AI: Semi-Capable? Absolutely!
November 20, 2019
Reporter Jeremy Kahn at Fortune ponders an important question—“A.I. Is Everywhere—But Where Is Human Judgment?” Kahn recently spent a week at the Web Summit in Lisbon, where he learned just how much machine learning has taken over at many companies. From product recommendations to delivery-drone operation to the prevention of crime, algorithms are making many real-world decisions. For example, Amazon CTO Werner Vogels explained at the conference that machine learning is at the heart of “absolutely everything” at his company. This takeover really took off in 2015, when Amazon took up deep learning and found its forecasts became 15 times more accurate. The article also notes that Mastercard uses predictive analytics to foil cyber attacks and fraudulent activity by organized crime.
After the rah-rah conference, however, Kahn found some sobering news in a report from the National Transportation Safety Board. He writes:
“While Web Summit was all about the promise of A.I., this news from last week ought to give people pause: the National Transportation Safety Board released its preliminary report investigating how one of Uber’s self-driving cars came to strike and kill 49-year old Elaine Herzberg as she crossed the road in Tempe, Arizona, last year. The NTSB found in its ‘Vehicle Automation Report’ that while the car’s sensors did detect Herzberg six seconds before hitting her, the self-driving system failed to correctly classify her as a pedestrian, in part because Uber had trained its computer vision system to only expect pedestrians in designated cross-walks. What’s more, the agency concluded that Uber’s engineers had programmed the car to only brake or take evasive maneuvers if its computer systems were highly confident that a collision was likely. Humans decided to train the system in this way and set these tolerances. Most likely, this was done to prioritize the comfort of Uber’s passengers, who would have found sudden braking and unexpected swerves annoying and alarming. And it’s ultimately these human decisions that doomed Herzberg.”
The piece concludes with a simple but important suggestion—anyone involved with deploying A.I. with real-world impact should read the NTSB report. That is a good idea.
Cynthia Murrell, November 20, 2019
Palantir and Sompo: Is a $150 Million Deal Big Enough, Too Small, or Just Right
November 19, 2019
Palantir Technologies has ingested about $2 billion in a couple of dozen investment rounds. Now a $150 million deal is very important to a services firm with a few million in sales. To an outfit like Booz, Allen or Deloitte, $150 million means a partner will keep her job and a handful of MBAs will be making regular flights to wonderful Narita.
“Thiel Marks Palantir’s Asia Push with $150 Million Japan Venture” reports that Sompo Holdings is now Palantir’s partner, noting that the $150 million may be more of an investment. We noted this passage:
The billionaire entrepreneur [Peter Thiel] was in Japan Monday to unveil a $150 million, 50-50 joint venture with local financial services firm Sompo Holdings Inc., Palantir Technologies Japan Co. The new company will target government and public sector customers, emphasizing health and cybersecurity initially. Like IBM Corp. and other providers, Palantir’s software pulls together a range of data provided by its customers, mining it for patterns and displaying connections in easy-to-read spider web-like graphics that might otherwise get overlooked.
Bloomberg reported:
Palantir is very close to breaking even and will end 2019 either slightly in the black or slightly in the red, Thiel said at the briefing. The company will be “significantly in the black” next year, he added.
A few comments from the DarkCyber team:
- The money in the headline is not explained in much detail. There is a difference between setting up a new company and landing a cash deal.
- Bloomberg seems indifferent to the revenue challenge Palantir faces; namely, there are quite a few investors and stakeholders who want their money plus interest. The announcement may not put these individuals’ minds at ease.
- The news story does not mention that new, more agile companies are introducing solutions which make both IBM Analysts Notebook and Gotham look a bit like Vinnie Testaverde or Bart Starr throwing passes at a barbeque.
Singapore is the location of choice for some of the more agile intelware and policeware vendors. Is Japan is a bit 2003?
To sum up, Palantir is to some a start up. To others Palantir is an example of a company that may lose out to upstarts which offer a more intuitive user interface and slicker data analytics. It is possible that an outfit like Amazon and its whiz bang data market place could deliver a painful blow to a firm which opened for business in 2003. That’s more than 15 years ago. But next year? Palantir will be profitable.
Stephen E Arnold, November 19, 2019
The Google Will Pursue the Pentagon
November 19, 2019
Despite last year’s Project Maven kerfuffle, Google cannot pass up the (lucrative) chance to collaborate with the US military. Roll Call reports, “Google Looks Past Project Maven to Work Anew with the Pentagon.” We’re told the renewed partnership will not involve the development of weapons or unlawful surveillance, and stays true to guidelines the company has now adopted. Writer Gopal Ratnam cites Kent Walker, senior vice president for global affairs at Google:
“‘It’s right that we decided to press the reset button until we had an opportunity to develop our own set of AI principles, our own internal standards and review processes,’ Walker said last week at an event organized by the National Security Commission on Artificial Intelligence [NSCAI]. The decision to stop working with the Pentagon on the drone video contract was a ‘discrete’ one and not indicative of a ‘broader principle or an unwillingness’ to work with the Defense Department, he said. “The commission was created by Congress in the 2019 Pentagon policy bill to figure out how the Pentagon can harness artificial intelligence and related technologies for national security purposes. The panel is chaired by Eric Schmidt, the former chairman of Google’s parent company Alphabet, where he continues to be an outside adviser. “Google is working with the Pentagon on a ‘number of national mission initiatives,’ Walker said, listing cybersecurity, health care, tools to identify deep fake videos, and other AI projects with the Defense Advanced Research Projects Agency, or DARPA, to ensure the ‘robustness of AI.’”
Note that former Google CEO Eric Schmidt chairs the NSCAI. Perhaps that is just a coincidence. Walker paints their work with the Pentagon as motivated by their patriotism and concern for world safety. I suppose, then, that government dollars are just the icing on the AI cake.
Cynthia Murrell, November 19, 2019
The Flu Tracking Thing: A Hollywood-Style Repeat?
November 19, 2019
Guess not.
Perhaps we should be asking more often, “Do we really need an app for that?” Engadget announces, “Weather Channel App Uses Watson to Tell You When Flu Season Is Coming.” The write-up specifies:
“The latest version of the Weather Channel app for Android and iOS now includes a Flu Insights feature that uses IBM’s Watson to warn you when influenza is likely to be on the warpath. The team-up wields machine learning to provide a 15-day flu forecast as well notifications that pop up at key moments, such as the start of flu season or confirmed outbreaks. Ideally, this will keep you from going to a house party full of soon-to-be-sick virus carriers. Whatever the conditions are like, you’ll get prevention advice as well as CDC flu reports.”
On the other hand, Google tried its flu map back in 2008, and what happened to that? It was not too useful. We all know that flu season is upon us. According to the CDC, it happens every year from October through March, peaking December through February. And anyone can navigate directly to that site for prevention advice. Granted, that Watson-powered Flu Insights feature gives more details, like when and where confirmed outbreaks have occurred and a zip-code centered “flu forecast” that predicts risk from low to high. However, as the Engadget piece concedes:
“It’s easy to see Flu Insights fueling some paranoia. Do you really want to be afraid of going outside simply because Watson predicted trouble?”
For us, the answer is no. I, for one, prefer to simply get my flu shot in the fall and be done with it. No AI-prompted angst required.
Cynthia Murrell, November 19, 2019
DarkCyber for November 19, 2019, Now Available
November 19, 2019
The November 19, 2019, DarkCyber discussed Amazon’s patent US 10,296,764 B1 “Verifiable Cryptographically Secured Ledgers for Human Resource Systems.” Stephen tries his best to make this patent discussion thrilling. Well, perhaps “thrilling” may be stretching the discussion of the system and method disclosed in this 24 page disclosure. But there are some graphics and a number of statements which are probably too simple to satisfy a patent attorney. Nevertheless, if you are curious about Amazon and its invention for human resources, navigate to www.vimeo.com/373810982 and check out the program. This week’s program marks the start of “season two” of DarkCyber. More patents, an interview, and news stories will feature in the coming weeks. After celebrating three quarters of a century of semi-coherent thinking, DarkCyber will appear every two weeks. The interfaces implemented in the software Stephen uses slows him down. The team just tells him, “Okay, Boomer, work harder.” His response cannot be printed in this prestigious blog.
PS. In August, Stephen was quoted by the New York Times, in October by MIT’s Technology Review (yep, the Epstein friendly organization), and this month by Le Monde (that’s in Paris and in French no less). The subjects? Intelligence, Amazon, and the lack of awareness among certain residents of Harrod’s Creek to Stephen’s research. Hey, he lives in Kentucky which holds a proud place in the lower quartile of literacy in the US.
Kenny Toth, November 19, 2019
Open Source Goodness? Not So Fast
November 18, 2019
DarkCyber does not have a dog in the fight. Open source software has been an interesting sector. However, there may be some tension in open source land. If you have a stake in open source software, you will want to read “Venture Capital Shillscapegoating Free Software’s Failures.” I noted this statement in the article:
Venture capitalists and hireling lawyers make convenient scapegoats. The old, creaky pillars of the free software movement need convenient scapegoats, because the facts on the ground raise serious doubt about the effectiveness of their leadership and the byzantine, insular ideology of copyleft they impose. When the facts don’t help, substitute narrative. You can rule on narrative alone, at least for a while.
Free software has found its way into some interesting products and services. Some of these are backed by big money; for example, LucidWorks, Palantir, and even IBM Watson.
So what?
No answers shall be forthcoming from DarkCyber. You, gentle reader, are on your own to ponder the Amazon open source plays, the future of proprietary software stripped of open source goodness, and venture firms betting that the “community” will keep on being communal.
Stephen E Arnold, November 18, 2019
The Sharp Toothed MSN Gnaws on the Google Search Carcass
November 18, 2019
Search and retrieval is fraught with challenges. In the enterprise search sector, fraud has been popular as a way to deal with difficulties. In the Web search sector, the methods have been more chimerical.
MSN, a property of Microsoft, published “How Google Interferes With Its Search Algorithms and Changes Your Results.” The write up appears to recycle the work of the Wall Street Journal. The authors allegedly are Kirsten Grind, Sam Schechner, Robert McMillan and John West. It is unlikely that Alphabet Google will invite these people to the firm’s holiday bash this year.
What’s in the write up? The approximately 8,500 word article does the kitchen sink approach to sins. Religious writers boil evil down to seven issues. Google, it seems, requires to words to cover the online advertising firm’s transgressions.
DarkCyber will not engage in the naming of evils. Several observations are warranted:
- Google’s waterproof coating has become permeable
- After decades, “search experts” are starting to comprehend the intellectual impact of search results which has been shaped
- The old-fashioned approach of published editorial policies, details about updating indexes, and user control of queries via Boolean logic is not what fuels the Google method.
But so what? With more than 60 percent of search queries to the Google flowing from mobile devices, old school approaches won’t work. Figuring out what works depends on defining “works”.
Finding information is a big deal. What happens when one tries to hide information? The answers may be observed in the action of Google employees who have forced the company to stop communicating in “all hands” Friday meetings.
What’s Microsoft doing? For one thing, poking Googzilla in the eye with MSN articles is one example of Microsoft’s tactical approach. The other is to ignore problematic Windows 10 updates and “ignite” people to embrace a hybrid cloud paradigm.
And what about Microsoft’s own search technologies. One pundit apologist continues to explain that Microsoft search is just getting more efficient, not better.
Net net: Google and Microsoft may have more in common than some individuals realize. Maybe envy? Maybe techno-attraction? Maybe two black holes circling? Whatever. The situation is interesting.
Stephen E Arnold, November 18, 2019
Amazon Product Search: A Challenge for the GOOG
November 18, 2019
Amazon is gaining ground in the search-based advertising arena. ZDNet reports, “Amazon Search Ad Business to Whittle Away at Google Market Share Through 2021, Says eMarketer.” Citing a recent eMarketer report, writer Larry Dignan tells us that, though Google will remain top dog by a wide margin for the foreseeable future, Amazon is positioned to increase its share. He writes:
“The report finds that Google will continue to dominate search advertising, but its share will fall over time. Amazon is expected to show search ad revenue growth of 29.5% in 2019, 30.7% in 2020 and 26.2% in 2021. Amazon’s advertising business has surged past Microsoft to be No. 2 behind Google, which has 73.1% of the search ad market. Amazon will end 2019 with 12.9%, followed by Microsoft at 6.5%. Verizon Media and Yelp round out the top five with market share of about 2%.In addition, Amazon’s advertising business is closely watched among Wall Street analysts. The search ad business falls into Amazon’s ‘other’ revenue category and many analysts expect it to be a break out business like Amazon Web Services. Google’s market share in the search advertising market is expected to drop to 70.5% by 2021, according to eMarketer estimates.”
Amazon, you see, has a unique advantage—many active shoppers begin their product searches there, so they are already poised to make a purchase. Dignan adds that other retail sites like Wal-Mart, Target, and eBay are also nipping at Google’s search-ad market share.
Cynthia Murrell, November 18, 2019
AI Education: India Emulates UAE
November 17, 2019
The Indian Institute of Technology (IIT) located in Madras, India is dedicated to helping its students advance their knowledge and careers. India is the second most populous country in the world and it is still considered a developing nation, but it has one of the fastest growing technology economies. IIT Madras wants to help young Indians succeed and plans to double the amount of students enrolled in their Inter-Disciplinary Dual Degree for Data Science, which is an ambitious goal. IIT Madras is even more ambitious with its recent announcement published in Analytics India Mag: “IIT Madras Researches Develop ‘AISoft,” An Algorithm To Solve Engineering Problems.”
Assistant professor Vishal Nandigana, based in the Fluid Systems Laboratory in the Mechanical Engineering Department, led the development team that designed AISoft. AISoft was designed to solve problems across all engineering fields, including semiconductors, automobiles, thermal management, aerospace, and electronics cooling applications. It has already been used to solve problems in thermal management and the test group said it was faster at solving problems compared to existing solutions, such as CNN or C-GAN.
IIT Madras believes there is a brand new market for AI that has not been developed yet:
“The institute believes that AI, Machine Learning and Deep Learning are now being used for over a decade but traditionally only in areas such as signal processing, speech recognition, image reconstruction and prediction. Very limited attempts have been made globally in using these algorithms in solving engineering problems such as thermal management, electronics cooling industries, automobile problems like fluid dynamics prediction over a bonnet or inside the engine, aerospace industries like aerodynamics and fluid dynamics problems across an aero-foil or turbine engine.”
AISoft offers solutions on generalized rectilinear and curvilinear input geometry, but its biggest advantage is saving computational time. Computational time eats up a lot of hours, but AISoft can eliminate some of that time which can be better spent on research and applying knowledge.
IIT Madras hopes its startup plans will assist industries as well as be used in education. AISoft, if the costs are kept low, would also generate income for IIT Madras that can be used for scholarships, growing educational programs, and on research projects.
Whitney Grace, November 17, 2019
Google: The Emerging Cancel Culture
November 16, 2019
Google has terminated a number of products and services. My favorite is Web Accelerator, but you may have other candidates. The cancel phenomenon — whether practiced by Microsoft with its wonderful Zune product or Hewlett Packard’s fascinating Autonomy deal — means that big companies change their minds. Poof. Time, money, and maybe a customer are two are vaporized.
Cancelled. Some in government may use the phrase “with extreme prejudice” to signal this approach to an ill-advised decision, a wonky product, or a troublesome entity.
The Verge, a real news outfit, published “Google Is Scaling Back Its Weekly All-Hands Meetings after Leaks, Sundar Pichai Tells Staff.” The write up approach this cancel culture move as “scale back”, noting that the Verge stumbled upon an email from Google’s CEO to the Googlers. The Verge revealed:
In the note, Pichai begins by praising what Google has achieved through its large workforce. “But in other places — like TGIF — our scale is challenging us to evolve,” he writes. “TGIF has traditionally provided a place to come together, share progress, and ask questions, but it’s not working in its current form.” He writes that employees “come to TGIF with different expectations,” with some looking to hear about “product launches and business strategies” and others looking for “answers on other topics.” Only about 25 percent of the company watches the meeting each week, Pichai says. He also says that there has been “a coordinated effort to share our conversations outside of the company after every TGIF” and that those efforts have “affected our ability to use TGIF as a forum for candid conversations on important topics.”
“Google Will No Longer Hold Weekly All-Hands Meetings Amid Growing Workplace Tensions” explains:
Google is getting rid of one of its best-known workplace features: TGIF, its weekly all-hands meeting. The company confirmed to CNBC that it will instead hold monthly all-hands meetings that will be focused on business and strategy while holding separate town halls for “workplace issues.”
Yep, unfriended, terminated, modified, or cancelled. Mostly the same action spun in different ways.
Several observations:
- What’s the best way to avoid problematic staff? Avoid them? That’s one approach, and a path less fraught with legal hassles than firing the un-Googley.
- Google’s challenges span numerous legal hassles from US jurisdictions. Is it 50 for 50 now? Not even major leaguers can bat 1,000. Google can and is. How many strike outs await?
- The chest X-ray matter (please, see Fast Company’s story)
- The billion dollar dust up with Oracle is back in court, the Supreme Court no less. See the Silicon Angle story, please).
What’s up?
Google’s activities are increasingly interesting. My phrase for the firm’s approach to management is HSSCMM which is short hand for high school science club management method. What adds a handful of kokum to the digital stew served in the employees’ only cafeteria.
How many Googlers enjoy this rare and hard to find spice? Perhaps Googler’s analysts can quantify their data and provide some insight. A Google Trends diagram might show a curve like this one from Scientist Cindy?
Just cancel that. Unfriend!
Stephen E Arnold, November 16, 2019
-
- Subscribe to Beyond Search
Feature archive
News archive
-
