Baidu released a video search app called Kankan to find short videos and live streams, a move that may start to challenge Douyin, the Chinese version of TikTok — With Kankan, Baidu is competing for a larger slice of China's massive online video market, which has more than 880 million viewers …
Enlarge (credit: Nick Wright. Used by permission.)
For months, Apple’s corporate network was at risk of hacks that could have stolen sensitive data from potentially millions of its customers and executed malicious code on their phones and computers, a security researcher said on Thursday.
Sam Curry, a 20-year-old researcher who specializes in website security, said that, in total, he and his team found 55 vulnerabilities. He rated 11 of them critical because they allowed him to take control of core Apple infrastructure and from there steal private emails, iCloud data, and other private information.
The 11 critical bugs were:
Read 16 remaining paragraphs | Comments
The recently thwarted Michigan kidnapping plot highlights social media's role in promoting violent extremism, as well as building a case against extremists — In June, one of the suspects in the plot to kidnap Michigan Gov. Gretchen Whitmer took to the relative privacy of a Facebook group to make clear his brewing hatred.
Dispo, a disposable camera app co-created by YouTube star David Dobrik, raises $4M seed led by Alexis Ohanian's new fund Seven Seven Six — Seven Seven Six fund leads a $4 million seed round for social-media platform Dispo, also known as David's Disposable
Documents show that Airbnb spent $1.2B between mid-2019 and mid-2020, with the biggest portion of cash burn coming in Q1 2020 as it issued refunds amid COVID-19 — Airbnb burned through more than $1.2 billion in cash between mid-2019 and mid-2020, according to previously undisclosed figures …
A UK parliamentary inquiry concludes clear evidence of collusion between Huawei and the CPC apparatus exists; Huawei says report is built on opinions not facts — There is “clear evidence of collusion” between Huawei and the “Chinese Communist Party apparatus”, a parliamentary inquiry has concluded.
Federal judge ordered Twitter to reveal identity behind @whyspertech, who allegedly forged an FBI doc to spread conspiracy theory on killing of DNC's Seth Rich — A federal judge in California has ordered that Twitter reveal the identity of an anonymous user who allegedly fabricated …
A look at the scrutiny Google and Oracle faced from Supreme Court justices during the companies' arguments Wednesday, where no clear winner emerged — - Things got technical at the Supreme Court as the justices heard arguments from Google and Oracle in a blockbuster copyright dispute that has captivated Silicon Valley for a decade.
Amazon has filed a legal notice against Future Group, India's second largest retail chain, for selling a large portion of its business to Reliance Retail — Amazon has sent a legal notice to Future Group, India's second largest retail chain, for breaching the terms of its contract by selling …
Television screens airing the first presidential debate at the Walters Sports Bar in Washington on September 29. | Sarah Silbiger/Getty Images
The temporary ban doesn’t solve Facebook’s organic content problem or the problematic political ads appearing on its platform before voting.
Facebook is going to temporarily ban all political ads … but only after the 2020 election, a move that solves neither its organic content problem nor the problematic political ads appearing on its platform prior to voting.
On Wednesday, the social media giant announced that it will temporarily stop running social, electoral, and political ads in the United States after the polls close on Election Day, November 3. The measure is an effort “to reduce opportunities for confusion and abuse,” wrote Guy Rosen, Facebook’s vice president of integrity, in a blog post announcing the decision. The company will notify advertisers once it lifts the policy post-election, but it didn’t indicate when that would be. In early September, Facebook said it would ban new political ads the week before the election, but ads that have already been in the mix prior to then will continue to appear in News Feeds.
Also on Wednesday, Facebook said it would ban and remove posts that seek to intimidate voters, including ones that encourage poll watching “when those calls used militarized language or suggest that the goal is to intimidate, exert control, or display power over election officials and voters.”
These announcements are the latest in a series of small, slow, and iterative measures Facebook has introduced in recent months related to US politics and elections. And while they’re better than doing nothing, they are also too little, too late.
There’s a lot a political ad ban doesn’t do — it doesn’t stop politicians from lying in ads in the days leading up to the election, and it doesn’t stop giving political campaigns the ability to hyper-target ads to tiny groups of voters with very specific messaging. (Microtargeting makes it super easy to precisely target negative and misleading ads to certain voters, and it makes it harder for opponents and other groups to know those ads are out there and counter them.) Plus, banning political ads after the election doesn’t solve the, you know, before-the-election problem.
Some political strategists also argue that clamping down on political ads online hurts small campaigns more than it does the big ones. Facebook ads are a lot less expensive to run than television commercials — which means campaigns with big budgets can go to TV, while campaigns with small budgets can’t.
Also: Ads are just a small part of the equation. Facebook’s role in presenting voters with political information, disinformation, and conspiracy theories stretches far beyond advertising, and focusing too much on advertising allows it and other tech platforms to avoid the bigger problem: organic content. That means the type of stuff that goes on the platform for free — such as a false story in 2016 claiming Pope Francis had endorsed Donald Trump, or an unsubstantiated claim made by the president about mail-in-voting over the summer.
Misleading and polarizing organic content spreads fast and far on the platform all the time because social media thrives on engagement, and what engages people is content that evokes strong emotions. A political campaign doesn’t need to pay for a political ad to spread lies claiming Elizabeth Warren wasn’t born in the US or Marco Rubio has six secret love children — they can just post it.
The dangerous, preposterous QAnon conspiracy movement, which has shifted from the fringes to the mainstream, is a perfect example of social media’s failures. It has flourished on places like Facebook, Instagram, and Twitter — not because of ads, but because of organic content. Facebook finally banned QAnon this week, but it has already reached far and wide on its platform, as Recode’s Shirin Ghaffary recently explained:
The theory continues to grow online, in both the number of followers and the strength of its political influence in the Republican Party. The growing political clout of the movement is especially worrisome for misinformation researchers who say QAnon is potentially becoming one of the largest networks of extremism in the United States. QAnon is gaining broad appeal not just with the extremely online, male-dominated, 4chan message board crowd, where QAnon was first born; it’s also increasingly popular with suburban moms and yoga-loving wellness gurus on Instagram and Twitter.
Misinformation about voter fraud and the election is spreading — and it’s not relying on paid ads to do it. While Facebook tries to catch misinformation and put warnings on it, falsities travel a lot faster than its content moderators. In a hypothetical post-election scenario where Joe Biden wins the election but President Trump refuses to concede or insists the election was rigged, he doesn’t need an ad to spread that sort of lie — again, he can just do it in a post. Facebook says it will attach an “informational label” to content seeking to delegitimize the outcome of the election. However, it seems unlikely it would take down such posts or outright fact-check them, given how reluctant it’s been to take such actions in the past.
Will it work? When I consider my own anecdotal experience, I doubt it. I spent hours browsing the Facebook pages of anti-maskers for a story over the summer and encountered multiple people who had such labels on content they shared. They just dismissed them by claiming Facebook was censoring them or hiding the truth. They had also often developed their beliefs because of content they saw on Facebook or other online platforms.
In a September 3 post, Facebook CEO Mark Zuckerberg wrote that the 2020 election is “not going to be business as usual.” But the iterative measures Facebook has introduced so far seem to be exactly that — business as usual.
Will you help keep Vox free for all?
The United States is in the middle of one of the most consequential presidential elections of our lifetimes. It’s essential that all Americans are able to access clear, concise information on what the outcome of the election could mean for their lives, and the lives of their families and communities. That is our mission at Vox. But our distinctive brand of explanatory journalism takes resources. Even when the economy and the news advertising market recovers, your support will be a critical part of sustaining our resource-intensive work. If you have already contributed, thank you. If you haven’t, please consider helping everyone understand this presidential election: Contribute today from as little as $3.
Etsy says it will ban all merchandise related to the QAnon conspiracy theory, citing its policies against hate speech, inciting violence, and misinformation — A ghost. … - Digital marketplace Etsy has announced a ban on all items related to the QAnon conspiracy theory.
Google Ads introduces an Insights page, automated Performance Max campaigns and will bring Video Action out of beta to all advertisers in the coming weeks — The new products and features are further indication of Google's machine learning goals. — Google introduced more automation features …
Apple says it “vehemently disagrees” with the conclusions reached in US House report and doesn't have dominant marketshare in any categories of its businesses — Earlier today, the U.S. House Judiciary Antitrust Subcommittee completed its ongoing antitrust investigation …
Facebook CEO Mark Zuckerberg testifying in Congress in 2019 | Photo By Bill Clark/CQ-Roll Call, Inc via Getty Images
The report scrutinizes the ways that the four biggest tech companies have amassed enormous market power.
A long-awaited report from top Democratic Congressional lawmakers about the dominance of the four biggest tech giants had a clear message on Tuesday: Amazon, Apple, Facebook, and Google engage in a range of anti-competitive behavior, and US antitrust laws need an overhaul to allow for more competition in the US internet economy.
“To put it simply, companies that once were scrappy, underdog startups that challenged the status quo have become the kinds of monopolies we last saw in the era of oil barons and railroad tycoons,” the report’s introduction states.
The 400-plus page report, written by the majority staff of the Democratic members of the House Judiciary Subcommittee on Antitrust, is the result of a 16-month investigation into whether these corporate giants abuse their power, and whether the country’s antitrust laws need to be reworked to rein them in. The report released Tuesday cites numerous examples of each tech titan engaging in acts that the lawmakers believe have hurt innovation and impede competition. While the anti-competitive behaviors the report cites vary from company to company, they are all linked by the allegation that the four giants abuse their gatekeeper status in various internet industries to secure and grow their market power in those sectors and others.
So what’s the solution? The report from Democratic lawmakers recommends creating new laws that would potentially break up tech companies and make it harder for them to pursue acquisitions; it also calls for clarifying existing antitrust laws with the goal of making them easier to enforce, particularly for tech companies. For now, the report’s recommendations are only high-level guidance to Congress for potential future legislation; it won’t lead to immediate action against these companies.
The release of the report was complicated on Tuesday by news that the Republican lawmakers in the house antitrust committee refused at the last minute to sign the report with their Democratic colleagues. Instead, Rep. Ken Buck (R-CO) and Jim Jordan (R-OH) each plan to release their own reports. Buck’s report, a draft of which Politico published on Monday, largely agrees with the Democrat’s conclusion that the big four tech firms have amassed too much power. But he disagrees with Democrats on how to fix the problem: Instead of creating new laws, Buck’s memo calls on Congress to fund and empower regulatory agencies and government departments like the Federal Trade Commission (FTC) and Department of Justice (DOJ) to go after Big Tech under existing laws. Jordan’s report hasn’t yet been released, but Reuters coverage indicates it will focus on so-far unproven claims of tech companies’ supposed anti-conservative bias, which he has shouted over his colleagues about in previous hearings.
These partisan divides are somewhat besides the point: Regardless of the specifics of how they advise to go after Big Tech, the fact that Republicans and Democrats agree that these companies pose a threat to the free market is significant.
“This is the first time since the 1970s that a congressional committee has devoted this kind of attention to dominant firms … and changing the structure of a major American industry,” former FTC Commissioner William Kovacic, who was appointed by George W. Bush, told Recode.
Here’s a breakdown of some of the key claims the report makes about each of the four major tech giants:
With Amazon accounting for nearly 40 percent of all e-commerce sales in the US — making it more than seven times larger in this arena than No. 2 Walmart — the Democratic report argues that the tech giant has used its powerful position in anti-competitive ways. (The report also alleges that Amazon’s US e-commerce market share is closer to 50 percent or more in the US, rather than the near-40 percent figure commonly cited based on estimates from the research firm eMarketer). The report argues that the company unfairly gleans data and information from its third-party sellers that it uses to strengthen the retail side of its business, including by favoring its own product brands over those of competitors, giving this merchandise exclusive merchandising space on its virtual shelves, and prioritizing it in search results.
Another criticism is that Amazon can charge sellers ever-increasing fees because of its dominant position, and that most sellers and brands have practically no negotiating power because of their reliance on the Amazon sales channel. Amazon also penalizes sellers for selling their merchandise for lower prices on other retail sites.
Amazon released a company blog post in response to Tuesday’s report, calling it “flawed thinking” that Amazon is engaging in anti-competitive business practices, and that antitrust regulatory action “would have the primary effect of forcing millions of independent retailers out of online stores.”
“All large organizations attract the attention of regulators, and we welcome that scrutiny. But large companies are not dominant by definition, and the presumption that success can only be the result of anti-competitive behavior is simply wrong,” reads the post.
The report from Democrats argues that Facebook has expanded its monopolistic power in the social media industry by using a “copy, acquire, kill” strategy against its competitors and by unfairly hurting rival companies like Instagram (which the company purchased in 2012).
Specifically, the report argues that Facebook’s acquisition of Instagram was a blatant attempt to “neutralize a nascent competitive threat.” The report alleges that after Facebook bought Instagram, it intentionally stymied the photo-sharing app’s success so that it wouldn’t compete with Facebook internally.
The report cites a slew of internal emails, memos, and testimony from senior-level Facebook employees, including CEO Mark Zuckerberg, which support the argument that Facebook crushed Instagram by exerting monopoly power.
In one email, Zuckerberg told Facebook’s former CTO that “ that he had “been thinking about ... how much [Facebook] should be willing to pay to acquire mobile app companies like Instagram ... that are building networks that are competitive with our own.” The report argues this proves that Zuckerberg had anti-competitive interests from the beginning.
The report also cites a former senior-level Instagram employee who told Congress that Facebook CEO Mark Zuckerberg oversaw “brutal infighting between Instagram and Facebook” after the acquisition, with Zuckerberg slowing down Instagram’s natural growth to benefit Facebook proper. The Instagram whistleblower went so far as to call it “collusion, but within an internal monopoly. … It’s unclear to me why this should not be illegal.”
As part of their investigation, the subcommittee found an internal Facebook document called “The Cunningham Memo,” written in 2018 by Thomas Cunningham, a senior data scientist and economist at Facebook, which allegedly shows that Facebook has knowingly “tipped” its company toward becoming a monopoly, acknowledging that social media apps have tipping points where “either everyone uses them, or no-one uses them,” according to the memo. This memo was a key part of Zuckerberg’s acquisition strategy ahead of the Instagram purchase, according to internal documents and an interview the subcommittee conducted with a former Facebook employee involved with the project.
In a statement to Recode on Tuesday, Christopher Sgro, a spokesperson for Facebook, disagreed with the report’s conclusions. “Facebook is an American success story. We compete with a wide variety of services with millions, even billions, of people using them. Acquisitions are part of every industry, and just one way we innovate new technologies to deliver more value to people. Instagram and WhatsApp have reached new heights of success because Facebook has invested billions in those businesses. A strongly competitive landscape existed at the time of both acquisitions and exists today. Regulators thoroughly reviewed each deal and rightly did not see any reason to stop them at the time,” Sgro wrote.
The Democrats’ report argues that Google has a monopoly in the online search and marketing industry, creating an “ecosystem of interlocking monopolies” — which it has maintained through anti-competitive practices in two key ways.
The first is by launching an “aggressive campaign to undermine” what the report calls “vertical search providers” — which are search engines for specific topics, such as Yelp for restaurants, or Expedia for travel. The report says Google uses its dominance to “boost Google’s own inferior” content over some of these other companies’ content in its search results.
The second major way, that Google has demonstrated anti-competitive behavior, the report argues, is through “a series of anti competitive contracts” that pushed people to rely on Google search when using phones with the Android operating system (Google purchased Android in 2005).
“Documents show that Google required smartphone manufacturers to pre-install and give default status to Google’s own apps,” the report states.
Unsurprisingly, Google told Recode it disagreed with Tuesday’s reports, saying that they “feature outdated and inaccurate allegations from commercial rivals about Search and other services.
Americans simply don’t want Congress to break Google’s products or harm the free services they use every day,” read a statement in part from Julie McAlister, a spokesperson for Google.
According to the Democrats’ report on Tuesday, Apple exerts monopoly power through its oversight of software that’s downloaded on half of all mobile phones in the US. That’s a direct reference to Apple’s App Store — if you have an iPhone, you can only use apps that you download from the company’s tightly controlled Store. The subcommittee staff investigating Apple say in the report that the company has exploited its dominance to exclude some rivals from its store, unfairly favor its own apps, and charge fees that some app developers told the subcommittee are “exorbitantly high.”
Such a battle between Apple and developers over in-app fees exploded into public spotlight earlier this year when the maker of Fortnite, Epic Games, told its users they could buy the game’s virtual currency directly from Epic rather than through the Apple iOS version of the app. The reason? Epic wanted to avoid the 30 percent fee Apple charges for such in-app purchases. Dueling lawsuits ensued, and Apple even banned the game from the App Store. This is just one example of many cases like this that the report cites.
Apple, of course, refuted the conclusions in Tuesday’s report, telling Recode in a statement, “Our company does not have a dominant market share in any category where we do business. ... Last year in the United States alone, the App Store facilitated $138 billion in commerce with over 85% of that amount accruing solely to third-party developers. Apple’s commission rates are firmly in the mainstream of those charged by other app stores and gaming marketplaces.”
Depending on the results of the November election, Democrats may not need Republicans’ support on antitrust legislation — if Democrats sweep Congress and win the White House. (The latest polls show Democrats and Biden currently have an edge, but poll-based predictions are far from certain.)
If Biden does win the presidency, “this [report] is a roadmap for how you would tackle this under a President Joe Biden … administration,” a staff member for a Democratic member of the subcommittee told Recode.
Rep. Pramila Jayapal (D-WA), a member of the subcommittee, told Recode in an interview on Tuesday, “I do anticipate...that we will have signed pieces of legislation pass the House of Representatives next year.” The bi-partisan subcommittee will meet later this year to debate and potentially amend the report.
And Tuesday’s congressional reports are just the beginning of upcoming antitrust regulatory proceedings against Big Tech. The Department of Justice is imminently expected to file a lawsuit against Google for anti-competitive business practices, which several state attorney generals may sign on to. Separately, the FTC is also investigating the business practices of the tech giants over antitrust concerns.
Facebook says it is launching an emotional health resource center, with expert guides and information, available globally with location-specific resources — Instagram, WhatsApp and Messenger will also get new tools to help support users' emotional health during the coronavirus pandemic.
Uni, a new startup by PayU co-founder Nitin Gupta, announces $18.5M seed round led by Lightspeed and Accel for “building the modern age consumer credit card” — Even as close to a billion debit cards are in use in India today, only about 58 million credit cards are in circulation in the world's second most populous nation.
Researcher says that Macs with T2 chips are vulnerable to a variant of the checkm8 exploit, which could jailbreak certain iPhones and was unpatchable — Jailbreak involves combining last year's checkm8 exploit with the Blackbird vulnerability disclosed this August.
To operate successfully in unstructured open-world environments, autonomous intelligent agents need to solve many different tasks and learn new tasks quickly. Reinforcement learning has enabled artificial agents to solve complex tasks both in simulation and real-world. However, it requires collecting large amounts of experience in the environment, and the agent learns only that particular task, much like a student memorizing a lecture without understanding. Self-supervised reinforcement learning has emerged as an alternative, where the agent only follows an intrinsic objective that is independent of any individual task, analogously to unsupervised representation learning. After experimenting with the environment without supervision, the agent builds an understanding of the environment, which enables it to adapt to specific downstream tasks more efficiently.
In this post, we explain our recent publication that develops Plan2Explore. While many recent papers on self-supervised reinforcement learning have focused on model-free agents that can only capture knowledge by remembering behaviors practiced during self-supervision, our agent learns an internal world model that lets it extrapolate beyond memorized facts by predicting what will happen as a consequence of different potential actions. The world model captures general knowledge, allowing Plan2Explore to quickly solve new tasks through planning in its own imagination. In contrast to the model-free prior work, the world model further enables the agent to explore what it expects to be novel, rather than repeating what it found novel in the past. Plan2Explore obtains state-of-the-art zero-shot and few-shot performance on continuous control benchmarks with high-dimensional input images. To make it easy to experiment with our agent, we are open-sourcing the complete source code.
At a high level, Plan2Explore works by training a world model, exploring to maximize the information gain for the world model, and using the world model at test time to solve new tasks (see figure above). Thanks to effective exploration, the learned world model is general and captures information that can be used to solve multiple new tasks with no or few additional environment interactions. We discuss each part of the Plan2Explore algorithm individually below. We assume a basic understanding of reinforcement learning in this post.
Plan2Explore learns a world model that predicts future outcomes given past observations $o_{1:t}$ and actions $a_{1:t}$. To handle high-dimensional image observations, we encode them into lower-dimensional features $h$ and use an RSSM model that predicts forward in a compact latent state-space $s$. The latent state aggregates information from past observations and is trained for future prediction, using a variational objective that reconstructs future observations. Since the latent state learns to represent the observations, during planning we can predict entirely in the latent state without decoding the images themselves. The figure below shows our latent prediction architecture.
To learn an accurate and general world model we need an exploration strategy that collects new and informative data. To achieve this, Plan2Explore uses a novelty metric derived from the model itself. The novelty metric measures the expected information gained about the environment upon observing the new data. As the figure below shows, this is approximated by the disagreement of an ensemble of $K$ latent models. Intuitively, large latent disagreement reflects high model uncertainty, and obtaining the data point would reduce this uncertainty. By maximizing latent disagreement, Plan2Explore selects actions that lead to the largest information gain, therefore improving the model as quickly as possible.
To effectively maximize novelty, we need to know which parts of the environment are still unexplored. Most prior work on self-supervised exploration used model-free methods that reinforce past behavior that resulted in novel experience. This makes these methods slow to explore: since they can only repeat exploration behavior that was successful in the past, they are unlikely to stumble onto something novel. In contrast, Plan2Explore plans for expected novelty by measuring model uncertainty of imagined future outcomes. By seeking trajectories that have the highest uncertainty, Plan2Explore explores exactly the parts of the environments that were previously unknown.
To choose actions $a$ that optimize the exploration objective, Plan2Explore leverages the learned world model as shown in the figure below. The actions are selected to maximize the expected novelty of the entire future sequence $s_{t:T}$, using imaginary rollouts of the world model to estimate the novelty. To solve this optimization problem, we use the Dreamer agent, which learns a policy $\pi_\phi$ using a value function and analytic gradients through the model. The policy is learned completely inside the imagination of the world model. During exploration, this imagination training ensures that our exploration policy is always up-to-date with the current world model and collects data that are still novel. The figure below shows the imagination training process.
We evaluate Plan2Explore on the DeepMind Control Suite, which features 20 tasks requiring different control skills, such as locomotion, balancing, and simple object manipulation. The agent only has access to image observations and no proprioceptive information. Instead of random exploration, which fails to take the agent far from the initial position, Plan2Explore leads to diverse movement strategies like jumping, running, and flipping, as shown in the figure below. Later, we will see that these are effective practice episodes that enable the agent to quickly learn to solve various continuous control tasks.
Once an accurate and general world model is learned, we test Plan2Explore on previously unseen tasks. Given a task specified with a reward function, we use the model to optimize a policy for that task. Similar to our exploration procedure, we optimize a new value function and a new policy head for the downstream task. This optimization uses only predictions imagined by the model, enabling Plan2Explore to solve new downstream tasks in a zero-shot manner without any additional interaction with the world.
The following plot shows the performance of Plan2Explore on tasks from DM Control Suite. Before 1 million environment steps, the agent doesn’t know the task and simply explores. The agent solves the task as soon as it is provided at 1 million steps, and keeps improving fast in a few-shot regime after that.
Plan2Explore (—) is able to solve most of the tasks we benchmarked. Since prior work on self-supervised reinforcement learning used model-free agents that are not able to adapt in a zero-shot manner (ICM, —), or did not use image observations, we compare by adapting this prior work to our model-based Plan2Explore setup. Our latent disagreement objective outperforms other previously proposed objectives. More interestingly, the final performance of Plan2Explore is comparable to the state-of-the-art oracle agent that requires task rewards throughout training (—). In our paper, we further report performance of Plan2Explore in the zero-shot setting where the agent needs to solve the task before any task-oriented practice.
Plan2Explore demonstrates that effective behavior can be learned through self-supervised exploration only. This opens multiple avenues for future research:
First, to apply self-supervised RL to a variety of settings, future work will investigate different ways of specifying the task and deriving behavior from the world model. For example, the task could be specified with a demonstration, description of the desired goal state, or communicated to the agent in natural language.
Second, while Plan2Explore is completely self-supervised, in many cases a weak supervision signal is available, such as in hard exploration games, human-in-the-loop learning, or real life. In such a semi-supervised setting, it is interesting to investigate how weak supervision can be used to steer exploration towards the relevant parts of the environment.
Finally, Plan2Explore has the potential to improve the data efficiency of real-world robotic systems, where exploration is costly and time-consuming, and the final task is often unknown in advance.
By designing a scalable way of planning to explore in unstructured environments with visual observations, Plan2Explore provides an important step toward self-supervised intelligent machines.
We would like to thank Georgios Georgakis and the editors of CMU and BAIR blogs for the useful feedback.
This post is based on the following paper:
Spotify updates its iOS and Android apps to let users search for songs by lyrics, a feature Apple Music has had since 2018 — Spotify has rolled out a useful new feature today for iOS and Android that allows users to search for songs by its lyrics, something that Apple Music users have enjoyed for a couple of years.
Kaspersky researchers spot malware embedded in UEFI firmware on motherboards of victims' devices, affecting diplomats working on issues related to North Korea — The tool attacks a device's UEFI firmware—which makes it especially hard to detect and destroy.
Enlarge / Computer hacker character stealing money online. Vector flat cartoon illustration (credit: GettyImages)
If you’re in the market for a new mobile phone plan, it’s best to avoid turning to Boom! Mobile. That is, unless you don’t mind your sensitive payment card data being sent to criminals in an attack that remained ongoing in the last few hours.
According to researchers from security firm Malwarebytes, Boom! Mobile’s boom.us website is infected with a malicious script that skims payment card data and sends it to a server under the control of a criminal group researchers have dubbed Fullz House. The malicious script is called by a single line that comprises mostly nonsense characters when viewed with the human eye.
When decoded from Base64 format, the line translates to: paypal-debit[.]com/cdn/ga.js. The JavaScript code ga.js masquerades as a Google Analytics script at one of the many fraudulent domains operated by Fullz House members.
Read 5 remaining paragraphs | Comments
Apple removed headphones and speakers from Bose, Logitech, and Sonos from its online store at the end of Sept., asked retail employees to do the same — - Apple is working on first over-ear headphones, smaller HomePod — Products from Sonos, Bose, Logitech pulled from online store
