but will likely change its name so students have the challenge of guessing which teaching assistant isn’t human.

• Cowen: It Looks Even More Like Amazon Will Become America's Top Clothing Retailer in 2017: It's been a tough few weeks for retailers, and today is no different. But there is one exception: Amazon.com Inc. Whereas XRT, the SPDR S&P Retail ETF is down over 4 percent,  Amazon shares are up over 2 percent. One driver of the retail selloff is Macy's, whose earnings report sent shares tumbling this morning after the retailer cut profit forecasts and missed analyst estimates on revenue. That's the opposite of what happened to Amazon last week when the web giantblew past analyst estimates. Given this recent action in the retail space, analysts at Cowen and Company are reiterating their call for Amazon to displace Macy's as the number one U.S. apparel retailer by next year. "[Our calculation] implies a gain in U.S. Apparel & Accessories market share from 5 percent to 14 percent," the team, led by John Blackledge write in their note, adding, "We have seen a continued shift away from more traditional retailers...In the first quarter of 2016, Amazon Apparel purchasers were up ~19 percent year-over-year, while Apparel purchasers fell ~1 percent year-over-year and ~5 percent year-over-year at Wal-Mart and Target, respectively." The team previously made this call in July of 2015, saying they were confident that Macy's would lose its top spot within two years. Looking at revenue from electronics and general merchandise at Amazon, the trend is obvious. Amazon has seen continued growth in its retail business with the expansion of its Prime offerings. Their stock prices have certainly reflected these trends. Amazon is up 65 percent over the past year, whiles Macy's is down more than 50 percent. 

• Google’s answer to Amazon's Echo is code-named ‘Chirp’ and is landing soon: A product team at Google is working on a hardware device that would integrate Google's search and voice assistant technology, akin to the Amazon Echo, Recode has learned. Google's device will resemble its OnHub wireless router, according to several sources. We don't know if it has a name yet, but internally the project goes by "Chirp." Google declined to comment. The Information previously reported that Google was plotting a competing version of Echo, a portable speaker with voice assistant tech. Sources said the device is unlikely to launch next week at Google's I/O developer conference, but plans are for it to land at some point this year. We should, however, get a peek at it and its potential next week — voice search and intelligent personal assistance will occupy center stage at the company's splash show, along with virtual reality. Google has long had voice assistant tech in its Android phones — beckoned by the words "Okay, Google" — that many in the industry see as leading the pack. (People inside Google think so, too.) But it has yet to bake that into the home, a key growing marketing for Google and its rivals. Its OnHub router, released last summer, does not have voice recognition capabilities. Amazon, on the other hand, has moved headlong into the home with Echo. One analyst estimated that Amazon has sold three million units. And Echo is collecting the type of data — what consumers search for, listen to and buy, and how they talk to machines — that Google loves. Amazon has long been considered a big threat to Google's core business as web and mobile app users go to the online retailer for product searches.

Daily Tech Snippet: Thursday, April 14

• Regulators plan to revoke Theranos’ federal license and ban founder Elizabeth Holmes: Theranos might find itself homeless soon. A federal agency plans to force founder Elizabeth Holmes out of her blood analysis startup for two years and take away the California lab’s federal license. First reported in the Wall Street Journal, the Centers for Medicare and Medicaid Services sent a letter dated March 18 proposing sanctions barring Holmes and company president Sunny Balwani from owning or running operations in labs for at least two years – including in  both California and Arizona – and taking away federal licensing for Theranos’ California facilities in Newark and Palo Alto after Theranos’ continued failure to correct major problems with accuracy and competence. These actions would be a major financial blow to the startup valued at $9 billion. Theranos has the runway to keep working with approximately $700 million in the bank but the two labs make a good portion of the money for Theranos’ operations and a loss of the founder and president would strangle any hope of recovery.


• Artificial Intelligence for Everyday Use: How four programmers with almost no knowledge of Japanese designed software to read handwriting. Real-world artificial-intelligence applications are popping up in unexpected places—and much sooner than you might think. While winning a game of Go might be impressive, machine intelligence is also evolving to the point where it can be used by more people to do more things. That's how four engineers with almost zero knowledge of Japanese were able to create software, in just a few months, that can decipher handwriting in the language. The programmers at Reactive Inc. came up with an application that recognizes scrawled-out Japanese with 98.66 percent accuracy. The 18-month-old startup in Tokyo is part of a growing global community of coders and investors who are harnessing the power of neural networks to put AI to far more practical purposes than answering trivia or winning board games. "Just a few years ago, you had to be a genius to do this," said David Malkin, who has a Ph.D. in machine learning but can barely string two Japanese sentences together. "Now you can be a reasonably smart guy and make useful stuff. Going forward, it will be more about using imagination to apply this to real business situations." While handwriting recognition might be considered deep-learning 101, Japanese is a whole other ballgame. That's because the language includes symbolic characters such as kanji, which is composed of elements that can be read independently, making it difficult to know where one ends and another begins. There are also more than 2,000 common pictograms made up of dozens of strokes. The trick is to tackle one squiggle at a time. Reactive’s algorithm queries the neural network for a match, adds another stroke and repeats, all the while refining the probability of an accurate hit. The startup trained its model on about 1.8 million characters. Unlike a typical program built around rigid rules, deep-learning AI is modeled on how humans process information. Given enough data as inputs and a set of desired outputs, neural networks figure out what goes in the middle. This allows them to find solutions that have bedeviled traditional approaches, like interpreting speech or tagging images. And once built, a neural network doesn’t have to be limited to language applications. In their spare time, the four Reactive engineers showed the program 5,000 dresses downloaded from Google Images, then gave it a picture of a woman in a revealing outfit. "Sexy clothes," the software responded.

• Some Online Bargains May Only Look Like One: Amazon has some unbelievable bargains on its virtual shelves. A cat litter pan with a list price of $2,159 can be yours for a mere $28. A bag of doggy treats, normally $822, is only $8. A windshield wiper blade, which the unwary pay $1,504 for, has been knocked down 99 percent. You say you don’t believe that a plastic cat pan could ever have been sold to anyone for a couple of thousand bucks? Or that a six-ounce bag of Zuke’s Lil’ Links pork and apple sausage bits ever cost more than dinner at a five-star restaurant? It’s all part of the bizarre world of Internet “discounts,” which let retailers and brands assert that you are getting a stupendous deal because someone somewhere else — exactly where is never explained — is being charged much more. Boomerang Commerce, a retail analytics firm, compared the list prices of dozens of pet items on Amazon and the specialist pet site Chewy.com. In only a handful of cases did the retailers even agree on what the list price was. So a 22-pound bag of Blue Buffalo Basics Limited Ingredient Grain-Free Duck and Potato dog food had a list price of $131 on Amazon and $84 on Chewy. Yet the retail price at both sites was the same: $49.49. “A perceived deeper discount creates a higher conversion event — in other words, more buyers,” said Boomerang’s chief executive, Guru Hariharan, who previously worked at Amazon. Another consultant, Ripen eCommerce, analyzed 746,000 product searches on Amazon. Ripen’s goal was to help third-party clients who sell on the giant retailer jockey for a better position — say, on the first page of results rather than further back. A little over 44 percent of the products — some sold directly by Amazon, others by third parties — were billed as discounted, Ripen said. “It’s less than I expected, actually,” said Dave Rekuc, Ripen’s director of marketing. “Considering you can basically name your own list price.”

• Verizon bets on Armstrong, M&A savvy in Yahoo bid:  Verizon is the clear favorite in the upcoming bidding for Yahoo's core Internet business, according to Wall Street analysts, in large part because the telecommunications company's efforts to become a force in Internet content have gone relatively well under the leadership of AOL Inc Chief Executive Tim Armstrong. Verizon acquired AOL last June for $4.4 billion - its first big foray into the advertising-supported Internet business - and it is not yet clear how well the unit is performing financially. Subsequent moves, including the takeover of much of Microsoft Corp's advertising technology business, a deal to buy Millennial Media for about $250 million and the recent launch of the mobile video service go90, are also too recent to assess. Yet analysts have given the big phone company high marks for allowing AOL to operate independently and folding in other recent acquisitions without much drama. And they said Armstrong seems to be driving Verizon's recent moves in go90 and recent acquisitions. Verizon showed interest in Yahoo's core business as early as December, when Chief Financial Officer Fran Shammo said the company would "see if there is a strategic fit" for Yahoo's holdings, which include mail, news, sports and advertising technology. Yahoo, under pressure from activist investors, launched an auction of its core business in February after it shelved plans to spin off its stake in Chinese e-commerce giant Alibaba. The first round of bidding is slated for next week, and Verizon plans to make a bid, sources familiar with the matter have told Reuters. Verizon is already working on increasing revenue through its ad-supported mobile video service go90, targeted at millennials and built on video streaming technology acquired from Intel in 2014. The app, which launched in October, offers videos from Comedy Central and Vice, among others, as well as basketball and football games. However, analysts cautioned that even a combined Yahoo-AOL would lack the unique data, such as user interests and tastes, that powers its rivals in online ads, chiefly Google and Facebook. Armstrong, who made his name leading sales at Google, is highly regarded in the advertising community - in contrast to Yahoo CEO Marissa Mayer, another former Google high-flyer, who has been struggling to revive Yahoo. Mayer would likely leave after a Verizon-Yahoo deal, analysts sa

• More Startups Are Getting Lower Valuations Than Joining the Billion-Dollar Club: According to a new report from KPMG International and CB Insights, 2016 has seen a larger number of startups taking new lower valuations than those earning the billion-dollar badge. “The first quarter of 2016 has borne witness to high-profile unicorn company issues, layoffs, down rounds and mutual fund valuation markdowns,” according to the report. Only five venture capital–backed companies entered the $1 billion club in the same period, less than half the number from any quarter since the first quarter of 2015. Meanwhile, CB Insights’ Downround Tracker shows there were 19 "down events"—or companies raising new money or being acquired at a lower valuation—during that same time frame, including big names such as Foursquare Labs Inc., Gilt Groupe Inc., and Jawbone Inc. Those downgrades may also cause other startups to wait to raise new money if they anticipate having to take cuts themselves.

Daily Tech Snippet: Tuesday, April 5

• Amazon Plans Big Push to Expand Prime Now Fast Delivery: Amazon.com Inc. plans to broaden the reach of its fast delivery service Prime Now, and is selling major brands promotional deals connected to the expansion, a sign the world’s largest Internet retailer is satisfied with early results from the nascent offering. The service -- now only available through the Prime Now app on smartphones -- will be run on Amazon’s website starting in May, according to documents reviewed by Bloomberg. Getting Prime Now on the Web puts the service in front of a larger audience, many of whom may not have downloaded the app on their phones. While shopping on mobile devices is expected to reach $96.2 billion in the U.S. this year, that represents a quarter of all e-commerce, according to the research firm EMarketer. Amazon is trying to sell advertising space to major brands for the Web launch, promising them visibility with tens of millions of Amazon shoppers. The premium "Launch Hero Package" would cost $500,000 for about two weeks of placement on Amazon’s website associated with the rollout. That price includes e-mail promotions sent to Amazon customers, which Amazon said have a stand-alone value of $100,000, according to the documents reviewed by Bloomberg.

• Salesforce.com Acquires Deep Learning Startup MetaMind: Cloud business software company Salesforce.com has acquired MetaMind, a startup focused on artificial learning that had been funded by Salesforce CEO Marc Benioff and venture capital firm Khosla Ventures. Terms were not disclosed, but it has all the markings of an “acqhire” sort of deal. Founder and CTO Richard Socher announced the deal in a post on the MetaMind website. Socher says on his personal website that his new title is Chief Scientist at Salesforce. MetaMind’s area of expertise is deep learning, the subset of artificial intelligence focused on data processing that is en vogue with Google, Facebook and other tech companies. The startup’s specialty is natural language processing — allowing computers to analyze relationships between words. Some of its capabilities disclosed in published researchdescribe advancements in the field that outrank those of some of the larger tech giants. Socher, a Stanford PhD, has said MetaMind’s plan was to sell this technology as a service to other companies. At MetaMind, he led a tiny group of researchers who gained a reputation for some rapid breakthroughs in the area of artificial intelligence. A Re/codereport from 2014 described how after only four months, the team came in only slightly behind Google in the ImageNet competition, in which companies compete to build systems that can recognize images.

• Toyota expands Microsoft partnership in connected vehicle services: Toyota Motor Corp is expanding a five-year-old partnership with Microsoft Corp to develop new internet-connected vehicle services for owners and dealers, Toyota said on Monday. The automaker has established Toyota Connected at its U.S. headquarters in Plano, Texas, to consolidate its existing connectivity services and serve as the company's "data science" hub. Microsoft has a 5 percent stake in the venture. Among the services to be developed or expanded under Toyota Connected are insurance coverage and rates based on owners' actual driving patterns; connected vehicle networks that can share information on traffic and weather conditions, such as icy roads; and information services tailored to a driver's habits and preferences, including monitoring heart rate, glucose level and other personal health data. The new wireless services will use Microsoft's cloud-based Azure platform. Toyota launched its initial partnership with Microsoft in 2011. Ford Motor Co introduced a similar program with Microsoft a year ago, and BMW AG and Nissan Motor Co announced Azure-based services earlier this year. Toyota Connected will use Microsoft's cloud technology to develop "predictive, contextual and intuitive services" to "humanize the driving experience while pushing the technology into the background," Toyota said. Toyota Connected also will consolidate the automaker's current initiatives in data analytics, data management and data services for dealers and fleet customers. In addition, the new organization will provide support for Toyota's ongoing research in robotics and artificial intelligence, as well as development of self-driving cars.

• Microsoft’s mobile problem may not be a problem at all: When Microsoft announced its Windows 10 strategy last year, the thinking was that the unified platform would drive Windows Mobile and finally bring the Windows phone out of the doldrums where it’s been virtually forever. The idea was you could develop once for Windows 10 desktop and easily share that code on any device, making it impossibly attractive for developers, which would finally drive Windows Mobile popularity in a beautiful virtuous development cycle. Unfortunately, it hasn’t worked out that way, and Microsoft finds itself in an unusual position, developing software for iOS and Android because it simply doesn’t have a viable Windows mobile ecosystem. The question remains; can Microsoft succeed without a strong Windows mobile position? From the looks of things, they don’t seem to have much choice. Nadella appears to be staking his position in the cloud, which is a perfectly reasonable way to play it, while opening up his company’s tools to iOS and Android in the absence of any meaningful Windows phone adoption. When you look at the beauty of the mobile-cloud connection, it’s understandable Microsoft would want to be there with Windows, but perhaps Nadella is beginning to understand that Windows is not necessarily the future of the company — Azure and Office 365 are — and that could explain why the company stayed firmly focused on these two areas at Build. When you combine that with the idea of bots created by Microsoft, including Cortana (Microsoft’s talking virtual assistant), that can run in Microsoft’s tools or external platforms like Slack and LINE, you start to see a vision where Microsoft thrives even without an in-house mobile platform. As the world moves swiftly to that mobile-cloud intersection, perhaps the underlying OS becomes less important. If that’s the case — if Microsoft can have a piece of the underlying cloud-mobile plumbing and have apps and bots created in its ecosystem, run anywhere on any device — it renders the Windows phone gap irrelevant.

Daily Tech Snippet: Monday, March 21

• A beginner’s guide to finally buying a virtual reality headset: if you are interested in being an early adopter,  here's a quick guide of the basics, plus a little input from my experiences with these products. Sony Playstation VR (PS VR) Buy if: You have a PlayStation already, or are looking to make a slightly smaller investment. Oculus Rift: After a long wait, Oculus opened preorders for the Rift headset, the first of which are expected to arrive at the end of March. The Rift is due to hit store shelves in April. Buy if: You want a stellar experience over everything else. Oculus was the first really big name to come out of the VR space and has probably done the most to minimize motion sickness. Samsung Gear VR: Powered by Oculus's technology and Samsung's smartphones, the Gear VR was first released in 2015 and is getting a renewed PR push with the new Galaxy S7 and S7 Edge smartphones. Buy if: You're really watching your budget and are happy with some smaller-scale experiences. HTC Vive: The product of a partnership between Taiwanese tech giant HTC and the video game company Valve, the HTC Vive is due to ship its first orders in April. Buy if: You really want an early version of a Star Trek-style Holodeck and have the room to make one.

• Uber seeking to buy self-driving cars: source: Ride-hailing service Uber has sounded out car companies about placing a large order for self-driving cars, an auto industry source said on Friday. "They wanted autonomous cars," the source, who declined to be named, said. "It seemed like they were shopping around." Loss-making Uber would make drastic savings on its biggest cost -- drivers -- if it were able to incorporate self-driving cars into its fleet. Earlier on Friday, Germany's Manager Magazin reported that Uber had placed an order for at least 100,000 Mercedes S-Class cars, citing sources at both companies. The top-flight limousine, around 100,000 of which Mercedes-Benz sold last year, does not yet have fully autonomous driving functionality. Auto industry executives are wary of doing deals with newcomers from the technology and software business who threaten to upend established business models based on manufacturing and selling cars. "We don't want to end up like Nokia's handset business, which was once hugely profitable...then disappeared," a second auto industry source said about doing a deal with Uber. A key hurdle to driverless cars has been the question of liability in the event of an accident. Most countries are signatories to the 1968 United Nations Convention on Road Traffic which stipulates that a person, rather than a computer, must be in control of a vehicle. In February this year, U.S. vehicle safety regulators softened the rules to allow driverless cars, by saying an artificial intelligence system piloting a self-driving Google car could be considered the driver under federal law, a major step toward ultimately winning approval for autonomous vehicles on the roads.

• How real businesses are using machine learning: There is no question that machine learning is at the top of the hype curve. And, of course, the backlash is already in full force: I’ve heard that old joke “Machine learning is like teenage sex; everyone is talking about it, no one is actually doing it” about 20 times in the past week alone.But from where I sit, running a company that enables a huge number of real-world machine-learning projects, it’s clear that machine learning is already forcing massive changes in the way companies operate. So where is it happening? Here are a few behind-the-scenes applications that make life better every day. Making user-generated content valuable: The average piece of user-generated content (UGC) is awful. It’s actually way worse than you think. It can be rife with misspellings, vulgarity or flat-out wrong information. But by identifying the best and worst UGC, machine-learning models can filter out the bad and bubble up the good without needing a real person to tag each piece of content. Pinterest uses machine learning to show you more interesting content. Yelp uses machine learning to sort through user-uploaded photos. NextDoor uses machine learning to sort through content on their message boards. Disqus uses machine learning to weed out spammy comments. Finding products faster: Successful e-commerce startups from Lyst to Trunk Archive employ machine learning to show high-quality content to their users. Other startups, like Rich Relevance and Edgecase, employ machine-learning strategies to give their commerce customers the benefits of machine learning when their users are browsing for products. Engaging with customers: You may have noticed “contact us” forms getting leaner in recent years. That’s another place where machine learning has helped streamline business processes. Instead of having users self-select an issue and fill out endless form fields, machine learning can look at the substance of a request and route it to the right place. Understanding customer behavior: Machine learning also excels at sentiment analysis. And while public opinion can sometimes seem squishy to non-marketing folks, it actually drives a lot of big decisions. For example, say a movie studio puts out a trailer for a summer blockbuster. They can monitor social chatter to see what’s resonating with their target audience, then tweak their ads immediately to surface what people are actually responding to - that puts people in theaters.
  
  
• Why unicorns falter: In early February 2016, a study of financing deals reported by The Wall Street Journal found that investors are increasingly protecting themselves from IPOs that don’t perform as expected. This fallout is a continuation of the demise of the so-called “unicorn,” a tech startup with a pre-IPO valuation of over one billion dollars. As these companies secure late-stage funding before their public market exit, smart private investors are setting terms that ensure they don’t lose a dime if the IPO falls short of expectations. This comes at a great cost to the startup if the exit doesn’t deliver, as was the case for many of the IPOs of 2015. The unicorn investment cycle has been consuming the growth ramp of an IPO-bound company. Unlike previous eras when a public exit occurred earlier in the company’s growth, leaving the best days ahead of the company, the fastest growth for an IPO-bound startup now happens in the last funding rounds before an IPO. This leaves a 20-30 percent growth rate post-IPO, which is pretty good for a company at $100-$200 million/year revenue, but bad for anyone looking for greater than 2X ROI from an IPO investment. Addressing these issues requires a little course correction as companies work toward an IPO. To ensure ample room for future growth, a startup should be careful not to push its market cap too high by taking more funding rounds than needed during the growth-stage period before IPO. This can be a challenge because funding often generates media interest and credibility, which are certainly not things a young company wants to leave on the table. However, leaving a portion of its growth for the IPO will ensure that the company has enough runway to continue to grow and deliver for its public market investors, just as the company has done for its VCs. Otherwise, you create yet another unicorn where the late-stage investors garner all the potential gains, and even force guarantees on returns. This is bad for new investors in the open market, and worse for the employees of the company who only receive poor post-lockup stock performance as compensation for years of hard work and sacrifices.

• Facebook's Zuckerberg meets propaganda czar in China charm drive: Facebook's co-founder and CEO Mark Zuckerberg met China's propaganda tsar Liu Yunshan in Beijing on Saturday as part of a charm offensive in one of the few markets where the social network cannot be accessed. The rare meeting, reported by China's state news agency Xinhua, suggests warming relations between Facebook and the Chinese government, even as Beijing steps up censorship of and control over the Internet. Liu, who sits on the Communist Party's Politburo Standing Committee which is the apex of power in China, praised Facebook's technology and management methods, Xinhua said. Zuckerberg was in Beijing for the China Development Forum, a government-sponsored conference bringing together top business executives and the country's ruling elite. China "hopes (Facebook) can strengthen exchanges, share experiences and improve mutual understanding with China's Internet companies", Xinhua quoted Liu as telling Zuckerberg. On Friday, Zuckerberg posted an image of himself running through smog in Beijing's Tiananmen Square, past the portrait of the late Chairman Mao Zedong hanging over the Forbidden City. The 31-year-old has achieved celebrity status in China, one of the few markets where Facebook and other foreign Internet platforms, including Alphabet Inc's Google services and Twitter Inc, are not available due to tight government controls. He has long sought to improve his company's relationship with the Chinese authorities, and now sits on the advisory board of the School of Economics and Management at China's elite Tsinghua University.

.