The Japanese anime studio, Khara, is moving to Blender, the the open-source 3D creation software. "It'll begin partially using the software for its current development 'EVANGELION:3.0+1.0' but will make the full switch once that project is finished," reports Neowin. "The current project is expected to end in June next year, so after that point, its employees will start using Blender for the majority of their work." From the report: At the moment, Khara uses 3ds Max from Autodesk on a subscription basis; however, the company found that it had to reach out to small and medium-sized businesses for its projects. Due to the limitations of those companies, it's harder for them to afford 3ds Max. By switching to Blender, Khara says it can work better with external firms.

While Blender will be used for the bulk of the work, Khara does have a backup plan if there's anything Blender struggles with; Hiroyasu Kobayashi, General Manager of Digital Dpt. and Director of Board of Khara, said: "There are currently some areas where Blender cannot take care of our needs, but we can solve it with the combination with Unity. Unity is usually enough to cover 3ds Max and Maya as well. Unity can be a bridge among environments."

An anonymous reader quotes a report from CNBC: Nvidia's CEO, Jensen Huang, has reason to be concerned about other chipmakers, like AMD. But he's not worried about Nvidia's own big customers turning into competitors. Amazon, Facebook, Google and Tesla are among the companies that buy Nvidia's graphics cards and have kicked off chip-development projects. "There's really one I know of that have silicon that's really in production," Huang told CNBC in an interview on Thursday. That company would be Google, he said. "But our conversation with large customers is intensifying," Huang said. "We're talking to more large customers."

Google first announced its entrance into the data center AI chip-making world in 2016. As it came up with new versions, the web company pointed to performance advantages over graphics cards that were available at the time. Google hasn't started selling data center chips for training AI models to other companies, though. (Google has started offering various products that use its Edge tensor processing unit chips, but those chips aren't as powerful as the TPU chips for training AI models in Google's cloud.)

A survey of more than 6,000 firmware images spanning more than a decade finds no improvement in firmware security and lax security standards for the software running connected devices by Linksys, Netgear and other major vendors. The Security Ledger reports: "Nobody is trying," said Sarah Zatko, the Chief Scientist at the Cyber Independent Testing Lab (CITL), a non-profit organization that conducts independent tests of software security. "We found no consistency in a vendor or product line doing better or showing improvement. There was no evidence that anybody is making a concerted effort to address the safety hygiene of their products," she said. The CITL study surveyed firmware from 18 vendors including ASUS, D-link, Linksys, NETGEAR, Ubiquiti and others. In all, more than 6,000 firmware versions were analyzed, totaling close to 3 million binaries created from 2003 to 2018. It is the first longitudinal study of IoT software safety, according to Zatko. CITL researchers studied publicly available firmware images and evaluated them for the presence of standard security features such as the use of non-executable stacks, Address Space Layout Randomization (ASLR) and stack guards, which prevent buffer overflow attacks.

The results were not encouraging. Time and again, firmware from commonly used manufacturers failed to implement basic security features even when researchers studied the most recent versions of the firmware. For example: firmware for the ASUS RT-AC55U wifi router did not employ ASLR or stack guards to protect against buffer overflow attacks. Nor did it employ a non-executable stack to protect against "stack smashing," another variety of overflow attack. CITL found the same was true of firmware for Ubiquiti's UAP AC PRO wireless access points, as well as DLink's DWL-6600 access point. Router firmware by vendors like Linksys and NETGEAR performed only slightly better on CITL's assessment. CITL researchers also "found no clear progress in any protection category over time," reports The Security Ledger. "Researchers documented 299 positive changes in firmware security scores over the 15 years covered by the study... but 370 negative changes over the same period. Looking across its entire data set, in fact, firmware security actually appeared to get worse over time, not better."

On the bright side, the survey found that almost all recent router firmware by Linksys and NETGEAR boasted non-executable stacks. "However, those same firmware binaries did not employ other common security features like ASLR or stack guards, or did so only rarely," says the report.

Uber and Lyft tapped transportation consultancy Fehr & Peers to examine their combined vehicle miles traveled (VMT) in six cities in September 2018, and compare that number to the total VMT in each area for the same month. "The results show that while they are vastly out-stripped by personal and commercial vehicles, Uber and Lyft are still responsible for significant shares of VMT in those cities," reports The Verge. From the report: The analysis looks at Boston, Chicago, Los Angeles, San Francisco, Seattle, and Washington, DC. The results are presented on two levels: the regional area, including the surrounding towns and suburbs, and the "core" county of each region that contains the main part of the city with the densest concentration of jobs. The findings show that Uber and Lyft account for just 1-3 percent of VMT in the broader metropolitan areas of each city. But those numbers spike when zooming in on the core county of each city. In San Francisco County, for example, Uber and Lyft make up as much as 13.4 percent of all vehicle miles. In Boston, it's 8 percent; in Washington, DC, it's 7.2 percent.

These figures suggest that Uber and Lyft are hitting some cities harder than previously thought. An independent study commissioned by the San Francisco County Transportation Authority looked at 2017 traffic patterns in the county and concluded that TNCs generated about 6.5 percent of the total VMT on weekdays, and 10 percent on weekends. (TNC, which stands for transportation network company, is an industry term used to describe ride-hailing apps like Uber and Lyft.) The findings from Fehr & Peers show totals "nearly twice that previous estimate," said Gregory Erhardt, a professor of civil engineering at the University of Kentucky who has researched Uber and Lyft's effects on public transit ridership. "This difference may be due to the continued increase in TNC use over the intervening two years."

A computer issue is preventing U.S. immigration officials from processing arriving passengers at several airports across the country. "It wasn't immediately clear what caused the problem at U.S. Customs and Border Protection, but the agency said it was investigating," reports CNBC. Agents will be processing people manually until the systems are resolved.

One Twitter user posted a video of a long line of an estimated 5,000+ passengers at Dulles International Airport. Another traveler tweeted that passengers on her flight from London weren't allowed to deplane upon arrival in Boston because of the issue.

UPDATE 21:28 UTC: NBC News has confirmed that the system shutdown is nationwide, and U.S. Customs and Border Patrol officials are working to determine the cause. You can view the thread here.

Story is developing...

For years, Naval Research Laboratory meteorologist David Peterson has been obsessed with one of Earth's rarest atmospheric spectacles: thunderclouds formed by raging wildfires. Last week, he became one of the only people on Earth to fly straight through one. From a report: Peterson is the lead forecaster for Fire Influence on Regional to Global Environments and Air Quality (FIREX-AQ), a joint NASA and NOAA-led field campaign that's spending the summer intensively studying wildfire smoke from the ground, the air, and satellites. On August 8, he rode shotgun as NASA's DC-8 research aircraft passed directly through an anvil cloud as it was developing over the 45,000-acre Williams Flats fire currently burning in the Pacific Northwest.

Over the next few hours, the plane would conduct the most detailed reconnaissance ever from within a pyrotechnic weather system, making observations and collecting samples that will help researchers to better understand the nature of these dramatic events and how they can impact Earth's climate. "Just being there was the most amazing experience I've ever had while working in science," Peterson said. Pyrocumulonimbus clouds (pyroCbs) only form when conditions are just right -- you need a special combination of atmospheric instability, moisture, and loads of wildfire heat to create an updraft.

Sam Biddle, reporting for The Intercept: Nearly 1,500 miles from the Menlo Park headquarters of Facebook, at a company outpost in Austin, Texas, moderators toil around the clock to screen and scrub some the most gruesome, hateful, and heinous posts that make their way onto the social network and its photo-sharing subsidiary, Instagram. They are required to view as many as 800 pieces of disturbing content in a single shift, and routinely turn to on-site counselors to help cope with the procession of stomach-turning images, videos, and text. But some members of this invisible army have complained, in a statement widely circulated within Facebook, that the outsourcing giant that officially employs them, Accenture, has repeatedly attempted to violate the confidentiality of these therapy sessions.

The moderators work from within a special section for outsourced staffers at Facebook Austin. The Texas outpost is designed to mimic the look and feel of the company's famously opulent Silicon Valley digs, but Accenture workers say they're reminded daily of their secondary status and denied perks, prestige, and basic respect. This second-class tier at Facebook, a sort of international shadow workforce, has been well documented in the media, from Manila to Arizona, and it's not clear whether the company has done anything to address it beyond issuing defensive PR statements. Moderators in Austin say their job is a brutalizing slog and that Facebook remains largely indifferent to their struggles. Access to on-site counseling is one of the few bright points for this workforce.

But now even this grim perk has been undermined by corporate prying, according to a letter drafted by a group of about a dozen Austin moderators who work across Facebook and Instagram. The letter alleges that, starting in early July, Accenture managers attempted to pressure multiple on-site counselors to share information relating to topics discussed in employee trauma sessions. This information was understood by both counselors and Accenture employees to be confidential, said several Accenture sources interviewed by The Intercept. It is not clear what specific information related to the sessions was sought by the managers.

An anonymous reader shares a report: Platforms like Facebook, YouTube, and Twitter are banking on developing artificial intelligence technology to help stop the spread of hateful speech on their networks. The idea is that complex algorithms that use natural language processing will flag racist or violent speech faster and better than human beings possibly can. Doing this effectively is more urgent than ever in light of recent mass shootings and violence linked to hate speech online. But two new studies show that AI trained to identify hate speech may actually end up amplifying racial bias. In one study [PDF], researchers found that leading AI models for processing hate speech were one-and-a-half times more likely to flag tweets as offensive or hateful when they were written by African Americans, and 2.2 times more likely to flag tweets written in African American English (which is commonly spoken by black people in the US). Another study [PDF] found similar widespread evidence of racial bias against black speech in five widely used academic data sets for studying hate speech that totaled around 155,800 Twitter posts.

This is in large part because what is considered offensive depends on social context. Terms that are slurs when used in some settings -- like the "n-word" or "queer" -- may not be in others. But algorithms -- and content moderators who grade the test data that teaches these algorithms how to do their job -- don't usually know the context of the comments they're reviewing. Both papers, presented at a recent prestigious annual conference for computational linguistics, show how natural language processing AI -- which is often proposed as a tool to objectively identify offensive language -- can amplify the same biases that human beings have. They also prove how the test data that feeds these algorithms have baked-in bias from the start.

Police have urged everyone to download a smartphone app they say has already saved several lives. What is it and how does it work? From a report: Kicked. Converged. Soccer. These three randomly chosen words saved Jess Tinsley and her friends after they got lost in a forest on a dark, wet night. They had planned a five-mile circular stroll through the 4,900 acre (2,000 hectare) woodland Hamsterley Forest, in County Durham, on Sunday evening, but after three hours they were hopelessly lost. "We were in a field and had no idea where we were," the 24-year-old care worker from Newton Aycliffe said. "It was absolutely horrendous. I was joking about it and trying to laugh because I knew if I didn't laugh I would cry."

At 22:30 BST they found a spot with phone signal and dialled 999. "One of the first things the call-handler told us to do was download the what3words app," Ms Tinsley said. "I had never heard of it." Within a minute of its download, the police said they knew where the group was and the soaked and freezing walkers were swiftly found by the Teesdale and Weardale Search and Mountain Rescue Team. "I have told everyone I know to download this app," Ms Tinsley said. "You never know when you are going to get lost and need it." What3words essentially points to a very specific location. Its developers divided the world into 57 trillion squares, each measuring 3m by 3m (10ft by 10ft) and each having a unique, randomly assigned three-word address. For example, the door of 10 Downing Street is slurs.this.shark, while the area across the road where the press congregate is stage.pushy.nuns.

On Wednesday, a horde of Amazon employees rushed to defend the honor of the world's richest man. From a report: The employees -- known as "FC ambassadors," for Amazon's Orwellian-sounding "fulfillment center" warehouses -- flooded into the mentions of several Twitter users. This isn't the first time these accounts have attracted attention. Earlier this year, the accounts descended on Twitter with coordinated anti-union talking points. The FC Ambassador program made its debut last August and also seemed to coordinate talking points about how great it was to work at Amazon. The program backfired most recently after a Twitter thread of FC ambassadors went viral, with many people saying the tweets were dark or dystopian (others raised doubts about their legitimacy).

When TechCrunch first covered the Amazon FC Ambassador accounts in August 2018, there were around 14 accounts. That army has grown. This week, Motherboard found more than 40 FC Ambassador accounts on Twitter which appear to be genuine. Open source intelligence collective Bellingcat did their own investigation, and found close to 60 accounts. The accounts are spread across the world, with users not just based in the U.S., but Spain, the UK, Germany, and elsewhere too.

The self-driving freight truck startup TuSimple has been carrying mail across the state of Arizona for several weeks. From a report: UPS announced on Thursday that its venture capital arm has made a minority investment in TuSimple. The announcement also revealed that since May TuSimple autonomous trucks have been hauling UPS loads on a 115-mile route between Phoenix and Tucson. UPS confirmed to Gizmodo this is the first time UPS has announced it has been using TuSimple autonomous trucks to deliver packages in the state. Around the same time as the UPS and TuSimple program began, the United States Postal Service and TuSimple publicized a two-week pilot program to deliver mail between Phoenix and Dallas, a 1,000 mile trip. TuSimple claims it can cut the average cost of shipping in a tractor-trailer by 30 percent. In an announcement about the new partnership, UPS Ventures managing partner, Todd Lewis, said the venture arm "collaborates with startups to explore new technologies and tailor them to help meet our specific needs."

A new Google study this week confirmed the obvious: internet users need to stop using the same password for multiple websites unless they're keen on having their data hijacked, their identity stolen, or worse. From a report: It seems like not a day goes by without a major company being hacked or leaving user email addresses and passwords exposed to the public internet. These login credentials are then routinely used by hackers to hijack your accounts, a threat that's largely mitigated by using a password manager and unique password for each site you visit. Sites like "have I been pwned?" can help users track if their data has been exposed, and whether they need to worry about their credentials bouncing around the dark web. But it's still a confusing process for many users unsure of which passwords need updating.

To that end, last February Google unveiled a new experimental Password Checkup extension for Chrome. The extension warns you any time you log into a website using one of over 4 billion publicly-accessible usernames and passwords that have been previously exposed by a major hack or breach, and prompts you to change your password when necessary. The extension was built in concert with cryptography experts at Stanford University to ensure that Google never learns your usernames or passwords, the company says in an explainer. Anonymous telemetry data culled from the extension has provided Google with some interesting information on how widespread the practice of account hijacking and non-unique passwords really is.

YouTube is making a change to its copyright enforcement policies around music used in videos, which may result in an increased number of blocked videos in the shorter term -- but overall, a healthier ecosystem in the long-term. From a report: Going forward, copyright owners will no longer be able to monetize creator videos with very short or unintentional uses of music via YouTube's "Manual Claiming" tool. Instead, they can choose to prevent the other party from monetizing the video or they can block the content. However, YouTube expects that by removing the option to monetize these sorts of videos themselves, some copyright holders will instead just leave them alone. "One concerning trend we've seen is aggressive manual claiming of very short music clips used in monetized videos. These claims can feel particularly unfair, as they transfer all revenue from the creator to the claimant, regardless of the amount of music claimed," explained YouTube in a blog post.

To be clear, the changes only involve YouTube's Manual Claiming tool which is not how the majority of copyright violations are handled today. Instead, the majority of claims are created through YouTube's Content ID match system. This system scans videos uploaded to YouTube against a database of files submitted to the site by copyright owners. Then, when a match is found, the copyright holder owner can choose to block the video or monetize it themselves, and track the video's viewership stats.

Is there room for another social media platform? ShareChat, a four-year-old social network in India that serves tens of million of people in regional languages, just answered that question with a $100 million financing round led by global giant Twitter. From a report: ShareChat serves 60 million users each month in 15 regional languages, Ankush Sachdeva, co-founder and CEO of the firm, told TechCrunch in an interview. The platform currently does not support English, and has no plans to change that, Sachdeva said. That choice is what has driven users to ShareChat, he explained. The early incarnation of the social media platform supported English language. It saw most of its users choose English as their preferred language, but this also led to another interesting development: Their engagement with the app significantly reduced.

Add another chapter to your internet revisionist history books: AOL held talks to buy both Facebook and YouTube in 2006 and considered taking a large minority stake in Tencent in 2004. From a report: Obviously none of this happened -- and the board of Time Warner is to blame, said ex-AOL CEO Jon Miller in an exclusive CNBC interview. Miller has never discussed the failed talks publicly before. Miller said he discussed buying YouTube from the founders, Chad Hurley, Steve Chen and Jawed Karim, in January and July 2006. He spoke with Facebook founder Mark Zuckerberg in the spring of that year, he said. The Tencent talks were held in 2004, Miller said. "We wanted to take some shots," Miller said. "We had a line on buying YouTube before anybody else. We had an opportunity to step in with Facebook when Yahoo stumbled. We had a chance to maybe to step in to Tencent."

DigiDay has reviewed a pitch deck detailing a 2017 bidding program for the "Amazon's Choice" label, a mark that denotes than an item is recommended and gives certain products higher placement in search results. While the mark is now believed to be operated by Amazon's algorithms, sources say that Amazon actually offered sellers the chance to bid on the mark back in 2017. An anonymous Slashdot reader shares the report: Amazon has previously offered vendors the ability to "bid" for an Amazon's Choice badge by lowering prices and spending more money on advertising, bringing into sharper focus how the program, which recently came under fire from senators, actually works. It's unclear whether or not this offer was taken up by any Amazon vendors, or how long the program was offered before it was discontinued. One source believed it was only offered for a few months. The criteria for each bid focused on reviews and in-stock rate. To be considered, Amazon required brands to be able to keep products in stock for a 12-month duration, keep customer review ratings above four stars, and maintain certain technical specifications for their respective subcategories. Essentially, Amazon was only interested in bids from brands already selling high-quality products on its platform. From there, Amazon wanted to drive attention to products in each category that would maximize its profitability. In an email to a brand client, an Amazon strategic vendor service manager recommended the brand drop its selling price by $30 in order to win the bid, a move that would increase Amazon's profits on the sale. For vendors, access to a strategic vendor service manager, considered to provide inside access to Amazon, costs hundreds of thousands of dollars per year. When asked if the program existed, an Amazon spokesperson denied that this program was offered.

Apple has filed a lawsuit against Corellium, accusing the software company of illegally selling virtual copies of iOS under the guise of helping discover security flaws. "Apple said the software company Corellium has copied the operating system, graphical user interface and other aspects of the devices without permission, and wants a federal judge to stop the violations," reports Bloomberg. From the report: Apple said it supports "good-faith security research," offering a $1 million "bug bounty" for anyone who discovers flaws in its system and gives custom versions of the iPhone to "legitimate" researchers. Corellium, the iPhone maker said, goes further than that. "Although Corellium paints itself as providing a research tool for those trying to discover security vulnerabilities and other flaws in Apple's software, Corellium's true goal is profiting off its blatant infringement," Apple said in the complaint. "Far from assisting in fixing vulnerabilities, Corellium encourages its users to sell any discovered information on the open market to the highest bidder."

Corellium creates copies of the Apple iOS, and says that it's all to help white-hat hackers discover security flaws. Instead, according to Apple, any information is sold to people who can then exploit those flaws. Corellium, in a posting dated July 4 on its website, said it "respects the intellectual property rights of others and expects its users to do the same." Corellium's products allow the creation of a virtual Apple device, according to the suit. It copies new versions of Apple works as soon as they are announced, and doesn't require users to disclose flaws to Apple, the Cupertino, California-based company said in the complaint. Apple also wants a court order forcing Corellium to notify its customers that they are in violation of Apple's rights, destruction of any products using Apple copyrights, and cash compensation.

Disney and Charter Communications are teaming up to fight account sharing in an attempt to prevent multiple people from using a single account to access streaming video services. Ars Technica reports: The battle against account sharing was announced as Disney and the nation's second-biggest cable company struck a new distribution agreement involving Disney's Hulu, ESPN+, and the forthcoming Disney+. Customers could still buy those online services directly from Disney, but the new deal would also let them make those purchases through Charter's Spectrum TV service. If you buy a Disney service through Charter, be aware that the companies will work together to prevent you from sharing a login with friends. Disney and Charter said in their announcement yesterday that they have "agreed to work together on piracy mitigation. The two companies will work together to implement business rules and techniques to address such issues as unauthorized access and password sharing."

The crackdown could target people who use Charter TV account logins to sign into Disney services online. Charter CEO Tom Rutledge has complained about account sharing several times over the past few years while criticizing TV networks for not fully locking down their content. "There's lots of extra streams, there's lots of extra passwords, there's lots of people who could get free service," Rutledge said at an industry conference in 2017. He argues that password sharing has helped people avoid buying cable TV. ESPN has also complained about account sharing, calling it piracy. Another possibility is that Charter could monitor usage of its broadband network to help Disney fight account sharing. For example, Disney could track the IP addresses of users signing in to its services, and Charter could match those IP addresses to those of its broadband customers.

Engineer.ai, an Indian startup claiming to have built an artificial intelligence-assisted app development platform, is not in fact using AI to literally build apps, according to a report from The Wall Street Journal. Instead, the company, which has attracted nearly $30 million in funding from a SoftBank-owned firm and others, is reportedly relying mostly on human engineers, while using hype around AI to attract customers and investment that will last it until it can actually get its automation platform off the ground. The Verge reports: The company claims its AI tools are "human-assisted," and that it provides a service that will help a customer make more than 80 percent of a mobile app from scratch in about an hour, according to claims Engineer.ai founder Sachin Dev Duggal, who also says his other title is "Chief Wizard," made onstage at a conference last year. However, the WSJ reports that Engineer.ai does not use AI to assemble the code, and instead uses human engineers in India and elsewhere to put together the app. When pressed on how the company actually employs machine learning and other AI training techniques, the company told the WSJ it uses natural language processing to estimate pricing and timelines of requested features, and that it relies on a "decision tree" to assign tasks to engineers. Neither of those really qualify as the type of modern AI that powers cutting-edge machine translation or image recognition, and it does not appear that any kind of AI agent or software of any kind is actually compiling code.

An anonymous reader quotes a report from The Economist: Many parents and grandparents will grumble about today's screen-obsessed youth. Indeed, researchers find that millennials look at their phones more than 150 times a day; half of them check their devices in the middle of the night; a third glance at them immediately after waking up. And yet, when all screens are accounted for, it is in fact older folk who seem most addicted. According to Nielsen, a market-research firm, Americans aged 65 and over spend nearly ten hours a day consuming media on their televisions, computers and smartphones. That is 12% more than Americans aged 35 to 49, and a third more than those aged 18 to 34 (the youngest cohort for whom Nielsen has data). American seniors "spend an average of seven hours and 30 minutes in front of the box, about as much as they did in 2015," the report says. "The spend another two hours staring at their smartphones, a more than seven-fold increase from four years ago."

Millennials have increased the time they spend on their mobile devices, but it's been largely offset by their dwindling interest in TV. As for teenagers, a report from 2015 by Common Sense Media "found that American teens aged 13-18 spent about six hours and 40 minutes per day on screens: slightly more than Nielsen recorded for 18- to 34-year-olds that year, but less than older generations."