One year after its first 1.0 release, Brave says it has hit the milestone of 20.5 million active monthly users. "At the same time last year, the browser had 8.7 million active monthly users, and of the 20 million monthly users, 7 million are daily users, which represents more than a doubling of last year's 3 million," reports ZDNet. "Brave added that since Apple allowed browsers other than its own to be the default option on iOS, it has seen its iOS user base increase by a third." From the report: One of the touted features of the browser is that it hates ads, and will go out of its way to block them, unless users decide to see Brave-powered advertisements. To that end, Brave has hit "2 billion ad confirmation events" and completed 2,215 campaigns from over 460 companies. The browser maker says its users have a click-through rate of 9%, way and away outstripping industry averages. The browser also has its own cryptocurrency, Basic Attention Tokens, that users use to "tip" content creators. Thus far, 26 million of the tokens have been sent to creators. At the time of writing, the blockchain-based token is trading for just under 18 cents, meaning $4.6 million has been sent from users.

Anonymouse Cowtard writes: When you're holding millions in bets on the favorite, you don't take the decision to payout lightly. Betting agencies are often good barometers for political polls, but Sportsbet has been stung before. They incorrectly paid out on bets for the Labor Party in Australia to win two days before the election was held, assuming a victory would come to the polling leaders. That party lost.

According to Bloomberg, "Sportsbet has decided to pay out early on Joe Biden to be elected as the U.S President even with the official result still not known, according to a statement. The online bookmaker has deemed Biden's lead as 'unassailable' after CNN projected Michigan would go to the Democrats and that he holds the ascendancy in Nevada and Arizona. Sportsbet has paid out over 100,000 bets to punters who backed Biden."

According to Phoronix, a Dell privacy driver is is being prepared for the Linux kernel. From the report: Beginning in Dell's 2021 laptop models they are providing hardware-based "privacy buttons" to disable microphone and camera support. These new Dell privacy buttons are basically hardware kill switches for the microphone and web camera video stream. The Dell privacy driver sent out on Tuesday for the Linux kernel is about manipulating the relevant LEDs and tracking the status of the hardware-based controls where as the actual toggling of the audio/video support is handled by the hardware.

The Dell privacy driver in its current form is talked about for the camera and microphone support but the patch does also note a "PRIVACY_SCREEN_STATUS" bit as well. Presumably they will be extending this privacy driver as well for privacy screen handling around reducing the horizontal/vertical viewing angles of the display. The dell-privacy Linux driver in its initial form can be found via the kernel mailing list. It's great seeing Dell working on this driver punctually for Linux ahead of their next-gen laptops.

An anonymous reader quotes a report from ZDNet: GrowDiaries, an online community where marijuana growers can blog about their plants and interact with other farmers, has suffered a security breach in September this year. The breach occurred after the company left two Kibana apps exposed on the internet without administrative passwords. Kibana apps are normally used by a company's IT and development staff, as the app allows programmers to manage Elasticsearch databases via a simple web-based visual interface. Due to its native features, securing Kibana apps is just as important as securing the databases themselves.

But in a report published today on LinkedIn, Bob Diachenko, a security researcher known for discovering and reporting unsecured databases, said GrowDiaries failed to secure two of its Kibana apps, which appear to have been left exposed online without a password since September 22, 2020. Diachenko says these two Kibana apps granted attackers access to two sets of Elasticsearch databases, with one storing 1.4 million user records and the second holding more than two million user data points. The first exposed usernames, email addresses, and IP addresses, while the second database also exposed user articles posted on the GrowDiaries site and users' account passwords. While the passwords were stored in a hashed format, Diachenko said the format was MD5, a hashing function known to be insecure and crackable (allowing threat actors to determine the cleartext version of each password). The company secured its infrastructure five days after Diachenko reported the exposed Kibana apps on October 10. It's unknown if someone else accessed the databases to download user data.

Virgin Hyperloop One announced its plan to build a $500 million certification center to advance its vision of the future of high-speed transportation in West Virginia. The state will serve as a locus for testing, developing, and validating the technology that underpins the still-theoretical hyperloop system. The Verge reports: There is no fully functional hyperloop in the world, and it has never been tested with human passengers. But the federal government has recently laid out the framework for regulating the hyperloop, giving hope to companies like Virgin Hyperloop One that it may eventually break ground on a full-sized operational hyperloop system. To do so, it will still need to raise millions of dollars in funding, acquire the enormous tracks of land, and certify that the hyperloop can be operated safely. Which is all to say, the hyperloop is still very far off.

The certification center is a first step toward that goal, said Richard Branson, founder of Virgin Group, which acquired Hyperloop One in 2017. "The Hyperloop Certification Center is the start of the hyperloop journey for West Virginia, for the United States, and for the world," Branson said in a statement. "We're one step closer to making hyperloop travel a reality for people everywhere." [...] Virgin Hyperloop One is seeking to build its first hyperloop system not in the US, but in India. Its project in the western district of Maharashtra aims to become the first system certified as safe for human passengers. Officials there have named the company and its partners as the originators of the multibillion-dollar infrastructure project.

Massachusetts overwhelmingly voted to extend its automobile right to repair law, in a huge win for consumers. From a report: Question 1 was the most expensive measure battle in Massachusetts history with the auto industry (and independent repair companies) spending tens of millions of dollars lobbying, according to the Boston Globe. The measure is an essential win for independent mechanics, auto-repair shops, and consumers, as it will require car manufacturers to continue to make diagnostic tools available for years to come. Under the law, car manufacturers will be required to use an open-data system in cars using telematics. This means mechanics will have access to wirelessly sent repair data -- whether they are associated with an official car dealership or an independent shop. While cars currently use a wired connection for diagnostics, there was concern among independent repair professionals that car manufacturers would switch to a wireless system in order to circumvent a 2012 right to repair law that required car dealers make wired repair codes universal. As new car models are produced in coming years, the thought is many will ditch physical diagnostic ports and instead, cars will wirelessly send repair information. Tuesday's ballot measure closes a loophole in the 2012 law that would have exempted wireless diagnostics from the law. Right to repair advocates have hailed the measure. iFixit's Kevin Purdy wrote of the significance of the vote, "that means that independent repair shops will have a level playing field with car makers and dealerships, which have turned increasingly to locked-down wirelessly collected repair data, or telematics. Car owners, too, will be able to see their cars' maintenance information through a smartphone app. And it opens the door for innovations, like wireless diagnostic apps for iOS and Android."

More than 23,000 hacked databases have been made available for download on several hacking forums and Telegram channels in what threat intel analysts are calling the biggest leak of its kind. From a report: The database collection is said to have originated from Cit0Day.in, a private service advertised on hacking forums to other cybercriminals. Cit0day operated by collecting hacked databases and then providing access to usernames, emails, addresses, and even cleartext passwords to other hackers for a daily or monthly fee. Cybercriminals would then use the site to identify possible passwords for targeted users and then attempt to breach their accounts at other, more high-profile sites. The idea behind the site isn't unique, and Cit0Day could be considered a reincarnation of similar "data breach index" services such as LeakedSource and WeLeakInfo, both taken down by authorities in 2018 and 2020, respectively.

T-Mobile is paying the price for the reported misdeeds of its recent acquisition. From a report: The carrier has agreed to pay a $200 million settlement over FCC allegations Sprint abused the Lifeline program for low-income communications, claiming subsidies for 885,000 customers that weren't using the initiative. The agreement also requires that T-Mobile honor the rules of a consent decree.The provider has to reform company procedures and training to ensure that it only makes legitimate Lifeline claims. A senior manager has to ensure T-Mobile complies with the order, and the company has to submit periodic reports for three years after the start of the decree.

An anonymous reader quotes a report from ZDNet: [E]ven today's most sophisticated GPS systems are still unable to map a huge chunk of the Earth: that which is located under oceans, seas, or rivers. The technology, in effect, doesn't mix well with water, which breaks down the radio waves GPS relies on to function. MIT scientists have been looking at ways to create a new type of underwater GPS, which could be used to better understand the mysteries that lie between surface and seabed. The researchers have now unveiled a device called an underwater backscatter localization (UBL) that reacts to acoustic signals to provide positioning information, even when it is stuck in oceanic depths. All of this, without even using a battery.

Underwater devices already exist, for example to be fitted on whales as trackers, but they typically act as sound emitters. The acoustic signals produced are intercepted by a receiver that in turn can figure out the origin of the sound. Such devices require batteries to function, which means that they need to be replaced regularly -- and when it is a migrating whale wearing the tracker, that is no simple task. On the other hand, the UBL system developed by MIT's team reflects signals, rather than emits them. The technology builds on so-called piezoelectric materials, which produce a small electrical charge in response to vibrations. This electrical charge can be used by the device to reflect the vibration back to the direction from which it came. In the researchers' system, therefore, a transmitter sends sound waves through water towards a piezoelectric sensor. The acoustic signals, when they hit the device, trigger the material to store an electrical charge, which is then used to reflect a wave back to a receiver. Based on how long it takes for the sound wave to reflect off the sensor and return, the receiver can calculate the distance to the UBL.

In practice, piezoelectric materials are no easy component to work with: for example, the time it takes for a piezoelectric sensor to wake up and reflect a sound signal is random. To solve this problem, the scientists developed a method called frequency hopping, which involves sending sound signals towards the UBL system across a range of frequencies. Because each frequency has a different wavelength, the reflected sound waves return at different phases. Using a mathematical theorem called an inverse Fourier transform, the researchers can use the phase patterns and timing data to reconstruct the distance to the tracking device with greater accuracy. Frequency hopping showed some promising results in deep-sea environments, but shallow waters proved even more problematic. Because of the short distance between surface and seabed, sound signals uncontrollably bounce back and forth in lower depths, as if in an echo chamber, before they reach the receiver -- potentially messing with other reflected sound waves in the process. [...] While the scientists acknowledged that addressing these challenges would require further research, a proof-of-concept for the technology has already been tested in shallow waters, and MIT's team said that the UBL system achieved centimeter-level accuracy.

Baby Shark, the infuriatingly catchy children's rhyme recorded by South Korean company Pinkfong, has become the most-watched video ever on YouTube. The BBC reports: The song has now been played 7.04 billion times, overtaking the previous record holder Despacito, the Latin pop smash by singer Luis Fonsi. Played back-to-back, that would mean Baby Shark has been streamed continuously for 30,187 years. Pinkfong stands to have made about $5.2 million from YouTube streams alone.

It took four years for Baby Shark to ascend to the top of YouTube's most-played chart, but the song is actually much older than that. It is thought to have originated in U.S. summer camps in the 1970s. One theory says it was invented in 1975, as Steven Spielberg's Jaws became an box office smash around the world. There are a huge number of variations on the basic premise, including one version where a surfer loses an arm to the shark, and another where the protagonist dies. There are also international versions - including the French Bebe Requin and the German Kleiner Hai (Little Shark), which became a minor hit in Europe in 2007. But none of them could match the phenomenal success of Pinkfong's interpretation, which was sung by 10-year-old Korean-American singer Hope Segoine and uploaded to YouTube in 2015.

Mint's programmers, led by lead developer, Clement "Clem" Lefebvre, have built their own take on Google's open-source Chromium web browser. ZDNet reports: Some of you may be saying, "Wait, haven't they offered Chromium for years? Well, yes, and no. For years, Mint used Ubuntu's Chromium build. But then Canonical, Ubuntu's parent company, moved from releasing Chromium as an APT-compatible DEB package to a Snap. The Ubuntu Snap software packing system, along with its rivals Flatpak and AppImage, is a new, container-oriented way of installing Linux applications. The older way of installing Linux apps, such as DEB and RPM package management systems for the Debian and Red Hat Linux families, incorporate the source code and hard-coded paths for each program.

While tried and true, these traditional packages are troublesome for developers. They require programmers to hand-craft Linux programs to work with each specific distro and its various releases. They must ensure that each program has access to specific libraries' versions. That's a lot of work and painful programming, which led to the process being given the name: Dependency hell. Snap avoids this problem by incorporating the application and its libraries into a single package. It's then installed and mounted on a SquashFS virtual file system. When you run a Snap, you're running it inside a secured container of its own. For Chromium, in particular, Canonical felt using Snaps was the best way to handle this program. [...]

Lefebvre wrote, "The Chromium browser is now available in the official repositories for both Linux Mint and LMDE. If you've been waiting for this I'd like to thank you for your patience." Part of the reason was, well, Canonical was right. Building Chromium from source code is one really slow process. He explained, "To guarantee reactivity and timely updates we had to automate the process of detecting, packaging and compiling new versions of Chromium. This is an application which can require more than 6 hours per build on a fast computer. We allocated a new build server with high specifications (Ryzen 9 3900, 128GB RAM, NMVe) and reduced the time it took to build Chromium to a little more than an hour." That's a lot of power! Still, for those who love it, up-to-date builds of Chromium are now available for Mint users.

An anonymous reader quotes a report from BleepingComputer: The data analysts firm NetMarketShare revealed that Windows 10 has seen another uptake in users and it went up to 64.04% from 61.26% last month. Linux (multiple distros) went from 1.14% to 1.65% and Ubuntu now holds a market share of 0.51%. The market share of Windows 7 has also dropped, but many users are still actively using outdated Windows 7, which could be due to its huge number of enterprise users. According to NetMarketShare, Windows 7 saw a drop from 22.77% to 20.41% last month. The report shows that 20.41% of desktops still use Windows 7. Even worse, some are still using Windows XP, according to the report. As of October 2020, the market share of Windows XP is 0.87%.

An anonymous reader quotes a report from Ars Technica: After years of building political pressure for antitrust scrutiny of major tech companies, this month Congress and the US government delivered. The House Antitrust Subcommittee released a report accusing Apple, Amazon, Google, and Facebook of monopolistic behavior. The Department of Justice filed a complaint against Google alleging the company prevents consumers from sampling other search engines. The new fervor for tech antitrust has so far overlooked an equally obvious target: US broadband providers. "If you want to talk about a history of using gatekeeper power to harm competitors, there are few better examples," says Gigi Sohn, a fellow at the Georgetown Law Institute for Technology Law & Policy. Sohn and other critics of the four companies that dominate US broadband -- Verizon, Comcast, Charter Communications, and AT&T -- argue that antitrust intervention has been needed for years to lower prices and widen Internet access. Analysis by Microsoft last year concluded that as many as 162.8 million Americans do not use the Internet at broadband speeds (as many as 42.8 million lack meaningful broadband), and New America's Open Technology Institute recently found that US consumers pay, on average, more than those in Europe, Asia, or elsewhere in North America.

The Department of Justice complaint against Google argues that the company's payments to Apple to set its search engine as the default on the iPhone make it too onerous for consumers to choose a competing search provider. For tens of millions of Americans, changing broadband providers is even more difficult -- it requires moving. The Institute for Local Self-Reliance, which promotes community broadband projects, recently estimated from Federal Communications Commission data that some 80 million Americans can only get high-speed broadband service from one provider. "That is quite intentional on the part of cable operators," says Susan Crawford, a professor at Harvard Law School. "These companies are extracting rent from Americans based on their monopoly positions." The United States has suffered, and broken up, telecom monopolies in the past. AT&T had a government-sanctioned monopoly for much of the 20th century, until it was broken up in 1984. The 1996 Telecom Act included rules for phone providers aimed at encouraging competition, but it excluded "information services," leaving broadband companies freer rein. The White House and Congress will both need to act in order to make US broadband more competitive. "Options worth considering include reversing some of the acquisitions that turned Comcast and others into nation-spanning giants and mandating that companies allow competitors to use their networks, as is common in Europe, [says Joshua Stager, a senior policy counsel at New America's Open Technology Institute.]

On Election Day, a subset of Instagram users woke up to a message at the top of their Instagram feeds that read, "Tomorrow is Election Day." For some, the message was still there by the early afternoon. Instagram chalked the outdated message up to a caching issue that caused some users to continue seeing Monday's election notification. From a report: "While we turned off the 'Tomorrow is Election Day' notice last night, it was cached for a small group of people if their app hadn't been restarted," Instagram wrote on Twitter on Tuesday, responding to a number of users who had received the message. "It's resolving itself as people restart."

The man on the trail went by "Mostly Harmless." He was friendly and said he worked in tech. After he died in his tent, no one could figure out who he was. Wired: It's usually easy to to put a name to a corpse. There's an ID or a credit card. There's been a missing persons report in the area. There's a DNA match. But the investigators in Collier County couldn't find a thing. Mostly Harmless' fingerprints didn't show up in any law enforcement database. He hadn't served in the military, and his fingerprints didn't match those of anyone else on file. His DNA didn't match any in the Department of Justice's missing person database or in CODIS, the national DNA database run by the FBI. A picture of his face didn't turn up anything in a facial recognition database. The body had no distinguishing tattoos.

Nor could investigators understand how or why he died. There were no indications of foul play, and he had more than $3,500 cash in the tent. He had food nearby, but he was hollowed out, weighing just 83 pounds on a 5'8" frame. Investigators put his age in the vague range between 35 and 50, and they couldn't point to any abnormalities. The only substances he tested positive for were ibuprofen and an antihistamine. His cause of death, according to the autopsy report, was "undetermined." He had, in some sense, just wasted away. But why hadn't he tried to find help? Almost immediately, people compared Mostly Harmless to Chris McCandless, whose story was the subject of Into the Wild. McCandless, though, had been stranded in the Alaska bush, trapped by a raging river as he ran out of food. He died on a school bus, starving, desperate for help, 22 miles of wilderness separating him from a road. Mostly Harmless was just 5 miles from a major highway. He left no note, and there was no evidence that he had spent his last days calling out for help.

The investigators were stumped. To find out what had happened, they needed to learn who he was. So the Florida Department of Law Enforcement drew up an image of Mostly Harmless, and the Collier County investigators shared it with the public. In the sketch, his mouth is open wide, and his eyes too. He has a gray and black beard, with a bare patch of skin right below the mouth. His teeth, as noted in the autopsy, are perfect, suggesting he had good dental care as a child. He looks startled but also oddly pleased, as if he's just seen a clown jump out from behind a curtain. The image started to circulate online along with other pictures from his campsite, including his tent and his hiking poles.

An excerpt from MIT Technology Review's interview with Geoffrey Hinton: You think deep learning will be enough to replicate all of human intelligence. What makes you so sure?
I do believe deep learning is going to be able to do everything, but I do think there's going to have to be quite a few conceptual breakthroughs. For example, in 2017 Ashish Vaswani et al. introduced transformers, which derive really good vectors representing word meanings. It was a conceptual breakthrough. It's now used in almost all the very best natural-language processing. We're going to need a bunch more breakthroughs like that.

And if we have those breakthroughs, will we be able to approximate all human intelligence through deep learning?
Yes. Particularly breakthroughs to do with how you get big vectors of neural activity to implement things like reason. But we also need a massive increase in scale. The human brain has about 100 trillion parameters, or synapses. What we now call a really big model, like GPT-3, has 175 billion. It's a thousand times smaller than the brain. GPT-3 can now generate pretty plausible-looking text, and it's still tiny compared to the brain.

When you say scale, do you mean bigger neural networks, more data, or both?
Both. There's a sort of discrepancy between what happens in computer science and what happens with people. People have a huge amount of parameters compared with the amount of data they're getting. Neural nets are surprisingly good at dealing with a rather small amount of data, with a huge numbers of parameters, but people are even better.

A lot of the people in the field believe that common sense is the next big capability to tackle. Do you agree?
I agree that that's one of the very important things. I also think motor control is very important, and deep neural nets are now getting good at that. In particular, some recent work at Google has shown that you can do fine motor control and combine that with language, so that you can open a drawer and take out a block, and the system can tell you in natural language what it's doing. For things like GPT-3, which generates this wonderful text, it's clear it must understand a lot to generate that text, but it's not quite clear how much it understands. But if something opens the drawer and takes out a block and says, "I just opened a drawer and took out a block," it's hard to say it doesn't understand what it's doing.

An anonymous reader shares a report: Less than 24 hours before a historic US presidential election day, Nick Clegg, Facebook's vice president of global affairs and communications and the former United Kingdom deputy prime minister, tried to rally employees at the embattled social networking corporation. Noting that the world would be watching the results, Clegg published a post on an internal message board about the work Facebook employees had done to prepare for the vote. Many things had changed since 2016, he said, alluding to an election in which Russian state actors used Facebook to sow discord, while the company and CEO Mark Zuckerberg stood by oblivious. "We have transformed the way we approach elections since the U.S. presidential election four years ago," Clegg wrote in the note titled "READY FOR ELECTION DAY." "Thanks to the efforts of far, far too many of you to mention by name, Facebook is a very different company today." It is indeed. Roiled by months of internal scandals and high-profile failures, the social network giant heads into Election Day with employee morale cratering and internal political discussion muzzled on internal message boards.

While Clegg took an optimistic tone in his post, Facebook released results of an internal survey on Monday that revealed a stark decline in employee confidence over the past six months. Its semi-annual "Pulse Survey," taken by more than 49,000 employees over two weeks in October, showed workers felt strained by office shutdowns and were continuing to lose faith that the company was improving the world. Only 51% of respondents said they believed that Facebook was having a positive impact on the world, down 23 percentage points from the company's last survey in May and down 5.5 percentage points from the same period last year. In response to a question about the company's leadership, only 56% of employees had a favorable response, compared to 76% in May and more than 60% last year. (A Facebook employee acknowledged in the announcement that the uptick in May's Pulse results were "likely driven by our response to COVID-19," which was widely praised.)

AmiMoJo writes: As we roll into 2021 the Free Software Foundation is looking to update its high priority free software projects list. These are the software projects that should be incorporating "the most important threats, and most critical opportunities, that free software faces in the modern computing landscape." For now the FSF is looking for help deciding what to include. The FSF high priority projects list is what once included PowerVR reverse engineering as being very important albeit never happened prior to PowerVR graphics becoming less common. In fact, many FSF high priority projects never panned out as they weren't contributing much in the way of resources to the causes but just calling attention to them. PDF support was among their high priority projects as well as another example as well as the likes of an open-source Skype replacement and reverse-engineering other popular technologies.

They overhauled the list in 2017 after forming a committee to maintain the list while now as 2021 is just around the corner they are looking to revise their high priority projects focus once more. They have issued a call for input to share with the High Priority Free Software Projects committee what you feel should belong on the list. Feedback is being collected through early January. Currently on the list are different "areas" they feel are high priority for free software as opposed to previously focusing on particular projects.

AmiMoJo writes: Fact checking is increasingly a mainstay of the modern internet, and that now includes "dead" web pages. The Internet Archive has started adding fact checks and context to Wayback Machine pages to explain just why they were removed. If a page was part of a disinformation campaign or pulled due to a policy violation, a conspicuous yellow banner will explain as much. The checks come from a variety of well-established outlets, including FactCheck.org, Politifact, the AP and the Washington Post.

The archivists saw the fact checks as striking a balance between historical preservation and acknowledging the problems with resurfacing false info. It hoped users would "better understand what they are reading" in the archives. It's also striving for neutrality -- one banner for context explained that including a page in the Wayback Machine "should not be seen" as endorsing the content.

The COVID-19 exposure notification app used in England and Wales failed to warn users if they were in close contact with potentially infectious COVID-19 patients. Because of the error, thousands of people were not told to quarantine even when they should have been, according to the Sunday Times, which first reported the flaw. The Verge reports: The app was launched on September 24th, and there have been 19 million downloads since. It's built using Google and Apple's Bluetooth Low Energy-based system, which monitors nearby phones. If someone tests positive for COVID-19, their app can alert the phones that they were in contact with. The UK originally planned to use its own app, sidestepping the Google and Apple system, but reversed course in June. Since its launch, the app flagged few users for possible exposure, according to the Sunday Times. Engineers figured out the reason why last week. The app was originally built to simply recommend that anyone closer than 2 meters for more than 15 minutes to someone who later tested positive should quarantine, according to The Guardian. But just before the launch, it was adjusted to take into account when the sick person's symptoms began.

Research shows that people tend to have high levels of virus in their nose and throat, and may be more contagious, around a day before they start to show symptoms. Levels stay high the first few days of symptoms, and then drop off. If someone was in contact with a sick person outside of that window, the UK app would consider the interaction less risky. If they were in contact inside of that window, on the other hand, it would only take three minutes of contact to trigger an alert. The adjusted app did calculate those new risk levels. But the thresholds at which a person would actually get an alert were left unchanged, according to a government blog post. Without the updated thresholds, a user could have spent up to 15 minutes with a highly infectious person and up to 40 minutes with a less-contagious person without getting an alert, according to The Guardian.