Twitter users aren’t the only ones getting updates from the micro-blogging social media site. One maker of Android malware is also using Twitter to communicate with infected smartphones, according to security firm ESET.
The company uncovered the feature in a malicious app called Android/Twitoor. It runs as a backdoor virus that can secretly install other malware on a phone.
Typically, the makers of Android malware control their infected smartphones from servers. Commands sent from those servers can create a botnet of compromised phones and tell the malware on all the phones what to do.
The makers of Android/Twitoor decided to use Twitter instead of servers to communicate with the infected phones. The malware routinely checks certain Twitter accounts and reads the encrypted posts to get its operating commands.
Lukas Stefanko, an ESET researcher, said in a blog post that this was an innovative approach. It removes the need to maintain a command and control server, and the communications with the Twitter accounts can be hard to discover.
“It’s extremely easy for the crooks to re-direct communications to another freshly created account,” he said.
ESET said this was first Twitter-controlled Android botnet it had ever found. Windows-based botnets using Twitter have been around since at least 2009.
ESET said Android/Twitoor hasn’t been detected in any app stores, so it probably spreads through malicious links sent to the victim. The malware pretends to be a porn player or multimedia messaging app, and it’s only been active for about a month.
So far, Android/Twitoor has been found downloading versions of mobile banking malware to users’ phones.
“In the future, we can expect that the bad guys will try to make use of Facebook statuses or deploy LinkedIn and other social networks,” Stefanko added.
It just became easier for HipChat customers to see one another whenever they want it. The company launched new group video calling and screen sharing functionality that lets up to 10 other people share a virtual face-to-face meeting.
Users can spin up a call in a HipChat channel, or bring additional people into a one-on-one video call. That way, people who work in far-flung teams can get onto the same page face-to-face, using the same software that they count on for text chat during the day.
HipChat’s announcement Thursday is a move to compete with both consumer services like Skype and Google Hangouts, as well as workplace videoconferencing systems like Lifesize and Skype for Business. The launch is particularly important for HipChat’s competition with Slack, which recently added group voice calls and has video calling on its roadmap.
Group video calls are only available for teams that pay for HipChat Plus, which costs $2 per user per month.
The new video calling features are based on technology HipChat vendor Atlassian acquired with the JitSi open source video-conferencing product. The company still makes the open source version available, but this integration brings video calling into HipChat natively.
Right now, group video calling is only available on HipChat’s desktop apps, but it will make its way to mobile in some form in the future.
It will be interesting to see how quickly Slack can answer with video calling features of its own, after the high-flying productivity startup acquired screen sharing company Screenhero in January 2015.
Some teams may still find themselves in need of dedicated videoconferencing services, if they use specialized hardware for video meetings or if their needs exceed what HipChat can offer. For example, meetings in HipChat can’t have moderators with special privileges, and are limited to 10 participants at launch.
Intel has acquired artificial intelligence (AI) startup Nervana Systems in a bid to future-proof its data centre business and shift focus away from the flailing PC market.
Intel hasn’t revealed the financial details of the deal, but Recode reported that the company paid “more than $400m”, citing an anonymous source.
Nervana, a 48-person firm based in San Diego, California led by co-founder Naveen Rao, a former Qualcomm researcher, was founded in 2014 and offers a fully optimized software and hardware stack for deep learning.
The firm’s cloud-based service allows businesses to build and deploy applications that make use of deep learning, and Nervana has developed a custom processor, known as an ASIC, especially for deep learning.
Intel is looking to the firm to bolster its own deep learning credentials, betting big on the fact that AI represents the next big shift in corporate data centres. The purchase also sees the firm moving away from the PC market, which hasn’t been going too well for Intel lately.
Diane Bryant, executive vice president and general manager of Intel’s Data Centre Group, said: “I’m excited to announce that Intel signed a definitive agreement to acquire Nervana Systems, a recognized leader in deep learning.
“Their IP and expertise in accelerating deep learning algorithms will expand Intel’s capabilities in the field of AI. We will apply Nervana’s software expertise to further optimise the Intel Math Kernel Library and its integration into industry-standard frameworks.
“Nervana’s engine and silicon expertise will advance Intel’s AI portfolio and enhance the deep learning performance and TCO of Intel Xeon and Xeon Phi processors.”
Rao added: “The semiconductor integrated circuit is one of humanity’s crowning achievements and Intel has the best semiconductor technology in the world.
“Nervana’s AI expertise combined with Intel’s capabilities and huge market reach will allow us to realize our vision and create something truly special.”
Intel’s acquisition of Nervana comes just days after Apple scooped up an AI startup called Turi. The firm handed over £150m for the Seattle-based firm, according to reports.
Google has set an early December deadline for removing most Flash content from its Chrome browser, adding that it will take an interim step next month when it stops rendering Flash-based page analytics.
In a post to a company blog, Anthony LaForge, a technical program manager on the Chrome team, said the browser would refuse to display virtually all Flash content starting with version 55, which is scheduled for release the week of Dec. 5.
Previously, Google had used a broader deadline of this year’s fourth quarter for quashing all Flash content except for that produced by a select list of 10 sites, including Amazon, Facebook and YouTube.
Another anti-Flash change will reach Chrome with version 53, now slated to ship the week of Sept. 5. At that time, Chrome will stop rendering very small Flash elements, which are invisible to users but generate data for Web analytics platforms.
LeForge’s latest deadlines were what will probably be among the closing moves in Chrome’s years-long campaign to eradicate Flash. Like other browser makers — including Apple, Microsoft and Mozilla — Google has championed the elimination of Adobe’s once-dominant media player by arguing that it results in longer laptop battery life, faster page rendering and improved security.
Apple’s Safari has frozen some Flash content since 2013, and will beat Chrome to the no-Flash milestone when it ships Safari 10 with macOS Sierra between now and October: Then, Safari will default to HTML5 and only alert users that a site supports just Flash with a message that they need to download the plug-in. Microsoft’s Edge — Windows 10’s default browser — froze some Flash content in the version bundled with last week’s 1607 upgrade.
Mozilla has only begun to restrict Flash content inside its Mozilla browser. While the open-source developer has said it will require users next year to manually activate the Flash Player plug-in, it has not revealed a timetable for more drastic constraints, like those Google announced.
The report said that the share of attacks from Linux botnets almost doubled (to 70 per cent) – and Linux bots are the most effective tool for the SYN-DDoS attack method. This is the first time that Kaspersky DDoS Intelligence has registered such an imbalance between the activities of Linux- and Windows-based DDoS bots.
SYN DDoS is one of the most common attack scenarios, but the proportion of attacks using the SYN DDoS method increased 1.4 times compared to the previous quarter and accounted for 76 per cent.
Oleg Kupreev, lead malware analyst at Kaspersky Lab said that it is Linux which is to blame.
“Linux servers often contain common vulnerabilities but no protection from a reliable security solution, making them prone to bot infections”, says. “These factors make them a convenient tool for botnet owners. Attacks carried out by Linux-based bots are simple but effective; they can last for weeks, while the owner of the server has no idea it is the source of an attack. Moreover, by using a single server, cybercriminals can carry out an attack equal in strength to hundreds of individual computers. That’s why companies need to be prepared in advance for such a scenario, ensuring reliable protection against DDoS attacks of any complexity and duration”.
Brazil, Italy and Israel all appeared among the leading countries hosting botnet Command and Control (C&C) servers. South Korea is the clear leader in terms of the number of C&C servers located on its territory, with its share amounting to 70 per cent. Brazil, Italy and Israel saw the amount of active C&C servers hosted in these countries nearly triple.
DDoS attacks affected resources in 70 countries over the report period, with targets in China suffering the most (77 per cent of all attacks). Germany and Canada both dropped out of the top 10 rating of most targeted countries, to be replaced by France and the Netherlands.
The report also identifies an increase in the duration of DDoS attacks. While the proportion of attacks that lasted up to four hours fell from 68 per cent in Q1 to 60 percent in Q2, the proportion of longer attacks grew considerably – those lasting 20-49 hours accounted for nine per cent (and those lasting 50-99 hours accounted for four per cent (one per cent in Q1).
The longest DDoS attack in Q2 2016 lasted 291 hours (12 days), an increase on the Q1 maximum of eight days.
The National Institute of Standards and Technology is recommending changes be made. The latest draft of its Digital Authentication Guideline, updated on Monday, warns that SMS messages can be intercepted or redirected, making them vulnerable to hacking.
Many companies, including Twitter, Facebook and Google, as well as banks, already use the phone-based text messaging to add an extra layer of security to user accounts.
It works like this: To access the accounts, the user not only needs the password, but also a secret code sent by the company by text message. Ideally, these one-time passcodes are sent to a designated phone number to ensure no one else will read them.
But even so, hackers have still found ways to trick the system. In the past, they’ve used malware to infect smartphones, and secretly redirect the SMS messages to another device.
Others have chosen to impersonate their victims. This can allow the hacker to call up the phone company and ask them to reroute the SMS text messages to another phone number.
NIST also suggested that phone numbers connected to software-based services, including VoIP, could be vulnerable to hacking, putting the SMS messages at risk of being read.
The microblogging service operator’s shares fell 11 percent in extended trading to $16.40. While Twitter struggles to find a way to boost user growth and win over advertisers, social media services such as Instagram and Snapchat are expanding their footprints.
Co-founder Jack Dorsey returned to the company as chief executive a year ago, but his plan for reviving Twitter is at best seen as unfinished.
The company’s second quarter revenue missed Wall Street estimates and the revenue forecast for the current quarter of $590 million to $610 million was well below the average analyst estimate of $678.18 million.
Twitter’s user base increased about 1 percent to 313 million average monthly active users in the second quarter from 310 million in the first quarter.
“Clearly, the turnaround is still a work in progress and the question of whether being a platform for a mass audience versus a niche audience needs to be answered,” said James Cakmak, analyst at Monness, Crespi, Hardt & Co.
Earlier this year, Twitter laid out a long-term strategy to turn around its business, focusing on five areas: its core service, live-streaming video, the site’s “creators and influencers,” safety and developers.
In live video, the company has signed deals with Major League Baseball and the National Basketball Association to revive user growth and attract more advertising dollars. Executives also said Twitter was investing more in user safety as the company continues to grapple with high-profile instances of abuse and harassment.
Struggling with flat user growth and lower spending by advertisers, Twitter has doubled down on attracting more people and encouraging existing advertisers to spend more as it tries to shape its stagnating business.
“We are a year into Dorsey coming back and there is really no end in sight of when it is going to start picking up to where investors are going to be happy,” said Patrick Moorhead, analyst at Moor Insights & Strategy.
Twitter is also working to better define its role in the social media landscape. This week it rolled out a video ad that showed it as the place to go for live news, updates and discussion about current events, which executives also emphasized on a call with analysts.
About 3.9 billion people, or 53 percent of the population still remains offline at the end of this year, according to the International Telecommunication Union estimates. Even in Europe, the most connected region, 20.9 percent of all people aren’t online. In Africa, the least connected continent, 74.9 percent are offline.
Those figures are part of the annual statistical report from the agency, which is part of the United Nations. The report also showed there’s still a huge divide between rich and poor countries, and a growing gap between men and women, when it comes to internet access. It shows that efforts by companies like Google and Facebook to get all people connected could take a long time.
While more than four out of five people in developed countries use the internet, just over 40 percent of those in developing countries have access. In the ITU’s “least developed countries” — places like Haiti, Yemen, Myanmar and Ethiopia — just 15.2 percent of the people are online.
Also, fewer women than men are on the internet, and that difference is getting worse. The worldwide difference between internet user penetration for males and females is 12.2 percent, up from 11.0 percent in 2013, the ITU says. It’s shrunk significantly in developed countries, from 5.8 percent to just 2.8 percent, but grown in poorer places.
Cost makes it harder to get online in some countries. The ITU says entry-level internet access has become affordable in many developing countries since 2011 but remains unaffordable in most of the poorest countries. By the ITU’s definition, that means internet service costs more than 5 percent of average monthly income.
Aquila, Facebook’s lightweight, high-altitude aircraft, flew at a few thousand feet for 96 minutes in Yuma, Arizona, Chief Executive Mark Zuckerberg wrote in a post on his Facebook page. The company ultimately hopes to have a fleet of Aquilas that can fly for at least three months at a time at 60,000 feet (18,290 meters) and communicate with each other to deliver internet access.
Google parent Alphabet Inc has also poured money into delivering internet access to under served areas through Project Loon, which aims to use a network of high-altitude balloons to made the internet available to remote parts of the world.
Yael Maguire, Facebook’s engineering director and head of its Connectivity Lab, said in an interview that the company initially hoped Aquila would fly for 30 minutes.
“We’re thrilled about what happened with our first flight,” Maguire said. “There are still a lot of technical challenges that need to be addressed for us to achieve the whole mission.” He said he hoped the system might be brought into service “in the near future.”
Zuckerberg laid out the company’s biggest challenges in flying a fleet of Aquilas, including making the plane lighter so it can fly for longer periods, getting it to fly at 60,000 feet and creating communications networks that allow it to rapidly transfer data and accurately beam down lasers to provide internet connections.
Maguire said Aquila will go through several more test flights and hopes it will soon break the world record for the longest solar-powered unmanned aircraft flight, which currently stands at two weeks.
Facebook, which has more than 1.6 billion users, has invested billions of dollars in getting more people online, both through an initiative called internet.org – which offers a pared-down version of the internet to poor areas – and by building drones.
It’s called Stream, and it’s supposed to let people easily work together with one another on videos and then share that content both inside and outside their company.
In the realm of consumer web services, video is ascending. Facebook has been emphasizing video posts on its popular social network, while YouTube is still going strong. Microsoft is trying to take some of that mojo and bring it to the business world with the launch of the open beta for Stream.
Stream allows users to log in to a video portal that lets them see all of the videos that are shared with them, and do things like subscribe to channels, search for subject matter they want to explore, and follow co-workers whose videos they want to see.
People who create videos can upload footage to the service by dragging and dropping files from their computers. Stream will handle the processing and let people add titles, descriptions, and even a caption file so that hearing-impaired viewers can read along with what’s being said.
The service also has the ability to set sharing permissions that can let anyone in an organization view a video, or lock it down to just a small group of people. That way, it’s possible for users to get feedback on a video from a small group before pushing it out to the wider company.
It’s all powered by Azure Media Services, a cloud-based video streaming system that Microsoft has been building up to host a variety of products including public cloud video encoding services used for the Olympics and Skype Meeting Broadcast, a service that lets Skype for Business customers send out a video feed to thousands of viewers.
Microsoft has a smorgasbord of planned features on the roadmap for Stream. IT managers, for example, will have access to greater management controls for the service. Microsoft also plans to add additional intelligence to Stream’s search, and let users of its nPowerApps software build applications that leverage its video viewing and capture capabilities.
Stream is similar to other business apps that Microsoft has recently launched, like Power BI, the company’s data visualization and business intelligence tool, and PowerApps, a service that lets employees build mobile applications that use company data. Like those applications, Stream is a subscription service that lets businesses get a particular capability without buying into one of Microsoft’s big suites.
The Winograd Schema Challenge is a competition intended to reward technologists who can build a system that understands the kind of ambiguous sentences humans come out with all the time, but which are simple for other humans, even stupid ones, to understand.
Get it right 90 per cent of the time and $25,000 is up for grabs. And with things like Apple’s Siri, Microsoft’s Cortana and Google Assistant, the Winograd Schema Challenge must surely be as good as obsolete by now.
The best two entrants at the event this week achieved correct scores only 48 per cent of the time, little better than randomly guessing the meaning of the sentences they were supposed to crack.
This is despite a decade of advances in the field of artificial intelligence (AI), which has barely shifted since the late 1950s, according to some.
The Challenge posed a series of ambiguously worded sentences to the entrants such as:
The trophy would not fit in the brown suitcase because it was too big (small). What was too big (small)?
The town councillors refused to give the demonstrators a permit because they feared (advocated) violence. Who feared (advocated) violence?
There is an ambiguity in the above examples, read literally, about what is too big (or small) and exactly who is fearing violence, although a semi-intelligent human should be able to work it out with ease.
The problem, according to Gary Marcus, a research psychologist at New York University, who acted as an advisor for the Challenge, is that computers lack common sense, and programming it into them is incredibly difficult.
Indeed, the MIT Technology Review said that most of the entrants in the Challenge used a combination of hand-coded grammatical understanding and a ‘knowledge base’ of facts. It still didn’t help much, though.
However, one of the two best-placed systems, led by Quan Liu, a researcher at the University of Science and Technology of China, together with researchers from York University in Montreal and the National Research Council of Canada, used neural network-based machine learning in a bid to train their computer to recognise the many different contexts in which words can be used.
Liu claimed that after fixing a problem in the AI, he was able to achieve a success rate closer to 60 per cent, which is still a long way from being able to go home with a cheque for $25,000.
The Challenge is deliberately designed to be different from the Turing Test, which tests only whether a human can be fooled into thinking that an AI program is human.
The trouble with this is that there are more than enough idiots who could be fooled into helping an AI system to pass that test. The language test, in contrast, provides a more objective test of genuine AI, argued Marcus.
The failure of the AI programs in the Challenge highlights how far chatbots and other supposedly revolutionary AI-based machines still have to go before humans can clock-off for the last time and leave running the planet to computers.
Some experts have claimed that its development will spark the next industrial revolution, while others, such as Apple co-founder and pontificator Steve Wozniak, suggest that we’ll be adopted as pets by robots.
Google, Microsoft and Facebook didn’t bother entering, perhaps because they feared outright humiliation. Maybe next year.
The limited testing on Messenger, which has more than 900 million users, comes three months after Facebook rolled out end-to-end encryption to its more popular WhatsApp, a messaging application with over 1 billion users that it acquired in October 2014.
The move comes amid widespread global debate over the extent to which technology companies should help law enforcement snoop on digital communications.
End-to-end encryption is also offered on Apple Inc’s iMessage platform as well as apps including LINE, Signal, Viber, Telegram and Wickr.
Facebook Messenger uses the same encryption technology as WhatsApp, which uses a protocol known as Signal that was developed by privately held Open Whisper Systems.
“It seems well designed,” said Matthew Green, a Johns Hopkins University cryptologist who helped review an early version of the protocol for Facebook.
While WhatsApp messages are encrypted by default, Facebook Messenger users must turn on the feature to get the extra additional security protection, which scrambles communications so they can only be read on devices at either end of a conversation.
Facebook said that it was requiring users to opt in to encryption because the extra security is not compatible with some widely used Messenger features.
“Many people want Messenger to work when you switch between devices, such as a tablet, desktop computer or phone,” the company said in an announcement on its website. “Secret conversations can only be read on one device and we recognize that experience may not be right for everyone.”
Facebook also said that Messenger users cannot send videos or make payments in encrypted conversations.
The world’s largest social network has announced that its own developers have built a multilingual composer. A user test of the service has begun.
The tool enables users to compose a single post that will appear in multiple languages. Other users will see that post in their preferred language.
“People use Facebook to share information and ideas in many different languages,” the Facebook team wrote in a blog post. “In fact, 50% of our community speaks a language other than English and most people don’t speak each other’s languages, so we’re always thinking about ways we can help remove language as a barrier to connecting on Facebook.”
Anyone in the test group can enable the multilingual composer by going to the Language section of their Account Settings.
The composer, Facebook noted, is only available for desktops now, but others can view the multilingual posts across all platforms.
With the multilingual composer, Facebook execs are aiming to let users connect with a broader group of people around the world.
According to Facebook, while the site is just beginning to test the service with individual users, they began testing it with Pages earlier this year.
The composer actually is being used by about 5,000 Pages today to post nearly 10,000 times per day on average, Facebook reported. Those posts are getting 70 million daily views, with 25 million of those views being seen in a language other than what it was originally posted in.
“This will absolutely help Facebook users connect to more people in more places, more easily,” said Dan Olds, an analyst with the Gabriel Consulting Group. “This new feature will give Facebook posters a much larger addressable audience and will save them quite a bit of time to boot.”
Language, according to Jeff Kagan, an independent industry analyst, continues to be the barrier that separates people the most. This new artificial intelligence-driven tool could help break down that wall.
“This is some of the magic that A.I. brings to the table that can change our world,” Kagan added. “This has always been a tough task, but with A.I., it’s actually getting much easier.”
According to Facebook, engineers used machine translation to change posts into different languages and language identification technology to determine which language individual users need to see posts in.
When creating a new post, users are given the option to have the post written in additional languages. They can specify each language they want the post written in using drop-down selections.
Skype Meetings is free and users can launch meetings for up to 10 people during their first 60 days of using it. After that, they’re limited to only hosting meetings for three or fewer people. Those meetings can take advantage of several features, including the option to bring in participants using a hyperlink and present a PowerPoint slide deck live.
The new service is something of a lure to try and get people hooked on Skype for Business. That’s why Microsoft imposed its user limit restrictions: small businesses can use Skype Meetings without paying, and growing companies are encouraged to buy an Office 365 subscription to hold big meetings.
Microsoft is facing tight competition in that arena, with Google pushing its Hangouts chat and calling software, while Slack is developing calling capabilities for its popular chat app.
Skype Meetings will give users a professional space to coordinate their work with one another and present to people outside their business without having to pay for Office 365. The ability to upload a PowerPoint presentation to a meeting and do things like wave over it with a virtual laser pointer and draw on it with digital ink ought to be of particular interest for people who do a lot of online presentations.
Microsoft says that users can join a meeting from any device that has a microphone, camera, speaker and web browser. It’s not clear if this product is only browser-based, or if it will also work with client applications like Skype or Skype for Business on smartphones and tablets.
DoNotPay is the brainchild of 19-year-old Stanford University student Joshua Browder, and it has already successfully contested some 160,000 parking tickets across London and New York. It’s free to use and has reportedly saved its users some $4 million in less than two years.
“DoNotPay has launched the UK’s first robot lawyer as an experiment,” the site explains. “It can talk to you, generate documents and answer questions. It is just like a real lawyer, but is completely free and doesn’t charge any commission.”
Earlier this week the bot was acknowledged on Twitter by the commissioner of the U.S. Federal Communications Commission.
DoNotPay’s artificially intelligent software uses a chat-like interface to interact with its users. It can also be used to help passengers on delayed airplane flights obtain compensation. Reportedly, Browder plans to extend the service to Seattle next. Meanwhile, he’s also working on helping HIV-positive people understand their rights and on a service for Syrian refugees.
All in all, Browder sees a bigger future for A.I. than the mundane tasks it typically handles today. As he said in a recent tweet, the “value in bots is not to order pizzas.”