The Very Mathematical History of a Perfect Color Combination

A couple of years ago, I fell in love with a color scheme: off-white text accented with a buttery yellow-orange and a neutral blue against a deep gray, the “color of television, tuned to a dead channel,” to borrow a phrase from Neuromancer author William Gibson. The colors were part of a theme called “Solarized Dark” for the popular MacOS code editor TextMate. To be honest, I didn’t think much of Solarized at first. But I soon found that I couldn’t work with any other color scheme. Staring at screens all day can make you particular about fonts and colors.

It turns out I’m not alone. I’m not a coder by trade, but I like to use code editors for writing and organizing notes. While hunting for tools after switching from a Mac to Windows, I started to see Solarized Dark and its sibling Solarized Light, which uses the same 16-color palette, practically everywhere I looked. It’s hard to say how many programmers use it. The design is free and open source, so there’s no tally of purchases. It’s available for every major code editor, and many other programming tools. Microsoft even bundled it with its popular code editor VS Code. Solarized has a loyal following.

“If I bring up a terminal window that doesn’t have Solarized, I feel out of place, I don’t feel at home,” says Zachery Bir, a Richmond, Virginia, programmer and artist who has been using Solarized since shortly after it was released in 2011. Bir likes Solarized so much he uses it as the color scheme for his computer-generated art. “I didn’t trust myself to come up with a palette that was balanced and looked good both in a dark and light medium,” he says.

The Solarized color scheme is no accident. It reflects the obsessive attention to detail of its creator, Ethan Schoonover. “I didn’t release it until I was 1,000 percent sure I loved all the colors and they were all dialed in mathematically,” Schoonover says. “I had multiple monitors, some were color calibrated, others were deliberately messed up. Sometimes I showed my wife, who thought I was a little nuts.”

Too Much Contrast

Schoonover was working as a designer and programmer in Seattle when he started work on Solarized in 2010. He’d recently switched operating systems was disappointed in the color schemes available for the tools he used. Many applications offered only a simple white-on-black scheme that harkens back to old-school text-based computer terminals. But Schoonover found these throwback color schemes much harsher than the retro displays they tried to emulate. That’s because the backgrounds displayed on old 1980s monitors weren’t truly black, Schoonover says. “They had less contrast.” Today’s LCD’s, on the other hand, are capable of displaying much darker, and much brighter, colors.

The optimal amount of contrast for text on a screen is controversial; many people prefer high-contrast themes. But contrast wasn’t Schoonover’s only concern. He found most low-contrast color schemes lacking as well. Even the best-designed themes tended to use at least one color that appeared distractingly brighter than others. That’s because the apparent brightness of a color varies depending on its background. In other words, a specific shade of blue will appear more or less bright, depending on the surrounding colors.

This phenomenon, known as the Helmholtz–Kohlrausch effect, is particularly aggravating for programmers because coding tools use color to distinguish different parts of code. In the code for a web page in a typical text editor using the Solarized Dark theme, for example, web links appear in green; the syntax for formatting, such as adding italics, is blue, and comments that developers write for themselves are gray. Ideally, the colors should help tell these elements apart, but no single element should stand out more than others.

Schoonover set out to find a set of colors that would not only look good together, but have the same apparent brightness. That task was made more difficult because he wanted to use the same palette in both a light and a dark theme. Hence the need for all the monitors and testing.

Examples of the Solarized Dark (left) and Solarized Light (right) themes displaying HTML code in the code editor Vim.

Ethan Schoonover

Ethan Schoonover

Schoonover talks a lot about the mathematical nature of his color selections, but he picked the starting colors, a blue and a yellow, for very personal reasons. The blue reminds him of his long standing thalassophobia, the fear of very deep water. And though he says he doesn’t otherwise experience synesthesia—such as hearing colors or tasting words—the yellow invokes tastes and smells he associates with his childhood. “My parents are artists, I’m comfortable picking things for obscure reasons,” he says.

With those starting points, Schoonover sought out other colors that provided just enough—but not too much—contrast between elements, and that maintained the same level of contrast in the light and dark versions. The result is a palette of just 16 colors that retain the same relationships even when inverted. “I suppose it’s a little like composing music with only a limited number of notes,” Schoonover says. “There can be something sparse and beautiful about it.”

An Open Source Program Takes Off

Schoonover released Solarized for free in April 2011 on GitHub, a code hosting platform and collaboration service. He says he never intended to commercialize it. “It would kill something special about it, taint it,” he says. “I believe in open source software, I believe in giving something special to the world that anyone can use.”

Although he’d tested the color scheme in a variety of applications, Schoonover initially only released themes for a few tools he used in his own work, like the code editor Vim and the text-based email client Mutt. He announced the release of Solarized on the Vim mailing list; soon after, the project hit the front page of the online community Hacker News. It was an immediate hit with programmers, who soon went to work adapting it to other programming tools beyond those Schoonover initially supported. In 2013, Solarized Dark appeared on the monitors of developers in a Facebook commercial—watch for those dark rectangles on the screens and notice the faintly colored lines that cross them.

Solarized is slowly starting to find its way into applications for non-geeks. Ulysses, a writing application for MacOS, includes Solarized themes as an option. The color scheme was used for many of the graphics in the video game N++ in 2014. The note-taking app MicroPad even advertises Solarized as a feature on its website. “Solarized Dark for MicroPad is especially useful for late-night studying, which I do more often than I would like to admit,” says MicroPad creator Nick Webster, a computer science student at Victoria University in Wellington, New Zealand.

But it still hasn’t really crossed over to the mainstream as a color scheme for, say, a major web application or software suite. “When Apple introduced dark mode for MacOS, I thought it was cool,” says Bir, the Virginia programmer and artist. “But I wish it was Solarized.”

With more applications, like Google Chrome, Facebook Messenger, and Slack, releasing dark mode themes, though, Solarized just might have its day in the sun.


More Great WIRED Stories

AI cloudops is coming, whether you like it or not

As someone who’s worked with AI for the last 30 years (yes, it was a thing 30 years ago), I’ve often thought of its capabilities were overrated and used for the wrong things in many cases. Now that it’s cheap thanks to cloud computing, and much more effective thanks to the pace of innovation, AI as a solution is coming up again, including the use in cloud operations.

The idea is to replace people with AI to be both proactive and reactive to cloud operational issues such as outages, resource governance, security attacks, and performance. Cloudops involves largely repeatable problems, right?

There are of course some upsides and some downsides to this. Moreover, although the use of AI in cloud operations maybe a foregone conclusion, there will still be a learning curve that is required. As long as you understand that and know what to expect in terms of ROI for both the short term and long term, I’m okay with anything that that makes cloud operations more effective.

So, let’s look at the pros and cons.

The pros of AI for cloudops

The pros are that you can have a 7/24/365 monitoring and management program on the cheap. If you believe operational staff is expensive, try hiring them for shift work. AI-based monitoring and management systems never sleep, never take time off, and never ask for a raise. Once they are up and running, they cost almost nothing beyond their license fees and infrastructure costs. And they are self-learning at the same time; in other words, the more they run, the better that they get at the job.

Another pro is that these systems get smarter every day and share a common brain. People get smarter with experience as well, but they don’t do a good job sharing their experiences with others. People also retire and quit, with the knowledge and experience walking out the door with them.

The cons of AI for cloudops

One con is that the cost of rolling out these systems is high, even in the cloud. Vendors that have married AI and operational tools are going to charge a premium to get them up and running and in production. While the prices are all over the place, count on paying 50 percent more than for traditional tools, including consulting services for the first year or so to get the tools learning correctly.

Another con is that operations people don’t seem to like them no matter how well they perform. The number of passive-aggressive actions that I’ve seen over the years from people pushing back on AI-enabled operations tools has been huge.

They view this technology as not to be trusted, plus the fact that AI some day may displace their jobs does not make things better. Organizations that implement these tools need to have change agents, plus an understanding about the human factors with this technology.

Is the future AI-enabled cloud-operations tools? I don’t see how it won’t be. The pros will get better, and the cons will begin to diminish, like any other rollout of new technology. Hopefully, our new AI operations overlords will have mercy on us in a few years.

A Cab’s-Eye View of How Peloton’s Trucks ‘Talk’ to Each Other

Techno-optimist prognosticators will tell you that driverless trucks are just around the corner. They will also gently tell you—always gently—that yes, truck driving, a job that nearly 3.7 million Americans perform today, is perhaps on the brink of extinction. At the very least, on the brink of uncomfortable change.

A startup called Peloton Technology sees the future a bit differently. Based in Mountain View, California, the eight-year-old company has a plan to broadly commercialize a partially automated truck technology called platooning. It would still depend on drivers sitting in front of a steering wheel, but it would be more fuel efficient and, hopefully, safer than truck-based transportation today.

The company employs 10 professional truck drivers to help refine its tech, and I’m about to meet two of them out on Peloton’s test track in California’s Central Valley. Michael Perkins is tall, thin, and has been driving very big trucks for about 20 years. Jake Gregory is shorter and picked up truck driving in college, before taking a detour to the FBI.

We hit the highway first, because the rain has suddenly cleared. (Here’s an unfortunate reality about Peloton’s driver assistance tech: It doesn’t work great in the rain. Or snow. It’s a safety issue. More on that later.) Out on Interstate 5, Perkins’ long, white semitrailer cruises along in front of me. I’m on board the second, identical truck behind it, with Gregory behind the wheel. A small screen mounted on Gregory’s dashboard shows a camera view of what’s happening in front of Perkins’ rig. It’s like their trucks are connected. Which, in fact, they are about to be.

Peloton

Perkins radios in that he’s ready to go; Gregory says he is too. Inside the two truck cabs, each driver hits a button. Three ascending tones—la, la, la—means Peloton’s automated system has authorized the trucks to platoon on this stretch of highway. A dedicated short range communications (DSRC) connection is now established between the two vehicles. It’s like Wi-Fi but faster and easier to secure. Now, whatever the front truck does, the back truck will near-simultaneously “know”—and react accordingly.

Then Gregory speeds up, pulling his truck up so it’s tailgating about 70 feet from the leader. Sounds risky! But right now, the two trucks are platooning. Ours is on a kind of hopped-up cruise control, which means Gregory’s feet aren’t actually controlling the brakes or accelerator. At the same time, Gregory maintains control of his steering wheel. If Perkins were to brake hard, Gregory’s truck would too, faster than a human could. The robots have taken over. Kind of? Not really? More like, they’re collaborating, with some human oversight.

Peloton’s name, a reference to bicycle racing, helps explain how this platooning works. Just as the riders in the peloton, or main group of racing cyclists, preserve energy by drafting off of those around them, the following trucks in the truck platoon reduce their aerodynamic drag by drafting off the ones in front. The lead truck, meanwhile, get a little push. This saves fuel, according to Peloton—up to 10 percent for the following car and 4.5 percent for the first one, depending on the road and weather conditions and the following distance. It might also prevent crashes, since this tech has much faster reaction times (about 30 milliseconds) than puny humans (about 1 to 1.5 seconds).

Other companies in Europe, China, Japan, and Singapore are seriously experimenting with truck platooning. The American military has hosted platooning demonstrations. Just this week, the US Department of Transportation gave out $1.5 million in grants to universities studying the tech. And Peloton has tested in a bunch of US states: Arizona, California, Michigan, Florida, and Texas, where Peloton has immediate plans to run the majority of its routes.

Right now, the company says it does have paying customers, though it won’t reveal their names until later this year. According to Josh Switkes, the company’s CEO, some pair of US truck drivers are running a route while platooning on a Peloton-enabled truck every day.

And testing continues, on the software in its office, on its test track, and on actual highways, where it confirms the technology’s reliability. “The highway or field is not for testing,” Switkes says. “The goal of testing is to find failures, and you don’t want those failures to be on public roads.” In a report released today, the company lays out this approach to safety for regulators and interested industry parties alike. It borrows more from automotive processes than Silicon Valley–style software ones, amounting to something like easy does it.

It turns out, the linking-up move Perkins and Gregory just performed on the 5 is one of the most safety-critical parts of truck platooning, Switkes says. The moment when the following truck has to move faster than the one in front of it is the most dangerous part.

To make sure drivers like Perkins and Gregory don’t crash into each other, or anyone else, Peloton needs to make sure that the platooning drivers know how the tech works. (Right now, the company’s driver training process takes about half a day.) It also needs to understand exactly how heavy the trucks are when they start platooning, how their brakes are working, and how their tires function. For this reason, the company says, it has carved out partnerships with its suppliers, which means its trucks are built from the ground up with platooning in mind.

This is also why Peloton doesn’t platoon in the rain right now, or in the snow: The company can’t yet gauge exactly how tires deteriorate over time, which means it can’t quite predict how they’ll react in a hard-braking situation. Worn tires might slide in the moisture, leading to a domino chain of truck crashes. So no platooning in the Midwest in the winter, or anywhere during a rainy spring. “On certain routes, it’s a significant limitation,” says Switkes. “But we’re erring on the side of safety.”

And if that seems a little dull, Switkes would tell you that’s the point. His favorite word is “pragmatic,” and he doesn’t believe driverless trucks will prowl the highways any time soon. The technology is too complicated, he argues, and developers will have to go through years of safety testing before they’re ready for the roads—and before the public feels safe riding in their own bitty cars around 50,000-pound robot trucks. So Peloton is going all in on making human-based driving both safer and more efficient. With a bit of tech boost.

Not all manufacturers agree: In January, Daimler announced it would stop its platooning development to focus on autonomous trucking. Tests showed that “fuel savings, even in perfect platooning conditions, are less than expected,” the German company wrote in a press release. “At least for U.S. long-distance applications, analysis currently shows no business case for customers driving platoons with new, highly aerodynamic trucks.”

Platooning advocates disagree, but even the most supportive believe finding a market for this trucker assistance isn’t simple. Steven Shladover is researcher with the California Partners for Advanced Transportation Technology program at UC Berkeley. He has studied platooning for two decades, and points out that the truck industry would need to execute a fair bit of choreography to pull off platooning. Fleet operators would have to coordinate deliveries, matching up trucks heading in the same direction at the same time. “Does the truck industry see enough of a benefit in platooning to fit it into their operational strategies?” he says.

While everyone in trucking waits to find out, Perkins and Gregory head back to Peloton’s test track and proceed to show off a few, freakier moves: some hard braking, some driving side-by-side to prove that the trucks can still “talk” to each other in that position. At one point, another company employee in a white Toyota Tundra cuts into the 55-foot space between the two trucks, and they smoothly part to make room for him. Maybe platooning will improve life for truckers—too bad it can’t fix the problem of everyday reckless drivers, too.


More Great WIRED Stories

Google to prompt Android users to choose preferred browsers to allay EU concerns

FILE PHOTO: A 3D printed Android mascot Bugdroid is seen in front of a Google logo in this illustration taken July 9, 2017. REUTERS/Dado Ruvic/Illustration

BRUSSELS (Reuters) – Alphabet’s Google will prompt Android users to choose their preferred browsers and search apps, a senior Google executive said on Tuesday, as the company seeks to allay EU antitrust concerns and ward off fresh sanctions.

The European Commission last year handed Google a record 4.34 billion euro ($4.9 billion) fine for using the market power of its mobile software to block rivals in areas such as internet browsing.

By pre-installing its Chrome browser and Google search app on Android devices, Google had an unfair advantage over its rivals, EU enforcers said.

Google will now try to ensure that Android users are aware of browsers and search engines other than its own services, Kent Walker, senior vice-president of global affairs, said in a blog.

“In the coming months, via the Play Store, we’ll start asking users of existing and new Android devices in Europe which browser and search apps they would like to use,” he wrote without providing details.

The company, which introduced a licensing fee for device makers to access its app marketplace after the EU sanction, does not plan to scrap the charge.

Google could be fined up to 5 percent of Alphabet’s average daily worldwide turnover if it fails to comply with the EU order to stop anti-competitive practices.

Reporting by Foo Yun Chee; Editing by David Goodman

Samsung Electronics sees tough 2019 for component business: CEO

The logo of Samsung Electronics is seen at its office building in Seoul, South Korea January 7, 2019. REUTERS/Kim Hong-Ji

SEOUL (Reuters) – Samsung Electronics Co Ltd expects a tough year for its component business including memory chips due to sluggish growth in the smartphone market and reduced investment from data center companies, Chief Executive Kim Ki-nam said on Wednesday.

He was speaking at the South Korean tech giant’s annual general meeting where shareholders are expected to vote on the appointment of board directors.

Samsung is seeking new growth in areas such as network equipment manufacturing as sales of its mainstay chips and smartphones begin to drop.

The company would continue to make bold investments in semiconductor manufacturing in the face of stiffening Chinese competition, Kim said.

Reporting by Ju-min Park; additional reporting by Hyunjoo Jin and Heekyong Yang; Editing by Stephen Coates

PagerDuty Joins A Flurry Of Silicon Valley Companies Planning To Go Public This Year

POWERFUL WOMEN

Jennifer Tejada, chief executive officer of PagerDuty Inc., speaks during the Fortune’s Most Powerful Women conference in Dana Point, California, U.S., on Wednesday, Oct. 3, 2018. The conference brings together leading women in business, government,© 2017 Bloomberg Finance LP

PagerDuty took the next step forward to a planned IPO, joining a windfall of startups expected to go public this year. But the cloud-based software company’s debut will be an exception among the tech IPO wave—it’s one of the few enterprise companies run by a woman, CEO Jennifer Tejada.

Founded in 2009, San Francisco-based PagerDuty acts as a watchdog for technical issues. The operations management software identifies problems in real time and directs engineers to the root of the problem, an alert system that’s attracted 10,800 customers in 90 countries.

In 2018, PagerDuty scored unicorn status after a $90 million round led by T. Rowe Price Associates and Wellington Management. Its first nine months of revenue last year rose 48% from the period to $84 million. However, the company took a $34.5 million loss during that time,up $4.7 million from 2017. It didn’t reveal data on the full year.

The company’s institutional investors own more than half of its shares, including early investor, Andreessen Horowitz, which owns the largest share of the company at 18.4%, followed by Accel and Bessemer Venture Partners. PagerDuty’s cofounders, Baskar Puvanathasan, Andrew Miklas and Alex Solomon, each hold 7.1%.

PagerDuty landed a spot in the top 50 on the Forbes Cloud 100 list in 2017, just a year after Tejada took over as CEO. “It was a neat brand, even though it’s a small company,” Tejada told Forbes back in July 2016. Tejada owns over four million shares of the company.

U.S. prosecutors probing Facebook's data deals: New York Times

(Reuters) – U.S. federal prosecutors are conducting a criminal investigation into data deals Facebook Inc struck with some of the world’s largest technology companies, the New York Times reported on Wednesday.

A grand jury in New York has subpoenaed records from at least two prominent makers of smartphones and other devices, the newspaper reported, citing people familiar with the requests and without naming the companies.

Both companies are among the more than 150, including Amazon.com Inc, Apple Inc and Microsoft Corp, that have entered into partnerships with Facebook for access to the personal information of hundreds of millions of its users, according to the report.

Facebook is facing a slew of lawsuits and regulatory inquiries over its privacy practices, including ongoing investigations by the U.S. Federal Trade Commission, the Securities and Exchange Commission and two state agencies in New York.

In addition to looking at the data deals, the probes focus on disclosures that the company shared the user data of 87 million people with Cambridge Analytica, a British consulting firm that worked with U.S. President Donald Trump’s campaign.

Facebook said it was cooperating with investigators in multiple federal probes, without addressing the grand jury inquiry specifically.

“We’ve provided public testimony, answered questions, and pledged that we will continue to do so,” Facebook said in a statement.

Facebook has defended the data-sharing deals, first reported in December, saying none of the partnerships gave companies access to information without people’s permission.

A spokesman for the United States attorney’s office for the Eastern District of New York, which The New York Times reported is overseeing the inquiry, said he could not confirm or deny the probe.

Reporting by Ismail Shakil in Bengaluru and Katie Paul in San Francisco; Editing by Richard Chang and Leslie Adler

How the FAA Decides When to Ground a Jet Like Boeing’s 737 MAX 8

When an Ethiopian Airlines Boeing 737 MAX 8 jet crashed shortly after takeoff from Addis Ababa on Sunday morning, killing all 157 people aboard, observers quickly noted that the circumstances resembled those of another flight. In October, Lion Air Flight 610 crashed into the Java Sea, killing all 181 passengers and eight crew. Both flights plummeted a few minutes after takeoff, in good weather. And both were on 737 MAX 8 jets, the plane Boeing started delivering in 2017 to replace the outgoing 737 as the workhorse of the skies. Since 2017, Boeing has delivered 387 MAX 8s and 9s. It has taken orders for 4,400 more, from more than 100 customers.

As of Tuesday evening, various foreign aviation regulators and airlines had decided that after the two crashes, the plane shouldn’t be in the air. Officials in the European Union, China, Indonesia, Singapore, Australia, and the United Arab Emirates have all grounded the planes. Of the 59 operators that fly the new 737, at least 30 have parked it.

In the US, though, Boeing’s plane is free to fly. American Airlines, Southwest Airlines, and United Airlines are still putting their 737 MAX jets—74 in total—in the air. (So is Air Canada.) And the Federal Aviation Administration—the agency that oversees American airspace—says that’s just fine.

Which might seem strange, since the FAA is notoriously safety-conscious. Planes in search of an airworthiness certificate must meet stringent standards; the certification process usually takes years. And it gets results: Just one person has died in American airspace on a commercial airplane since 2009. But, it seems, the agency has not yet found reason to ground the new 737.

In a statement Tuesday, acting FAA administrator Daniel Elwell said the agency is looking at all the available data from 737 operators around the world, and that the review “thus far shows no systematic performance issues and provides no basis to order grounding aircraft.” Elwell said the FAA “would take immediate appropriate action” should such problems be identified. The FAA and the National Transportation Safety Board both have teams at the crash site outside Addis Ababa to investigate and collect data.

The agency did note in a directive published Monday that it would probably mandate flight control system enhancements that Boeing is already working on, come April. And after the Lion Air crash, the FAA made a Boeing safety warning mandatory for US airlines.

“We have full confidence in the safety of the 737 MAX,” Boeing said in its own statement Tuesday. “Based on the information currently available, we do not have any basis to issue new guidance to operators.”

A number of senators, including Ted Cruz of Texas, Elizabeth Warren of Massachusetts, and Dianne Feinstein of California, have called for the US to ground the aircraft. But it’s the FAA chief who has final say. (Elwell has been the acting administrator since January 2018, though Politico reports that the Trump Administration is close to nominating Delta Air Lines executive Steve Dickson as administrator.) He doesn’t make that decision alone, says Clint Balog, a flight test pilot and human factors expert with the College of Aeronautics at Embry-Riddle University. Any grounding goes through a “semi-formal” process, full of discussions with experts on the specific aircraft and crash situation, both in- and outside the federal government.

“The FAA looks at all of this information and decides, ‘OK, if it’s just likely that there’s a significant problem here, it doesn’t matter what the cost to the traveling public is—we have to put safety first and ground this aircraft,’” Balog says. “However, if they look and say, ‘Well, jeez, grounding this aircraft is going to be a monumental cost to the world and we simply don’t have enough information to know what the risk really is with this aircraft, do we really want to ground it at this point in time?’”

The FAA has grounded aircraft before. In 1979, the FAA grounded all McDonnell Douglas DC-10s (and forbid the aircraft from US airspace) after a crash in Chicago killed 273 people. An investigation found the problem was maintenance issues, not the aircraft design, the FAA lifted the prohibition just over a month later.

In early 2013, the FAA grounded Boeing’s 787 Dreamliner, after two lithium ion-battery related fires in the aircraft. “We are issuing this [directive] because we evaluated all the relevant information and determined the unsafe condition described previously is likely to exist or develop in other products of the same type design,” the FAA wrote in its emergency airworthiness directive. It didn’t let the jet take to the sky again until Boeing found and corrected its design issues. (That happened in April.)

So far, though, we have little concrete information on whatever might be going on with the 737 MAX. The investigation into the Ethiopia crash is in its earliest stages. Indonesia’s civil aviation authority has released a preliminary report on the Lion Air crash, but has not issued any findings on what caused it.

Based on its directives, the FAA hasn’t “seen any red flags that are significant enough” to ground the aircraft, Balog says. So he’d have no problem getting on a 737 MAX-8. “More importantly, I would have no problem having my family get on a 737 MAX-8 at this point.”


More Great WIRED Stories

Elon Musk shoots down U.S. regulator's complaint about his Tesla tweet

SAN FRANCISCO – Tesla Inc (TSLA.O) Chief Executive Elon Musk shot back against U.S. securities regulators on Monday, arguing in a filing that his recent tweet about the electric vehicle maker’s production volume did not violate his fraud settlement and he cannot be held in contempt.

FILE PHOTO: Tesla CEO Elon Musk attends the Tesla Shanghai Gigafactory groundbreaking ceremony in Shanghai, China January 7, 2019. REUTERS/Aly Song/File Photo

Musk’s “single, immaterial” tweet to his more than 24 million Twitter followers claiming the electric vehicle-maker would produce around 500,000 cars in 2019 also complied with company’s communication policy for senior executives, a condition of the settlement, lawyers for the Tesla chief wrote in a filing in federal court in Manhattan.

The U.S. Securities and Exchange Commission had asked the court to hold Musk in contempt, saying his Feb. 19 tweet violated a September fraud settlement barring him from sharing material information about Tesla on social media without the company’s pre-approval.

“This contempt action, following Musk’s sincerely-held criticism of the SEC on 60 Minutes, also reflects concerning and unprecedented overreach on the part of the SEC,” the filing said.

In December, Musk said in a “60 Minutes” interview that: “I do not have respect for the SEC.”

The settlement between Musk, Tesla and the SEC resolved an SEC lawsuit over claims Musk made on Twitter in August that he had “funding secured” to take Tesla private at $420 per share. The SEC called those tweets “false and misleading” and a go-private deal never materialized.

As part of that settlement, Musk stepped down as the company’s chairman and he and Tesla agreed to pay $20 million each in fines.

The renewed public battle between Tesla’s chief executive and the top U.S. securities regulator adds pressure on Musk, the public face of Tesla, who is struggling to make the company profitable after cutting the price of its Model 3 sedan to $35,000.

Tesla has backed off a plan to close all its U.S. stores and said it will instead raise prices of its higher-end vehicles by about 3 percent on average.

In the filing, Musk’s lawyers said his tweet was a “proud and optimistic restatement of publicly disclosed information.”

Musk corrected his tweet four hours later to say that the “annualized production rate” at year-end 2019 would probably be about 500,000, with deliveries expected to be about 400,000.

Moreover, Musk has exhibited self-censorship in dramatically reducing the volume of tweets since the settlement, they wrote, adding that the SEC’s request, if granted, would raise free speech issues.

“This self-censorship is reflective of his commitment to adhering to the Order and avoiding unnecessary disputes with the SEC,” they wrote in the filing.

Musk called the regulator the “Shortseller Enrichment Commission” on Twitter after the settlement, and tweeted that “something is broken with SEC oversight” just one day after the agency started pursuing the contempt order.

Legal experts have said the SEC could now pursue multiple avenues, including a higher fine, imposing further restrictions on Musk’s activities or removing him from Tesla’s board or helm.

Tesla published a new communications policy in December for senior executives as part of the settlement. It called for Tesla’s general counsel and a newly designated in-house securities law attorney to pre-approve any written statements about Tesla that could be material.

A disclosure controls committee, made up of board members Brad Buss, Antonio Gracias and James Murdoch, was tasked with overseeing compliance with the new policy.

Reporting by Alexandria Sage in San Francisco, and Rama Venkat and Subrat Patnaik in Bengaluru; Additional reporting by Sonam Rai in Bengaluru; Editing by Meredith Mazzilli and Lisa Shumamker

Uber not criminally liable in fatal 2018 Arizona self-driving crash: prosecutors

(Reuters) – Uber Technologies Inc is not criminally liable in a March 2018 crash in Tempe, Arizona, in which one of the company’s self-driving cars struck and killed a pedestrian, prosecutors said on Tuesday.

FILE PHOTO: National Transportation Safety Board (NTSB) investigators examine a self-driving Uber vehicle involved in a fatal accident in Tempe, Arizona, U.S., March 20, 2018. National Transportation Safety Board/Handout via REUTERS

The Yavapai County Attorney said in a letter made public that there was “no basis” for criminal liability for Uber, but that the conduct of the back-up driver, Rafael Vasquez, should be referred to the Tempe police for additional investigation.

Police said last year that Vasquez was streaming a television show on a phone until about the time of the crash and called the incident “entirely avoidable.”

An Uber spokeswoman declined to comment on the letter.

Vasquez could face charges of vehicular manslaughter, according to a police report last June. Vasquez has not previously commented and could not immediately be reached on Tuesday.

The Maricopa County Attorney, whose jurisdiction includes Tempe, referred the case last year to another office because of a conflict. In Tuesday’s letter Yavapai County Attorney Sheila Sullivan Polk said its investigation concluded that “the collision video, as it displays, likely does not accurately depict the events that occurred.”

The letter said an “expert analysis” is needed to “closely match what (and when) the person sitting in the driver’s seat of the vehicle would or should have seen that night given the vehicle’s speed, lighting conditions, and other relevant factors.”

The National Transportation Safety Board and National Highway Traffic Safety Administration are still investigating the fatal crash.

The Uber car was in autonomous mode at the time of the crash, but the company, like other self-driving car developers, requires a back-up driver inside to intervene when the autonomous system fails or a tricky driving situation occurs.

The Tempe police report said Vasquez repeatedly looked down and not at the road, glancing up a half second before the car hit Elaine Herzberg, 49, who was crossing the street at night.

Police obtained records from Hulu, an online service for streaming TV shows and movies, which showed Vasquez’s account was playing the TV talent show “The Voice” for about 42 minutes on the night of the crash, ending at 9:59 p.m., which “coincides with the approximate time of the collision,” the report said.

The Maricopa County Attorney’s Office did not immediately comment on Tuesday.

In December, Uber resumed limited self-driving car testing on public roads in Pittsburgh, nine months after it suspended the program following the Arizona crash.

The company is now testing with two employees in the front seat and more strictly monitor safety employees. The company also said last year it made improvements to the vehicles’ self-driving software.

In March 2018, authorities in Arizona suspended Uber’s ability to test its self-driving cars. Uber also voluntarily halted its entire autonomous car testing program.

The NTSB has said Uber had disabled a manufacturer-installed automatic emergency braking system in the 2017 Volvo XC90 while the car was under computer control in order to “reduce the potential for erratic vehicle behavior.”

Reporting by David Shepardson; Editing by Bill Berkrot and Grant McCool

Electric scooter sharing firm VOI raises $30 million for European expansion

STOCKHOLM (Reuters) – Electric scooter sharing firm VOI Technology has raised $30 million in another fundraising round since being set up seven months ago for its European expansion and investment in research to fend off growing competition, it said on Monday.

Uber Technologies Inc, Alphabet and several other high-profile investors are betting scooter-sharing will rise rapidly in Europe thanks to large commuter populations and lower levels of car ownership than in the United States.

Having already put many scooters on European roads, domestic startups such as Tier and Dott and U.S. rivals Bird and Lime raised thousands of dollars in 2018 to expand further into the crowded marketplace.

VOI, backed by investors such as BlaBlaCar CEO Nicolas Brusson and venture fund Balderton Capital, believes it can beat rivals by building closer relationships with city authorities.

“Asking ‘permission’ before we enter new towns and cities, unlike some of our competitors, means we can work with the authorities on the ground to offer not only a viable alternative to cars,” CEO Fredrik Hjelm said.

This could also “help people to combine their e-scooter journeys with the existing public transport network,” he added.

People can locate nearby VOI scooters via its app or maps and then ride it by paying a 1 euro unlocking fee plus riding costs of 0.15 euro per minute.

Since its August launch, VOI has built up over 400,000 riders, taking more than 750,000 rides, and it said it would use the new funds to expand in Italy, Germany, Norway and France.

Critics warn operators could face similar issues as bike sharing firms. Forced into price wars due to competition and facing backlash from authorities over rules and vandalism, bike operators GoBee and Mobike have retreated from Europe.

Reporting by Esha Vaish in Stockholm; editing by David Evans

Tencent, Kakao among shortlisted bidders for South Korean gaming firm Nexon: Korea Economic Daily

FILE PHOTO: A sign of Tencent is seen during the fourth World Internet Conference in Wuzhen, Zhejiang province, China, Dec. 3, 2017. REUTERS/Aly Song/File Photo

SEOUL (Reuters) – Chinese tech giant Tencent and South Korea’s biggest chat operator Kakao Corp are among five bidders shortlisted to buy South Korean gaming firm Nexon, the Korea Economic Daily newspaper reported on Monday.

The newspaper said NXC Corp, the holding firm which controls Nexon, also had shortlisted private equity firms Bain Capital and MBK Partners, as well as an unidentified private equity firm.

South Korean publisher Netmarble Corp was not offered to join the preliminary bidding, but formed a consortium with MBK Partners to bid for Nexon, according to the report which cited investment banking sources.

The deal could worth as much as 15 trillion won ($13.3 billion) and the formal bid process was set to take place in early April, it added.

Nexon founder Kim Jung-ju has been exploring a sale of a 98.64 percent stake in NXC Corp that is held by him and related parties including his wife.

MBK Partners declined to comment on the report while Tencent, Kakao, Netmarble and Bain Capital were not immediately available for comment.

Reporting by Heekyong Yang and Hyunjoo Jin; Editing by Stephen Coates

Exclusive: China regulator requests pause in new game applications to clear backlog – sources

BEIJING/SHANGHAI (Reuters) – China’s top content regulator has asked local authorities to stop submitting requests to monetize new video games while it processes a backlog of applications built up after a lengthy pause last year, three people with knowledge of the matter said.

FILE PHOTO: A man plays a computer game at an internet cafe in Beijing,China May 9, 2014. REUTERS/Kim Kyung-Hoon/File Photo

The General Administration of Press and Publications (GAPP) issued the notice this week, the people said, indicating the impact on gaming stocks of the nine-month hiatus could continue and dulling hopes raised by the recent resumption of approvals.

The regulator’s notice has not previously been reported.

Shares of industry leader Tencent Holdings Ltd, which were up 2.2 percent in morning trading, pared back gains to trade about 1 percent higher after the Reuters report was published. Shares of smaller players also slid.

China stopped approving the monetization of new titles last March amid a regulatory body reshuffle triggered by growing criticism of games being violent and addictive, as well as concern over the increase in myopia among young people.

Gaming firms such as Tencent – China’s most valuable listed company – were able to continue filing applications, building up a backlog. They could also distribute new titles but were unable to earn any income from them, such as through in-game purchases.

The regulator resumed processing applications in December, with industry insiders estimating at least 5,000 games await approval. In China, game companies file applications to local authorities which in turn submit them to the regulator.

“The regulator asked local authorities to stop submitting applications because there is too much of a backlog for it to deal with at the moment,” said one of the people, whose company was informed about the matter by its local authority.

Game companies will still be able to file applications but they will no longer be passed on to the Beijing regulator while it deals with applications already in hand, said a second person.

The people declined to be identified as they were not authorized to speak with media on the matter.

GAPP and the Propaganda Department of the Communist Party of China, which oversees GAPP, did not immediately respond to requests for comment.

The approval freeze dragged down shares in Tencent and wiped billions of dollars off its market value. Among titles for which Tencent is awaiting a license to monetize is “PlayerUnknown’s Battlegrounds Mobile”, which industry insiders estimate could generate annual revenue of up to $1 billion.

The freeze has also hit many smaller companies that rely on a number of game releases each year.

The regulator approved 1,982 domestic and foreign online games during January-March last year before the freeze, government data showed. It approved 9,651 domestic and foreign online games in all of 2017.

It has approved 538 games since December.

Reporting by Pei Li in BEIJING and Brenda Goh in SHANGHAI; Editing by Christopher Cushing