The Low-Down: 2019-09-15

Sep 21, 2019

Demand For More Powerful AI Is Spawning New Hardware Architectures

The intangible is changing the tangible. JL

Ben Dickson reports in Venture Beat:

As society turns to artificial intelligence to solve problems across ever more domains, we’re seeing an arms race to create specialized hardware that can run deep learning models at higher speeds and lower power consumption.Given the variety of industries and domains that are finding applications for deep learning, there’s little chance that a single architecture will dominate the market. But what’s certain is that the AI chips of the future will be very different from the classic CPUs that have been sitting in our computers and servers for decades.

As concerns about the vulnerabilities of anonymization grow, this new technique suggests there are technological solutions that can protect peoples' identities. JL

MIT Technology Review reports:

AI could generate faces that match the expressions of anonymous subjects to grant them privacy—without losing their ability to express themselves. The algorithm extracts information about the person’s facial expression by the position of the eyes, ears, shoulders, and nose. It uses a GAN to create an entirely new face with the same expression.Because GANs don’t use the subject’s original face at all, they eliminate risk of (discovery). It shows how GANs could also be applied to audio to anonymize voices.

It's a global brand, so has to translate. Plus, it has to pop on the smartphone screen as well as on the store shelf. JL

Robert Klara reports in Ad Week:

Globally, potato chips are a $29 billion industry. While Lay’s is the category leader—75.9% market share—there are 40 companies making chips in America plus regional brands. Why should it take two years to design a bag? The buyer of chips might as easily be an adolescent as an octogenarian and hail from any socioeconomic background. Hence, the design must have an accessibility. And because the purchase decision for a snack food is often made on the fly, a bag of chips has little time to distinguish itself. It’s selling everywhere from supercenter shelves to smartphones. So the design will be as recognizable on a store shelf as on a computer screen.

The University of Alabama has a dominant football team. But it's been winning so long - and running up the score against weak opponents - that droves of students leave games early.

The university is trying to address that by offering a location tracking app which offers better seats for later season playoff games if the phone's student owner stays through more regular season games. The assumption being they're not going to be studying on a Saturday afternoon anyway. JL

Tim Cushing reports in Tech Dirt:

The Alabama football coach has long been peeved that the student section empties early. This season, the university is rewarding students who attend games and stay until the fourth quarter with improved access to tickets to the SEC championship game and to the College Football Playoff championship game. But to do this, Alabama is taking an Orwellian step: using location-tracking technology from students’ phones to see who skips out and who stays. The reward for fourth quarter attendance -- 250 points -- is equivalent to 2.5 credit hours, showing how the university considers fourth quarter attendance in meaningless games.

Chinese youth are wealthier, on a relative basis, than their US counterparts. They have no student debt and are likely to have already been given a home by their parents.

All of which puts the US trade war in perspective, but makes sales prospects for luxury brands uncertain. JL

Carol Ryan reports in the Wall Street Journal:

The Chinese are the most important luxury shoppers in the world, buying one-third of all designer goods last year. They start making purchases in their 20s, two decades earlier than U.S. consumers. The U.S. savings rate is just 8%, and outstanding student loans hit $1.6 trillion in 2019. The average U.S. student has borrowings of $37,000, diverting incomes toward servicing debt rather than the latest designer collections.U.S. millennials are more likely to be saving for a deposit for a house than their Chinese peers. 34% of people under 35 own their home in the U.S. The figure for Chinese millennials is 70%

Almost all connected devices are delivering owners' usage data to third parties without permission or acknowledgement. JL

Karl Bode reports in Motherboard:

A new study has found that most “internet of things”devices - smart TVs, streaming dongles, smart speakers, and video doorbells made by vendors including Google, Roku, and Amazon - deliver sensitive data to partners around the world without making these data transfers secure or transparent to the user. “Nearly all TV devices contact Netflix even though we never configured any TV with a Netflix account.” This usage data can be used to build behavioral profiles of consumers who may not understand that daily habits gleaned from their TV set to smart electricity meter are being collected, cataloged, and monetized.

The growth of remote work is changing where people live, the way they live and how the spaces they rent or buy are designed.

As a result, it is having a profound impact on real estate layouts and geographic preferences. JL

Lisa Prevost reports in the New York Times:

“The importance of home offices has begun to rival the attention that buyers give to kitchens.”A survey of 23,000 new home shoppers found 30% worked at home between one and four days a week. Of employed persons worked part of the time at home; among those with advanced degrees, it was 42%. Americans are moving less often. If you work at home, you dont have to move if your job moves. New home sizes are trending downward. The home entertainment factor has gone down. Apartment designs provide buyers with extra work space without requiring them to pay for an additional bedroom. When choosing a town to live in, they want coffee shops with free Wi-Fi, co-working spaces, places to find community.

The Big One: AI Is Taking On Earthquake Prediction - And Appears To Be Succeeding

If consistently replicable, such AI-generated early warning would truly be 'The Big One.' JL

Ashley Smart reports in Quanta:

Groups are using machine learning to demystify earthquake physics and tease out warning signs of impending quakes. Using pattern-finding algorithms similar to those behind image and speech recognition and other forms of artificial intelligence, collaborators successfully predicted temblors in a model laboratory — a feat since duplicated in Europe. They’ve tested their algorithm on quakes in the Pacific Northwest. They indicate that the algorithm can predict the start of a slow slip earthquake to “within a few days — and possibly better.”

As the compilation and analysis of data expands exponentially, finding ways to make its storage and access more effective could be one of the greatest challenges - and financial successes - of the techno-socio-economic future.

And it appears AI can help achieve that goal. JL

Jim Salter reports in ars technica:

As the scale and complexity of storage workloads increase, it becomes more and more difficult to manage them efficiently. With thousands of services competing for resources with differing performance and confidentiality targets, management of storage outpaces the human ability to make informed and useful changes.An AI architect might choose a convolutional or recurrent neural network to discover patterns in storage availability. Neural networks learn to spot anomalies and performance problems. AI management may also be able to provide a degree of efficiency not otherwise possible.

These were not those cute - if sometimes annoying - little quadro-propellar drones seen at the beach or local park.

They were more like subsonic cruise missiles capable of sophisticated evasive maneuvers - and carrying large explosive payloads. JL

Kyle Mizokami reports in Jalopnik:

How did the Saudi military, with a defense budget of $67.6 billion, allow this drone attack to cause such massive economic damage? Saudi Arabia has large numbers of Patriot missiles, but most of the missile batteries were looking south. The limited coverage arc of the radar makes it easy to fly around—particularly if you’re coming from the east. Drones encompass a variety of objects, from quadcopters to high speed subsonic pilotless aircraft similar to cruise missiles, which fly at low altitudes, making detection with ground-based radars difficult.

There will be legal challenges based on the Constitutional right to be assumed innocent until proven guilty, a presumption undermined by current uses of facial recognition.

That, plus growing concern about abuses of the technology, may limit its spread. JL

Charlotte Jee reports in MIT Technology Review:

“Proper use” of facial recognition by government is supported by 80% of Americans. Without specific examples of what proper use is or is not, though, it’s hard to be sure of public opinion. “There will be legal challenges, and there will eventually be regulation. A constitutional right we have is innocent until proven guilty. Facial recognition could flip that around." A campaign in New York by tenants to stop a plan for using facial recognition instead of keys to access their apartments mostly affected poor, black, and brown women. The tenants involved human rights lawyers, and more affluent groups started to ally with them.

Chinese and foreign companies will be surveilled, evaluated and assigned a score, which will impact everything from export and import licenses to access to new business opportunities.

While a good score may not necessarily improve a corporation's lot, a bad score can definitely hurt it. JL

Nathaniel Taplin reports in the Wall Street Journal:

A key target of China’s coming “social credit” system, which triggers visions of “1984”-style monitoring of people, is misbehaving businesses. 80% of information on the data-sharing platform relates to companies. Social credit will make falling afoul of regulations more costly. Compliance or noncompliance with important regulations will be assigned a value and fed into an algorithm to produce a company’s overall rating shared across agencies through a central database.A bad rating will have ripple effects. Data will be gathered from company submissions and inspections, but also from video surveillance, instrument monitoring and third-party sources.

It's all about attempts to grab attention, engage and monetize. And if consumers want to mention your brand on their social media feed, so much the better. JL

Josh Sternberg reports in Ad Week:

“We’re at the nexus of content, technology and distribution.” A content strategy needs to fit into a company’s business strategy. All tie into a basic yet vital function: the relationship with customer. Marketers say you have to measure what you can measure, and trust the rest. “We never ask a member or anyone to post on our behalf. “This is about their experience, what they feel comfortable doing. We will see that they tag us or talk about us. And this is a content lever, with zero media spend." Media companies structure what content performs around topics that have audience behavior underneath them. Why wouldn’t brands insert themselves into that?”

The notion that algorithms, or data, or technology is somehow objective and unbiased, neutral in all it surveys, has become a matter of faith for those disillusioned with their human colleagues.

But the reality is that algorithms and the data from which they are derived, and the devices on which they are devised or reported, are infused with the education, training and belief systems of their creators. JL

Sidney Fussell reports in The Atlantic:

Algorithms interpret millions of data points, and the exact path from input to conclusion can be difficult to make plain. But the effects are clear. This is a powerful asymmetry: Anyone can notice a change in search results, but it’s difficult to prove what caused it. That gives algorithm designers deniability. Because of their opacity, algorithms can privilege or discriminate without their creators designing them to do so. Algorithms provide “strategic ignorance." Try as companies might to minimize personal accountability, it is humans who build, train, and deploy algorithms. Human subjectivities are encoded every step of the way.

First Increase In Human Lifespan Via Biological Aging Intervention Reported

Exciting, but be careful what you wish for. JL

Shelly Fan reports in Singularity Hub:

Measuring a person’s “true” age is surprisingly difficult due to genetics and lifestyle. Compared to chronological age, biological age better correlates with general health status, mental abilities, age-related diseases, and death. Because aging gradually deteriorates the entire body, scientists have struggled to find the best markers. A combination of markers may form the best “clock” that measures true age. But one stood out: epigenetic alterations. A three-drug combination can induce measurable anti-aging effects in humans. “This is the first report of an increase in human lifespan by means of an aging intervention.”

Because of increasingly ease with which they can be crafted and the speed with which they can be disseminated, deep fakes may overpower purely technological solutions.

A combination of social, legal and technical fixes will have to be employed to address the threat. JL

Zoe Schiffer reports in The Verge:

Deepfakes are unlikely to be fixed by technology alone. (And) deepfakes can’t be solved just through the courts. Almost all solutions fight manipulation at the point-of-capture or at the detection level. Allowing people to manipulate videos and images using machine learning, with results that are almost impossible to detect with the human eye can go viral on social media in a matter of seconds. AI could actually make things worse by concentrating more data and power in the hands of private corporations. “Designing new technical models creates openings for companies to capture all sorts of images and create a repository of online life.”

Xrays, cat scans, cardiograms, your baby's ultra sound, all available on the internet. And not because of a nefarious plot, but because the growing consolidation and transferability of medical information has made it less secure.

But what happens if employers, banks, insurance companies - or a hacker - could get them - and use them to make judgements or extract financial concessions from patients? JL

Jack Gillum and colleagues report in ProPublica:

Medical images and health data belonging to millions of Americans, including X-rays, MRIs and CT scans, are sitting unprotected on the internet and available to anyone with basic computer expertise.The records cover 5 million patients in the U.S. and millions more around the world. A snoop could use free software programs or just a typical web browser to view the images and data. As networks of medical centers became more complex and connected to the internet, the responsibility for security shifted to network administrators who assumed safeguards were in place. “Suddenly, medical security has become a do-it-yourself project,”

Yes, it matters a lot. Because human sight may be far superior. JL

Kevin Hartnett reports in Quanta:

The neural networks underlying computer vision receive an image as input and process it through a series of steps. They first detect pixels, then edges and contours, then whole objects, before eventually producing a final guess about what they’re looking at.There is a lot we don’t know about human vision, but we know it doesn’t work like that. The visual cortex feedback process is very different from the feed-forward methods that enable computer vision. “Learning in deep learning methods is as unrelated to human learning as can be. “The wall is coming. You’ll reach a point where these systems can no longer move forward in terms of development.”

The more connected and accessible, the more vulnerable. JL

James Rundle reports in the Wall Street Journal:

Cities are using sensors to collect data about public utilities, traffic, garbage collecting, road conditions and more, then using that data to deliver services to more people more efficiently. But the more connected a city is, the more vulnerable it is to cyberattacks. As cities add connectivity to their streetlights, power grids, dams, transit lines and other services, they are adding more targets to be hacked. Sensors are the building blocks of smart-city initiatives.The amount of data sensors collect is enormous. Smart meters, smart garbage bins and traffic control systems also mean points of entry. The potential for harm is pretty much endless.

By lowering prices and offering inclusive upgrades on all its hardware, the company is effectively doing just that - and making good on its strategic shift to a service rather than product oriented company. JL

Ben Thompson reports in Stratechery:

Over the last three years the company’s “Wearables, Home and Accessories,” dominated by the Apple Watch and AirPods, has doubled from $11.8 billion to $22.2 billion. Over the same span Services revenue has increased from $23.1 billion to $43.8 billion. While Apple doesn’t need to worry about iPhone customers outside of China switching to Android, they are competing with the iPhones people already have, and, their own new, cheaper phones. How long until there is an all-up Apple subscription? Pay one monthly fee, and get everything Apple has to offer. Nothing would show that Apple is a Services company more than making the iPhone itself a service

It is worth acknowledging that those polled might change their minds if such break-ups seriously impinged on convenience and pricing.

But assuming people have become more sophisticated about the consequences of tech usage, two-thirds is a very big number, especially across all demographics and ideologies. JL

Emily Stewart reports in Vox:

Americans are on board with breaking up Big Tech, especially if it means companies such as Amazon and Google stop showing them search results they make money off of first. Two-thirds of Americans would support breaking up tech firms by undoing mergers, such as Facebook’s acquisition of Instagram, if it means ensuring more competition in the future. Seven in 10 Americans say it’s a good idea to break up big tech companies when the content they’re showing people is ranked depending on whether the company is making money off of it or not. The results hold across age groups, education levels, demographics, and political ideologies.

How Artificial Intelligence Is Changing Every Aspect Of War

Not just outshooting, outflying and outlasting humans but out-strategizing them - and in real time. JL

The Economist reports:

AI is “poised to change the character of the future battlefield. " The Pentagon’s budget for 2020 has lavished $1bn on AI and four times as much on autonomous capabilities. The Pentagon established the “Algorithmic Warfare Team," which uses deep learning and other techniques to identify objects and suspicious actions. The aim is to produce “actionable” intelligence. It will revolutionise command and control, a “massive near-real-time simulation”.(to) allow a “single synthetic command tool from the national security council down to the tactical commander." (But) “What do we do when AI applied to military strategy has calculated the probabilistic inferences of multiple interactions many moves beyond that which we can consider,”

If weather affects your mood and your mood affects your shopping patterns...JL

Patrick Kulp reports in Ad Week:

As closing the loop between offline shopping behavior and online advertising has become a holy grail of sorts for digital advertisers, some retailers have zeroed in on weather-based marketing and consumer behavior as a key component. The extra level of granularity provided by private forecasts can come in handy for a brand to steer marketing messages.The private weather industry could grow fivefold as “attractive new product offerings” allow businesses to wring more cost savings and revenue growth out of weather information.

As consumers have come to expect more service directed specifically at their immediate needs and convenience, horizontals eventually break. JL

Jeff Jordan and D'Arcy Coolican report in Andreessen Horowitz:

In all but a few circumstances, the broad horizontals eventually break. They become a victim of their own success. As the platforms grow, their submarkets grow too; their product gets pulled in different directions. Users get annoyed with an experience that caters to the lowest common denominator. What was previously too small a market to care about is a very interesting place for a standalone; a new wave of innovation begins to swell, picking off the compelling verticals the new horizontal players cannot satisfy. Looking at verticals within the horizontal platforms near their breaking point will point to the next great unbundling

With the major economies opposed, skeptical - or contemplating their own (China), Facebook's cryptocurrency initiative is confronting the company's terrible reputation as an ethical enterprise is a hindrance to gaining acceptance to anything it proposes.

The other issue is that this comes just at the time when the the tech backlash is gaining global power. JL

Raymond Collitt and Alastair March report in Bloomberg via Tech Central:

“A core element of state sovereignty is the issuance of a currency; we will not allow private companies to do it.” Euro-area governments see stablecoins like libra as a big threat to the stability of the euro because the economically weaker countries in the euro area might have some citizens who believe it is better to have their money in a stablecoin. Among regulators’ questions is that the new digital currency will be used by smugglers, drug dealers and terrorists. Another is that the social media giant, which has run afoul of regulators over user data in the past, should not be trusted to handle sensitive financial information.

In a saturated market, with technological differences a decreasingly important differentiator, Apple has learned the hard way that its brand umbrella can only provide so much additional advantage, meaning that, at a certain level, price is an important factor, even for them. JL

Jeremy Horwitz reports in Venture Beat:

Pricing tweaks are helping Apple reverse the sales slide that began a year ago, as the iPhone XR debuted at $749. The iPhone 11 instead starts at $699, getting closer to the price points that were popular with buyers of the non-Plus-sized iPhone 6, 6s, 7, and 8. Early sales are stronger than expected — a good sign as the company fights to maintain market share against increasingly powerful Android rivals.

The evidence suggests that there is a backlash. While tech industry avatars point to growing use of tech devices as 'proof' there is no resistance, the more convincing argument appears to be that the backlash is derived precisely from that very saturation: as consumers become more familiar with and adept at using technology, their sophistication about its pluses and minuses has grown, leading to demand for restrictions on what big tech is permitted to do.

In democracies, governments tend to bend to popular demand, not lead it. That so many are now challenging tech is further evidence of growing insistence on change. JL

Rob Walker comments in the New York Times and Casey Newton comments in The Verge:

95% of consumers in the United States say they have or use a cellphone, and 89% have or use the internet. “New connected devices continue to emerge” and we continue to embrace them. Voice assistants, smart TVs and wearable devices are growing in popularity. (But) four years ago, technology companies were seen as having a positive impact on the United States. The share of Americans who hold this view has tumbled 21% points since then, from 71% to 50%. Negative views of technology companies have nearly doubled from 17% to 33%. (As) for investigations into big tech, there (are) 2 Congressional, 6 state and local, and 8 federal investigations now underway.

Resistance to change may be futile, but it is usually inevitable. Successful transformations are not about new technology per se, but about managing the organization - its abilities, hopes and fears - in order to optimize it.

The crucial leadership variable is grounded in common values because those are what has gotten the enterprise where it is - and will endure long after the technology has changed again. JL

Greg Satell reports in Digital Tonto, Image by Pixabay:

Only 26% of organizational transformations succeed. Any transformation strategy that doesn’t take into account those who oppose change is unlikely to succeed. (But) when resistance is anticipated, transformational efforts can achieve astounding results. What makes the difference is that those leading the transformation expect resistance and built a plan to overcome it. Change is hard and can’t be easily managed. Root efforts not in specific goals, but in common values. “Transformation is about values first and technology second."

America's 50 Most Promising Artificial Intelligence Companies

The list of 50 companies below includes company name, strategic purpose, founders, funding sources and valuation or amount raised.

While just one such compilation, it is indicative of where both research and funding are trending. JL

Jilian D'Onfro reports in Forbes:

To be included on the list, companies needed machine learning, natural language processing, or computer vision to be a core part of their business model and future success. The honorees span human resources, security, insurance, and finance, with healthcare, transportation, and infrastructure best represented on the list. While most of the 50 hail from Silicon Valley, New York City and Boston, there’s representation from Detroit and Austin, too. The startups are flush with cash; startups touting AI received $7.4 billion in funding in the second quarter of 2019. The winners below are listed in order of ascending valuation.

The battle over the efficacy of risk models is really about who should bear the risk - and the cost - of additional insurance required as wild fires, hurricanes and tornadoes appear to be increasing in number and intensity. JL

Nicole Friedman reports in the Wall Street Journal:

Finding home insurance in California is getting more expensive (due to) the use of algorithmic models for predicting catastrophe losses. Insurers have long used models to project losses from natural disasters (but) are pushing for broader permission to use catastrophe models when setting rates. Consumers say models fail to capture wildfire risk. “Insurers are acting as you would expect where risk is going up. They’re adjusting their pricing and deciding in some cases that the risk is too high. We’re marching towards a future where wildfire risk is uninsurable.”

Jonathan English reports in CityLab:

From ancient Rome to modern Atlanta, the shape of cities has been defined by technologies that allow commuters to get to work in about 30 minutes. This means the physical size of cities is a function of the speed of the transportation technologies available. And, as speed increases, cities can occupy more land, bringing down the price of land, and of housing, in new territory. The average one-way commute time in American metropolitan areas today is 26 minutes.

The practice is technically legal in most states. For now. The question is whether the greater scrutiny to which tech companies are being subjected regarding personal information will extend to government agencies. JL

Joseph Cox reports in Vice:

Departments of Motor Vehicles in states around the country are taking drivers' personal information and selling it to thousands of businesses, including private investigators that advertise they will surveil spouses to see if they're cheating. Some of the data access is done in bulk, while other arrangements allow a company to lookup specific individuals. The data sold typically includes a citizen's name and address, ZIP code, date of birth, phone number, and email address. "With minimum standards, convicted felons can and do access professional databases."

That's a big increase in the last year, evidence of the company's push into advertising which it may see as a bigger future revenue source than people actually buying what they want. JL

Rani Molla reports in Re/code:

It’s never been more difficult to sell stuff on Amazon. 11% of all product views on Amazon come from sponsored listings, up 3% points in just the past year. Companies that sell on Amazon are increasingly having to pay to show up in search results — even when people are searching for specific brands. Brands not only have to buy ads to place prominently in search results, but they also must compete against Amazon’s own brands — and Amazon has extensive data on what and at what price people purchase goods online.

The results would be comical if it werent already apparent that companies and governments want to use these systems now to make with far ranging financial, political and medical judgements about people. JL

Nick Statt reports in The Verge:

Ask your standard recognition bot to analyze and label a photograph using only its acquired knowledge, and you’ll get some nonsensical results. A training data set AI researchers have relied on for the last decade is generally bad at recognizing people. It’s mostly an object recognition set, but it has a category for “People” that contains thousands of subcategories, each trying to help software classify a human being. And guess what? It's bad at it. Within computer vision and AI systems, forms of measurement easily but surreptitiously turn into moral judgments.

A growing body of data suggests that gig work is a transitional or supplemental arrangement until something higher paid and more permanent comes along.

And just as importantly, as human and intellectual capital becomes more important to corporate success, companies also find that a more stable, engaged work force is better for optimizing their own prospects. JL

Neil Irwin reports in the New York Times:

Business has been on a multidecade campaign to shift more economic risk onto its work force. The gig economy was the apotheosis of that shift. (But) the gig economy is looking less like the future of labor, and more like a niche arrangement, applicable in a handful of industries and used primarily as a side hustle for people whose main earnings come from a more stable type of job. It has major drawbacks for those who want a predictable income or for employers who need a reliable work force. "People did more gig work until they got a more permanent job"

A House Built - Mostly - By Robots

But without the jobs it displaces, who will be able to live there? JL

Anne Quito reports in Quartz:

Erecting a new building ranks among the most inefficient, polluting activities humans undertake. The construction sector is responsible for nearly 40% of the world’s total energy consumption and CO2 emissions.partnering with robots means letting the result of machine processes inform the design. Instead of forcing machines to fake handmade surfaces, there’s a totally new aesthetic that results from working with digital fabrication. “How you build matters.”

More important decisions have to made with less information in less time. JL

Peter Moore reports in his blog:

In too many organizations, decision making authority is determined by the hierarchical management structure that supports a silo-based business unit operating model not by the type of decision that needs to be made. This approach favors short term operating performance decisions over long term business growth decisions. The speed and complexity of change from disruptive technologies puts an even greater pressure on the need to change the decision-making cadence of your company. The risks of taking too long to make a decision in many cases outweigh the risk of making the wrong decision

Remote work is becoming a more significant factor in driving behavioral economics. People who can work anywhere are optimizing their variables by making choices that reflect a broader set of considerations. JL

Ben Eisen reports in the Wall Street Journal:

People who do their jobs from home, freelance or travel for work are migrating away from expensive centers such as Los Angeles and San Francisco toward cheaper cities including Boise; Denver; Austin, Texas; and Portland, Ore.The role of remote work has become increasingly important. The biggest U.S. cities are still drawing more people than they are losing, but the rate is slowing.“The livability crises of the West Coast and the East Coast are a push factor.”Workers spread out geographically during an economic cycle’s later stages. "We attract a lot of people who have the ability to work anywhere,”

The Emerging Confluence of AI and Biotech

To optimize their impact requires more imaginative thinking about their combined potential. JL

Tej Kohli reports in Project Syndicate:

Biotechnology, in cost-benefit terms, has been improving by a factor of ten every year. The cost of deciphering the human genome has dropped from $3 billion in 2001 to $1,000 today; a process that took months ten years ago can now be completed in less than an hour. AI’s contribution to global output will reach $15.7 trillion by 2030. (But) the siloed nature of current analyses means that potential AI/biotech combinations have not been fully priced in.The biggest hurdle preventing AI from reaching full potential is biological. Synthetic biotechnology could expand the scope of feasibility. We have underestimated their potential by considering them in isolation.

Wait. Data is fallible? Who knew...JL

Glyn Moody reports in Tech Dirt:

"We should remember: data is created to help deliver telecom services, not to control citizens or for surveillance." It could be valuable to police, but its primary purpose was to facilitate communication between users. If the authorities wish to use this kind of data they need to take into account that it was never designed to track people, and therefore has limitations as evidence. "This has changed our mindset about cellphone data. We are going to question it as we question a witness or other types of evidence, where we consider circumstances like who produced the evidence, and why and how."

Because luxury car owners dont drive long distances, concerns about charging are not as much of an issue. And, if the trends in places of immense wealth like Silicon Valley are any indication, they like the statement it makes about how fashion forward they are. JL

Tensie Whelan reports in Worth:

The downside for zero emission vehicles is the limited driving range before charging and the time it takes to charge the car. Because luxury cars are typically not driven for long distances (an average Aston Martin owner drives his or her car under 10,000 miles a year), because the electric motor enables much more internal space without growing the exterior, because autonomous technology will enable people to enjoy their cars as an extension of their homes, and because an electric motor can deliver immense power, I think the future of luxury is electric.And, not incidentally, an EV allows the owner to make a statement.

Will the ceaseless copying of other people's images dull humanity's ability to innovate and create? JL

Alexander Lange comments in The New York Times:

“On social media, Antelope Canyon has become more than a beautiful image of rocks in the desert — it’s a status symbol, a luxurious backdrop over which to perform one’s adventurousness and wanderlust.” Tour guides tell visitors which filters to use to make their photo look exactly like all the others. With Instagram, you can contribute more visual delight to the digital world, or you can collect digital souvenirs picked out by other people.The result is a cycle that can be either vicicious, when Instagram produces copycats and environments are ravaged by selfie-taking; virtuous, when Instagram provokes an interest in new places

AI systems' capabilities do not yet possess sufficient data and are not sophisticated, intelligent or robust enough to function without human guidance. And may never be able to do so. JL

Joe McKendrick reports in Forbes:

ML systems face challenges with respect to robustness and explicability. “The data used to train ML systems must be as trusted as the systems themselves need to be. Explainability is a challenge, as ML systems “tend to be black boxes that offer no insight into how they make their decisions (and) is essential for systems that must be robust to failure, interact with humans, and aid in significant decisions.”The notion of 100-percent automation with no human input may make sense for situations where the product or process is mature and stable. But even the most automated plants require a large number of workers to set up, maintain, and repair equipment.

Stagnant household income and wealth creation among the bulk of the population means less money is circulating which reduces economic stimulation and growth. JL

Justin Lahart reports in the Wall Street Journal:

The median U.S. household had income of $63,200 last year. That was a bit more than in 2017.But the 2018 figure was even with 1999 considering that the U.S. economy grew by an inflation-adjusted 48% over the same period. It reflects shifts in the economy by which the top 20% accounted for 52% of household income last year. Inequality may change the way the economy works. Rich people are less likely to spend additional income, leading to reduced growth. And because the rich tend to invest that money, it can lead to increased asset prices. Inequality might have created the bind in which easy-money policies failed to ignite growth.

Rather than providing secure employment for the future, STEM skills are precisely those being targeted for replacement by robotics and algorithms. JL

Caitlin Zaloom reports in The Atlantic:

College, in this STEM view, amounts to higher-level vocational education for young people, anointing them the yeoman workers of the corporate economy. That a liberal-arts education prevents students from getting jobs is spurious. Graduates with a broad-based education are in demand. Employers are seeking skills that come from a more exploratory education. Workers’ capacity to communicate and work with others are in short supply. And the income benefits of STEM jobs are in decline. Tech companies are hard at work to oust white-collar workers by automating their jobs. When universities “become engines of lifelong learning,” they also “robot proof” education.

Sep 21, 2019

Sep 20, 2019

Sep 19, 2019

Sep 18, 2019

Sep 17, 2019

Sep 16, 2019

Sep 15, 2019

contact

Search This Blog

Blog Archive

Labels

links