|
|
|
|
|
Companies Adjust To Covid-Skewed Data; Smart Software Lifts Spending Plans
|
|
|
|
|
|
Welcome back. The marketplace chaos sparked by the Covid-19 pandemic has left many companies with jagged trend lines reflecting wild swings in sales, for better or worse. That makes much of their data from the past year unreliable for predicting consumer behavior as economies reopen. To compensate, retailers, home-delivery firms and travel-booking services—which were hit especially hard by economic turmoils—are downplaying pandemic-era data in predictive models. Instead, many are leaning more heavily on real-time data, or digging deeper into historic trends that predate the onset of the coronavirus at the start of 2020. As software developers say, smart algorithms are only as good as your data.
|
|
|
|
|
|
A driver for an independent contractor to FedEx scans packages outside a truck in San Francisco.
PHOTO: DAVID PAUL MORRIS/BLOOMBERG NEWS
|
|
|
|
Predictive Models Get a Makeover
|
|
Retail, travel and logistics companies are putting less weight on what they see as skewed data gathered during Covid-19 restrictions and other chaotic market conditions, which are distorting AI models and producing unreliable trend lines for customer demand, prices and supply chains, The Wall Street Journal reports.
Delivering results. FedEx is putting greater emphasis on real-time data to provide accurate delivery-time estimates, to offset abnormal historic data caused by the surge in home-deliveries during Covid lockdowns.
A bigger cart. Walmart is downplaying last year’s shopping trends by including a larger share of historic data, industry reports and broader economic forecasts in predicting future sales.
|
|
|
As pandemic restrictions ease and the economy improves, U.S. companies across several industries are greenlighting spending plans on information technology, with an emphasis on software related to process automation, AI and security, WSJ’s Jared Council reports.
|
|
|
U.S. companies are expected to spend more on AI and other software applications.
|
|
|
Back on track. Tech budgets at U.S. businesses and governments are expected to rise to nearly $2 trillion this year, up 7.4% from 2020, beating weaker projections made earlier in the year, according to market researcher Forrester Research Inc.
|
|
|
Back to business. Investments in data, analytics and AI, “help us to create new experiences for quoting, binding, payments and claims,” Liberty Mutual Group CIO James McGlennon says.
|
|
|
|
|
|
Intel’s ControlFlag uses statistical analysis and machine learning to automatically detect errors.
PHOTO: INTEL
|
|
|
|
|
Large technology providers like Intel, Amazon.com and Microsoft are creating smart-software tools designed to find and fix software bugs, as errors get harder to spot in increasingly complex lines of code, WSJ’s Sara Castellanos reports.
Time suck. About half of developers’ time is spent debugging software, which can take weeks to fix, says Justin Gottschlich, director of machine-programming research at Intel Labs.
Costly errors. Poor software, including operational failures, cost the U.S. economy more than $2 trillion last year, according to the Consortium for Information & Software Quality, an IT industry group.
|
|
|
|
$482 billion
|
The amount U.S. companies are expected to spend next year on AI, cybersecurity and other software applications, according to Forrester Research Inc.
|
|
|
|
|
|
|
|
ILLUSTRATION: CHASE VOORHEES FOR THE WALL STREET JOURNAL
|
|
|
|
Ghost in the Machine Learning
|
|
Within the expanding market for chatbots and animatronic robots, researchers and entrepreneurs are exploring ways AI could be used to create digital versions of people after their deaths that live with their loved ones and continue to work, The Wall Street Journal reports.
Taking many forms. Tech startup Replika has developed an app that learns to replicate a person in the form of a chatbot, while HereAfter AI records people’s life stories and uses them to create a replica embedded in a smart speaker.
Getting in the spirit. Microsoft Corp. recently patented a method of using chatbots to preserve historical figures and living people, though a company spokeswoman says as yet there is no plan to use it.
|
|
|
‘Roswell: The Final Verdict’, a new Discovery+ series, uses advanced AI in efforts to determine if tales of an alien spaceship crash landing in the New Mexico desert 75 year ago are true or an elaborate hoax, WSJ reviewer John Anderson writes.
|
|
|
A scene from ‘Roswell: The Final Verdict’
PHOTO: DISCOVERY+
|
|
|
Digital lie detector. The series analyzes decades-old testimony of people claiming to have witnessed or have knowledge of the event, deploying smart software designed to scrutinize body movements, facial gestures and language for biometric signs of lying.
|
|
|
The results. At least one witness who claims that evidence of Roswell aliens reached President Harry Truman is determined to be lying, though the software's developer says someone who sincerely believes he or she was telling the truth might be judged as having told the truth.
|
|
|
|
|
“Whereas photographs could substitute the body of the dead in a static way, today digital technologies give motion and activity to this ‘eternal’ image.”
|
— Davide Sisto, a philosopher at the University of Turin.
|
|
|
|
|
|
WSJ Pro AI Extra: Smart Homes Get Smarter
|
|
|
|
|
ILLUSTRATION: RUTH GWILY
|
|
|
|
Make Way For More Gadgets
|
|
Researchers are working on a wave of sophisticated devices for the home, from toilets that screen for viruses to a ‘Roomba on steroids’ that might detect your TV remote in the couch, The Wall Street Journal reports. Here are a few top picks:
-
Smart toilets. Researchers at Stanford University have developed a prototype toilet that uses an AI-trained camera to track the form of feces and monitor the color and flow of urine for health warning signs.
-
Electronic noses. Stratuscent, a Montreal-based tech startup, last year launched a chip that can detect and catalog toxic fumes and volatile chemicals in indoor spaces.
-
Lost and found. Researchers at the Massachusetts Institute of Technology’s Signal Kinetics research group have developed a robot that uses AI to find lost TV remotes or misplaced keys.
|
|
|
|
|
Jim Whitehurst will continue to serve as an adviser to IBM.
PHOTO: DAVID PAUL MORRIS/BLOOMBERG NEWS
|
|
|
|
Cloud leader leaves IBM. Jim Whitehurst, a key figure in International Business Machines Corp. Chief Executive Arvind Krishna’s cloud and AI strategy, is leaving the company two years after he joined through its acquisition of cloud-software firm Red Hat. (WSJ)
Free-speech app to use AI content screening. Former Trump advisor Jason Miller says Getr, a new free-speech-focused social-media platform backed by fugitive Chinese businessman Guo Wengui, will use software moderators to block threats, hate speech, harassment and other barred content. (WSJ)
Intel delays chip production. In a setback for Intel Chief Executive Pat Gelsinger’s turnaround efforts, the chipmaker is delaying production of its next-generation central processing units servers until next year, rather than a previous target of late 2021. (WSJ)
Astronomers tune out noisy galaxies. Japanese researchers are using AI to remove distortions in astronomical data caused by random variations in galaxy shapes, an effort that could lead to more accurate measurements of the universe. (SciTechDaily)
Federal lab explores atomic chips. Argonne National Laboratory researchers are looking at ways to use AI to make atom-sized films of materials, a process known as atomic layer deposition, to design advanced computer chips. (NextGov)
Startup Pits AI against AI. Parity, a New York-based startup, uses AI-enabled software to analyze the data, technologies and methods used by companies to develop AI models, in order to pinpoint and fix instances of bias. (The New York Times)
Seventh grader wins tech prize. Aariv Modi, a 13-year-old student at Walsh Middle School in Round Rock, Tex., was named the Voice/AI Pioneer of the Year at technology conference Project Voice, for blog posts teaching kids how to make Alexa do things it isn't programmed to do. (Austin American-Statesman)
|
|
|
The boss of the company at the heart of a widespread cyberattack hack, that started to unfold Friday and has affected hundreds of businesses, said he briefed the White House and that attackers are demanding a single $70 million ransomware payment. (The Wall Street Journal)
Facebook, Twitter and Alphabet’s Google have privately warned the Hong Kong government that they could stop offering their services in the city if authorities proceed with planned changes to data-protection laws that could make them liable for the malicious sharing of individuals’ information online. (WSJ)
A unit of China’s cybersecurity regulator launched data-security reviews of apps operated by two U.S.-listed Chinese companies, days after announcing a similar probe into ride-hailing giant Didi Global Inc. (WSJ)
Nextdoor Inc., which operates an app that connects people to information about their neighborhoods, is merging with a special-purpose acquisition company to go public in a deal valuing the social-media network at roughly $4.3 billion. (WSJ)
|
|
|
|
|
|