View From the Valley iconView From the Valley

Illustration: iStockphoto

New Data Hammers Home Problems With H-1Bs and Outsourcing Firms

The biggest critique of the United States’ H-1B visa program is that tech companies use it as a channel for importing cheap labor from overseas. New data released by the federal government affirms that argument. It shows that most of the visas go to a handful of IT outsourcing companies, which pay their guest workers well below average.

The reports on H-1B employers and salaries by the U.S. Citizenship and Immigration Services show that 69 percent of the 350,000-plus visas approved in 2016 were for computer-related occupations, and 74 percent went to people born in India.

The top 20 employers took 37 percent of the approved visas. The top five were all IT outsourcing firms: Cognizant Tech Solutions, Infosys, Tata Consultancy Services, Accenture, and Wipro. All together, these companies took 60,000 visas.

The average salary of H-1B visa holders was $91,000. But top outsourcing firms paid well below this number, with TCS paying as little as $72,000 on average.

Non-outsourcing tech companies seem to be more fair. Apple, Cisco, Microsoft, and Google, for instance, pay their H-1B workers average wages of over $120,000. These higher-paying companies generally hire more workers with Master’s degrees, while most outsourcing firms hire mainly Bachelor’s degree holders.

But even major tech firms are using the program to pay less than median wages in their area, points out Ron Hira, professor of political science at Howard University. Microsoft, for instance, pays H-1B software developers $126,000 on average, while the Bureau of Labor Statistics puts the average salary for that job at $132,000. “Any firm that is able to pay less money is going to pay less money,” Hira says.

Read More
A simulation of Zizmos' earthquake early warning system shows the progression of a temblor in the San Francisco Bay Area

Zizmos Continues Its Quest to Create an IoT Earthquake-Warning Network

I first met Battalgazi Yildirim two years ago. He had posted a request in my local online community: His startup, Zizmos, wanted volunteers willing to mount a sensor package inside their homes, preferably on a bearing wall, to test whether a network of cheap packages of electronics, based on the Android phone design and his algorithms, could give early warnings of earthquakes. He wasn’t looking to do long-term prediction, just 15 or 30 seconds—enough to allow people to grab their kids and move to the safest spot in their house.

Yildirim funded that first trial—which eventually involved 100 sensors—with an NSF research grant of $150,000. Like many first design attempts, it didn’t work out so well. It turned out, Yildirim told me last week, that the Android platform had a fatal flaw—it couldn’t pull in data from an external sensor and simultaneously recharge. The alpha testers might have been willing to deal with keeping the gadgets charged, but this approach wasn’t going to appeal to the masses. And Yildirim’s idea is going to need mass adoption to work; it relies on large numbers of low-cost sensors that report possible earthquake vibrations to the cloud, then eliminates false alarms by comparing the data with neighboring sensors.

The good news, however, Yildirim said, was that the internal sensors on phones were getting better and better—maybe, he thought, he and cofounder Greg Stillman could just design an app instead of dedicated hardware. He entered the Verizon Powerful Answers competition and his proposal won a grand prize--$1 million. The award also came with a lot of help, formal and informal, he says, from Verizon’s business team.

With enough money to carry Zizmos for a while, Yildirim hired contractors to work on apps for Android and iPhone and also went back to his original idea of producing a wall-mounted Internet of Things gadget. “The system works better,” he says, “if we gather data from a fixed device for which we have full control of the hardware and software.”

The free app is out and has had nearly 100,000 downloads so far. It works as an earthquake sensor only when it is plugged into power, is connected to WiFi, and is stationary, say, sitting on a bedside table.

“It’s mostly looking at frequency,” Yildirim says. “When an earthquake originates some distance away, the high frequencies quickly die off; instead you are just getting low frequency vibrations. If a truck passes by or you bump the table the phone is sitting on, you get a broader range of frequencies.”

If the app triggers on a possible earthquake, it sends its data to the cloud. The cloud software knows the locations of nearby sensors. If enough are in the same neighborhood, it analyzes data from all of them. “If it is an earthquake wave, multiple sensors will trigger in sequence,” Yildirim said.

The system is currently using the data for research only. It isn’t sending out actual warnings to users. “We are detecting earthquakes,” Yildirim said, particularly on a small network in Northern California. But the users worldwide are currently distributed too sparsely to be of much use.

In the meantime, the app has some utility for those who have downloaded it: It gives earthquake reports based on data from government agencies around the world and allows users to customize those alerts for specific geographic areas, both local, and where friends and family live.

Next out of the gate, Yildirim says, will be that IoT package. He’s got 25 prototypes installed and tested, and Zizmos hardware engineering lead Mark Johnston is working on the industrial design for the final product, which Yildirim says will be small and sleek enough that it wouldn’t look out of place sitting next to a Nest thermostat on the wall. Zizmos plans on launching a Kickstarter campaign in a month or two in hopes of generating enough preorders to guarantee that particular geographies, like California or Oklahoma, will have critical mass—before shipping. He estimates that, for California, that will mean at least 6,000 sensors installed around the state.

Meanwhile, more prize money funding is trickling in—Zizmos scored additional prize money last month (but has to keep quiet until a formal announcement is made). Yildirim is hopeful that some additional grant applications will bring in funding, and says his idea has gotten some traction from reality TV producers. The company was scheduled to be featured on Planet of the Apps but the episode was preempted and Yildirim has heard no rescheduling news. He was also a finalist on America’s Greatest Maker, with $1 million at stake, but the show has been cancelled.

“Competitions like this product,” he says, “and they don’t require giving up equity. But you can’t run a company forever on prize money.”

The long-term plan is turning Zizmos into a venture-funded company that turns a profit—not from purchasers of the hardware, but from structural engineers and insurance companies interested in how certain types of buildings respond to earthquakes in different areas. They would find value in the data, using it to provide more granular risk assessments. 

A photo of a finger pressing an image of the world with interconnections between images of people.

The Big Sprawl: Software Engineering Jobs Escape Into New Territories

Once upon a time, software engineers worked in tight tech bubbles, concentrated either by geography or by industry. But today, “every company is trying to transform itself into a tech company,” says Andrew Chamberlain, chief economist with job search site Glassdoor. And that, he says, has led software jobs to emerge from the tech bubbles and sprawl, well, just about everywhere.

Read More
The giant B is Blind's logo; figures gather to represent tech companies

Tech Workplace Gossip App Blind Opens to the Masses

Want to know what it’s really like at a tech company? You can check out reviews on Glassdoor, but those tend to be very limited in scope, along the lines of “hard to advance,” “too much office politics,” “good benefits.” 

So people ask Quora, “What’s the company culture like at LinkedIn” or “What’s it like to work at Google Research?” They get a few qualitative responses, usually from former employees, but sometimes from people reporting just on what friends of friends have said. These answers tend to avoid specifics like salaries, and couch any criticisms in politeness. And people currently at the companies in discussion generally keep their opinions to themselves.

But there is a place where current employees are saying more—a lot more. That’s a discussion app called Blind. Launched in the U.S. 2015 following a successful start in South Korea, it offers channels for individual companies, verifies that people actually work at the company via email address and LinkedIn, and then masks their identities. It also sometimes creates mashups: the company opened a joint Microsoft/LinkedIn channel after the LinkedIn acquisition news.

Read More
Stanford students experience Navy SEAL training as part of research for the university's Hacking for Defense class

U.S. Defense Budget May Help Fund "Hacking for Defense" Classes at Universities

In 2016, Stanford students started hacking for defense—that is, they took on real projects from National Security Agency, the Army, the Navy, the Air Force, the Army Cyber Command, the Veterans Administration, and other agencies with defense-related problems. The students actually came up with prototype solutions.

The innovative Hacking For Defense (H4D) class, which requires each student team to conduct at least 100 interviews with defense industry “clients,” caught on quickly. Today, according to Steve Blank, an instructor at Stanford and one of the creators of the curriculum, eight universities in addition to Stanford have offered or will offer a Hacking for Defense class this year: Boise State, Columbia, Georgetown, James Madison, the University of California at San Diego, the University of Pittsburgh, the University of Southern California, and the University of Southern Mississippi. The class has spun out Hacking for Diplomacy, Hacking for Energy, and other targeted classes that use the same methodology.

The snowballing effort is now poised to get a big push. This month, the U.S. House of Representatives passed an amendment originated by Rep. Dan Lipinski (D-Ill.) to support development of curriculum, best practices, and recruitment materials for the program to the tune of $15 million (a drop in the $700 billion defense budget but a big deal for a university program).

In arguing for the amendment, Lipinski said, “Rapid, low-cost technological innovation is what makes Silicon Valley revolutionary, but the DOD hasn’t historically had the mechanisms in place to harness this American advantage. Hacking for Defense creates ways for talented scientists and engineers to work alongside veterans, military leaders, and business mentors to innovate solutions that make America safer.”

Even though the first class was just a year ago, ideas coming out of it are moving quickly into the real world. One of the teams in that first group, Capella Space, is getting ready to deploy a fleet of tiny, cheap, synthetic aperture radar satellites that may prove instrumental in tracking missiles launched by North Korea. (The team had angel investors before its members got their grades for the quarter.)

Hacking for Defense efforts, like Capella’s, have “the potential to make our world a much safer place,” wrote Blank in a blog post.

The Defense Authorization Act is now awaiting action by the U.S. Senate.

A study by the American Immigration Council shows dramatic increase in foreign-born percentage of STEM workers

42% of California’s STEM Workforce Hails From Outside the U.S.

The proportion of U.S. STEM jobs held by people born outside of the country has grown dramatically since 1990, and California and New Jersey are leading the trend. So says a study by the American Immigration Council. The organization looked at data from 2015 and earlier, excluding STEM careers in higher education.

The council calls its STEM definition narrow; it includes computer and mathematics, engineering and surveying, physical and life sciences, and managerial careers, and it excludes health care and social sciences. Here are a few key takeaways:

  • STEM workers as a share of the total U.S. workforce increased from 3.4 percent in 1990 to 5 percent in 2015
  • The foreign-born share of the total U.S. STEM workforce increased from 11.9 percent in 1990 to 24.3 percent in 2015.
  • The computer and math fields showed the highest increase in the proportion of foreign born workers, from 11.9 percent in 1990 to 26.1 percent in 2015.
  • U.S. software engineering fields, in particular, have a high representation of foreign born workers: 39.2 percent of the workforce. Foreign-born electrical engineers make up 28.4 percent of that workforce.
  • Among the states, California and New Jersey, at 42.4 percent and 43.8 percent respectively, have the largest foreign-born representation in their STEM workforces. Wyoming has the lowest at 4 percent.

The report also looked at other, independent studies of STEM workers, and noted that data shows that 25 percent of high tech companies founded between 1995 and 2005 have at least one immigrant founder and that foreign-born people living in the United States are more likely than native-born people to obtain a patent.

The full report is available here.

This robot positions mobile devices against a dummy designed to absorb radiation at the same rate as a human; government regulations set limits for safe absorption rates

Testing the Internet of Things

That tsunami of new IoT gadgets? They all have to be tested before they roll out into the world, not only to meet government regulations but to verify adherence to a host of voluntary standards, like WiFi, Bluetooth, ZigBee, Thread and others. That is a lot of testing. And that’s why TUV Rheinland recently opened a huge Silicon Valley test facility in Fremont, Calif.

Read More
Singapore entrepreneurs in security and surveillance say being able to avoid NSA requirements give their companies a big advantage

Doing a Startup Involving Cryptography? Get Out of the United States

“There’s no better place than Singapore to do a deep tech startup, particularly anything involving cryptography.” So says Brijesh Pande, founder and managing partner of the Tembusu ICT Fund, a Singapore-based software-focused venture capital fund. Admittedly, he has a vested interest in enticing entrepreneurs to come to the island nation, but he—and two founders of companies in his portfolio, Lawrence Hughes from Sixscape and Raymond Looi from Vi Dimensions—make a solid argument.

Here in Singapore, Pande says, “We have no requirement for a security back door. The fact that the NSA [National Security Agency] requires U.S. companies to provide a back door makes technology developed in the U.S. less trusted around the world.”

Hughes, who before decamping to Singapore founded several companies, including U.S.-based Ciphertrust and Philippine-based Infoweapons, says it’s just too difficult to do cyber security products in the U.S. these days. “The NSA requires weakened algorithms and back doors, so you have to assume all IT products in the U.S. are compromised.” That, he says, makes it hard to market them around the world.

Sixscape, Hughes’ latest startup, has developed a certificate management protocol based on a distributed public key infrastructure that can manage, he says, billions of unique certificates. Web sites use server certificates to identify themselves as legitimate. In his scheme, individual users will also use client certificates, created on their own computers, as identification, instead of less secure usernames and passwords.  For additional security, banks and other particularly sensitive businesses can give their clients hardware “keys” containing certificates to even more reliably confirm identity. Hughes says Sixscape will soon be piloting the technology, issuing some 2 million hardware tokens for a government agency in a nearby country that wishes to allow secure access its website, without usernames or passwords.

Meanwhile, two-year-old Vi Dimensions, Looi explained, is developing AI to mine surveillance videos for anomalies. “With hundreds of millions of cameras out there,” he says, “the cost of human monitoring is just too high.”  The software, he says, can spot a big truck driving an unusual route, or a child lost in a subway station. The company, he says, has deployed the technology on 200 cameras in Sentosa, a resort island with 20 million visitors annually, has signed an agreement with one of Europe’s largest national railways to use the technology on its surveillance cameras, and has completed trials in an Abu Dhabi skyscraper. On Sentosa, he said, the technology was able to improve operations by spotting too-long lines at taxi stands and identifying points where parents were having trouble navigating strollers.

It’s not just escaping the NSA that makes Singapore more and more attractive to startups, these entrepreneurs say.

The potential tightening of H1B visas in the United States will push more companies to start elsewhere, says Pande. “If the U.S. develops an H1B issue,” he says, “that will be good for Singapore.”

“Here, there is no limit to H1Bs, you just have to demonstrate a need,” Pande says. “We have Singaporeans, Americans, Iranians, Indians, Russians—a veritable United Nations working at all our companies.”

“And,” pointed out Hughes, “Singapore has zero capital gains tax,” relying instead on income and value added taxes.

Starting a company in Singapore, Pande points out, is not without challenges. “We don’t have a deep bench,” he says. “So the second level of tech talent, just below the entrepreneur, can be hard to come by.”

And though the local universities are solid, he says, there just aren’t the big tech companies doing core R&D that spit out spinoffs. “And there isn’t a big domestic market,” says Pande, though government support helps.

Valuations for startups also tend to be low compared to those in the U.S., Hughes says, “but it’s early days here.”

The fictional Richard Hendricks explains an Internet of mobile devices in HBO's "Silicon Valley"

A $2 Million Contest Seeks a Real-World Pied Piper to Solve Big Internet Challenges

In Sunday’s season finale of HBO’s Silicon Valley, fictional startup Pied Piper’s attempt to create a decentralized Internet appears to have spectacularly failed, thanks to mobile phone explosions and a disastrous attempt to move a server. Then the distraught founders discover that their network is actually ticking along just fine. But how? It turns out that it has jumped to smart refrigerators. Now that’s resilient!

The Internet of refrigerators is, of course, fiction. But could an Internet that is this resilient—or nearly so—be a reality? Mozilla and the U.S. National Science Foundation think it’s possible, and aim to accelerate its creation by offering $2 million in prize money to teams that invent it—or at least get close.

“We’ve picked two of the most challenging situations in which people are disconnected from the Internet,” Mozilla program manager Mehan Jayasuriya told me. These are, “Connecting people in the U.S. who don’t have reliable or affordable Internet and connecting people as quickly as possible after a major disaster, when the traditional networks go down.”

 Mozilla and the NSF are addressing that first group—an estimated 34 million people—with the “Smart-Community Networks Challenge” that seeks wireless technology designed to enhance Internet connectivity by building on top of existing infrastructure.

For the second group, there’s the “Off-The-Grid Internet Challenge.” That contest seeks technology that can be quickly deployed after a disaster to allow people to communicate when Internet access is gone.

The teams submit initial designs, and then later, working prototypes. Prizes at the design stage range from $10,000 to $60,000. At the working prototype stage, the stakes range from $50,000 to $400,000, with one of the top awards given for each challenge category.

Judging criteria for both challenges include affordability, feasibility, social impact, and scalability. Off-the-grid technology also has to be portable and have a portable power source. The smart community networks technology will also consider density, range, bandwidth, and security. Potential entrants must submit an intent to apply form by 15 October; the whole thing wraps up next August.

“A lot of projects out there address some parts of these problems,” Jayasuriya says. “With $2 million on the table, we are hoping this challenge encourages people to fill their technologies out.”

Were Pied Piper a real company, it would have a decent chance at winning some of that cash. Says Jayasuriya:

It’s the kind of thing we are looking for—a big idea, a crazy idea, an idea about how you piggyback on things that already exist. Pied Piper’s approach is like that, looking at all the phones out there and thinking that these phones have radios, and power, and CPUs, so why wouldn’t you take them and turn them into nodes on a network.

Jayasuriya is hopeful that the contest will lead to real change. The structure of the competition, he says, is designed to push design teams to build working prototypes, so the winner, with $400,000 in seed money, will be well positioned to attract partnerships or investment capital, and get the product out into the world.

Chinese flags wave near high-rise buildings

China’s Big Tech Advantage Is Speed, Not Cost

China’s not just a copycat economy anymore; it has turned into a high-speed innovator. So says Ben Joffe, general partner of HAX, a five-year-old hardware accelerator and investor.

And the rest of the world had better watch out.

We call what China is doing “Xiaomization,” because smart-phone giant Xiaomi’s high-speed commoditization of product categories is driving a lot of people out of business. Right now, he says, Xiaomi has invested in 77 companies—and four are unicorns, meaning they have a market capitalization of over $1 billion.

Xiaomization, he says, knocked Fitbit out of the Chinese market when Xiaomi launched a $13 activity tracker. And no product category is safe.

“The advantage of China is not cost,” Joffe says. “It is speed. Everything works at startup speed. You will get a message from the supplier Sunday night telling you that they can ship your component in morning.”

It’s also becoming a research mecca, Joffe said, with top universities and companies from around the world starting research centers in China. And, he says, this year China will overtake Japan in terms of numbers of patents filed, making it second only to the United States.

There’s plenty of startup funding in China as well, Joffe indicated. “Currently, there is an overflow of funding,” he said, “particularly angel funding.” First round funding, he pointed out, tends to be larger than funding for comparable ventures in the United States, because, he said, China believes you have to win a market quickly.

“China today,” says Joffee, “is like Japan in the 1980s: incredibly creative and incredibly fast.”

As for particular technologies, he indicated, voice control is booming—as is virtual reality, with thousands of VR arcades introducing people to the technology. Rental ventures are booming, with, he said, some companies introducing rental bikes that don’t need to be docked at a specific station, others are renting umbrellas, batteries, and even basketballs.


View From the Valley

IEEE Spectrum’s blog featuring the people, places, and passions of the world of technologists in Silicon Valley and its environs.
Contact us:

Tekla Perry
Palo Alto, Calif.
Load More