Bard, Google’s AI collaboration tool, is getting a major update today, adding advanced capabilities and new features to a larger audience, including the ability to create images and more languages.
Perhaps the most exciting of the Bard improvements that were announced today via a blog post, is the brand new skill that puts it in a closer footing to OpenAI’s ChatGPT Plus: the ability to create images based on the user’s prompts. While ChatGPT requires a paid subscription in order to gain access to tools like DALL-E, Bard will now allow you to do the same for free. This feature is initially available in English in most countries and is powered by Google’s updated Imagen 2 model.
The feature is very easy to use: just enter a few words to describe what you want Bard to bring to life, and watch it do its magic. If the initial results received don’t quite match what you had in mind, you can click “Generate more” to get more alternatives and download the ones you like.
Image Source – Google
This is very impressive, especially considering that it comes free of cost. However, it is important to note that is not quiet perfect…yet. In my testing I attempted to have Bard generate an image for me based on the physical description of a person. Bard promptly returned a response stating that it was not yet capable of doing that.
Error message returned by Bard when asked to create an image based on human physical characteristics
Additionally, Google states that in order to stay consistent with the company’s AI principles when it pertains to image generation, any images generated by Bard will use SynthID to embed a digital watermark into the results. This watermark is not detectable to the human eye, but can be scanned for identification.
But that’s not all the smarts that Bard is gaining today: Previously only available in English, the enhanced Gemini Pro in Bard, is now available in over 40 languages and 230 countries and territories. This expansion makes Bard a more inclusive and accessible platform to people globally.
Keeping with the spirit of this new multi-lingual Bard, now even its double-check feature will be available in more than 40 languages. This can be accessed by clicking on the “G” icon that appears below the response to your prompt, and checks the information against what is found on the web.
These upgrades are definite improvements to Bard’s accessibility and capabilities. The free addition of an image generation tool, along with its multi-language support, makes it a more helpful and adaptable tool for creative projects, daily activities, and information verification in a global scale.
The ANY.RUN sandbox has now been updated with support for Linux, further enhancing its ability to provide an isolated and secure environment for malware analysisand threat hunting.
ANY.RUNallows malware analysts, SOC members, and DFIR team members to safely examine Linux-based samples and Windows malware in an interactive cloud environment.
A cloud malware sandbox called ANY.RUN performs the difficult malware analysis work for SOC and DFIR teams.
Linux malware analysis is necessary because Linux is a popular target for hackers, and Linux malware is sophisticated.
Many organizations utilize Linux for their IT infrastructure, meaning there are many files to analyze on Linux systems.
Researchers at IBM have noticed an increase in Linux malware. In 2020, the number of malware families related to Linux increased by 40%.
Compromising Linux-based cloud computing platforms could allow attackers access to massive resources, making the OS an appealing target.
Malware Sandboxing Leader ANY.RUN is a cloud malware sandbox that handles the heavy lifting of phishing and malware analysis for SOC and DFIR teams and also helps 300,000 professionals use the platform to investigate incidents and streamline threat analysis.
Creating A LINUX New Task
You can select Linux from the Operating System drop-down menu when creating a new task.
If you choose it, Ubuntu will run your sample. Ubuntu 22.04.2 will be supported upon launch—all ANY.RUN users, including those on the community plan, can access Linux.
Linux as an option in the Operating system drop-down
The Ubuntu logo is used to identify Linux samples for easy navigation, making it simple to distinguish between Windows and Linux-based tasks in the team’s homepage menu and sidebar quick menu.
DocumentAnalyse Shopisticated Malware with ANY.RUN
More than 300,000 analysts use ANY.RUN is a malware analysis sandbox worldwide. Join the community to conduct in-depth investigations into the top threats and collect detailed reports on their behavior..
Enhancing Linux Malware Analysis with ANY.RUN’s
The interactive analysis power of ANY.RUN is now available on Linux for the first time. With Windows samples, ANY.RUN enables analysts of all levels to enhance threat analysis while using fewer resources and delivering quicker training for entry-level analysts and reverse engineers.
The foundation of ANY.RUN is an interactive analysis that enables analysts to identify undetected threatsmore quickly by changing the analysis’s vector, even in the case of zero-day vulnerabilities. It’s accessible for Linux for the first time.
Process Graph view – clear reports of ANY.RUN
Additionally, it provides real-time alerts to the analyst about suspicious activities, ensuring that no crucial information is ignored.
Users receive concise reports upon task completion, ensuring analysts can access all relevant data and IOCs for additional investigation or incident response.
One simple approach to identifying the kind of family or threat you face is quickly aligning suspicious behaviors that the sandbox recorded in a Linux task with TTPs using ANY.RUN’s MITRE Matrix report.
Operating systems like Linux are, by nature, more secure than Windows. This indicates that the many malware families that can exploit Linux vulnerabilities are complex and challenging to identify.
ANY.RUN provides the easiest way to analyze Linux malware, providing information from the analysis in real-time. Analysts immediately understand the results, allowing them to proceed efficiently without context switching.
Not every security expert has the reverse engineering skill set to swiftly identify the behavior of complex Linux malware and extract the required IOCs.
To overcome this, ANY.RUN offers real-time information obtained through the analysis. Analysts can move forward effectively and without switching contexts because they immediately grasp the results.
ANY.RUN is a cost-effective solution that lowers business expenses by doing away with custom infrastructure requirements.
Because the Linux virtual machines (VMs) are preconfigured to gather IOCs, customers can avoid weeks of infrastructure setup time related to DevOps.
In addition to being a stand-alone research platform, ANY.RUN can also be utilized in conjunction with SIEM/SOAR.
Accurate analysis of malware for Linux is necessary for strong security. Because Linux is so widely used, particularly in cloud hosting, attackers find it a desirable target. Breaching Linux-based systems might provide access to a wealth of resources. As a result, Linux users need to be aware of the growing threats to their devices.
Try all features of ANY.RUN at zero cost for 14 days with a free trial.
Whitehat hackers from Pen Test Partners identified a critical issue in Airbus’ Flysmart+ Manager suite, which was remediated 19 months after the initial disclosure.
Cybersecurity researchers at penetration testing firm Pen Test Partners have been testing the security of various electronic flight bag (EFB), IoT and vehicle applications for several years. Due to their extensive research, a crucial issue was identified in the Flysmart+ Manager suite from Airbus and remediated 19 months after initial disclosure.
NAVBLUE, an Airbus-owned IT services company, developed the Flysmart+ Manager app for iPad, which synchronizes and installs airline data into other apps, including EFBs. According to a report from Pentestpartners, this app has a disabled security control, allowing it to communicate with servers using insecure methods, potentially allowing an attacker to modify aircraft performance data or adjust airport information.
For your information, Flysmart+ is a suite of apps for pilot EFBs. EFBs are crucial for storing critical flight data and information, but they can be exploited to disrupt operations or compromise aircraft systems. Airline EFBs can be exposed to untrusted networks due to known pilot layover hotels, and standard operating procedures may not detect tampering.
Research published on February 1, 2024, revealsthat one of the suite’s iOS apps has intentionally got the App Transport Security (ATS) feature disabled. This issue exposes it to Wi-Fi interception attacks, potentially tampering with engine performance calculations, leading to tailstrike or runway excursion.
The app, Flysmart+, was previously disabled due to a lack of ATS protection, which prevents unencrypted communications. This vulnerability allows attackers to intercept and decrypt sensitive information in transit. Due to disabled ATS, insecure communication occurs, making the app susceptible to interception. An entry in the info.plist file allows insecure HTTP loads to any domain.
Airlines often use the same hotel for layover pilots, allowing attackers to modify aircraft performance data through targeted Wi-Fi networks. That’s because pilots in layover hotels can be easily identified, along with the airline and the suite of EFB apps they will likely use.
This helped Pen Test Partners to access data from NAVBLUE Servers, including SQLite databases containing aircraft information and take-off performance data (PERF), with specific table names.
Researchers downloading the data from data the NAVBLUE Servers (Screenshot: Pen Test Partners)
It is worth noting that database tables are crucial for aircraft performance, including the Minimum Equipment List (MEL) and Standard Instrument Departure (SID). Misunderstandings in MEL and SID can lead to safety issues, such as fuel starvation in the Gimli Glider. Confusion between units like US gallons, imperial gallons, litres, kilograms, and pounds can also cause safety problems.
“We’ve now worked on disclosures with Boeing, Lufthansa, and Airbus We’re really pleased that the vulnerability was successfully closed which is a win for aviation safety and security.”
Antonio Cassidy – Pen Test Partners
The researchers shared the vulnerability report with Airbus on 28 June 2022 and the next day Airbus confirmed the issue. By 25th July 2022, the company had replicated the issue and promised a fix for the next version of Flysmart+ by the end of 2022.
On 22 February 2023, the Airbus VDP team confirmed fixing the issue in the latest version of Flysmart+, and the mitigation measure was communicated to customers on 26th May 2023. The findings were presented at DEF CON 31 in Las Vegas in 2023, as well as at the Aerospace Village and Aviation ISAC in Dublin.
Google Play is rolling out a new tool aimed at enhancing the app experience by encouraging them to update their apps to the latest versions. Google Play aims to accomplish this with the help of specifically designed in-app update prompts. Despite the convenience of auto-updates and in-app updates, some users may still be using outdated or unsupported versions of an app.
The newly introduced Play recovery tools provide developers with the ability to prompt users, running specific versions of an app, to update every time they restart the application. This proactive approach ensures that users are not stuck on older versions, potentially plagued with issues, and brings them closer to the intended app experience.
Developers can use the Google Play console to utilize this feature
To utilize the update prompts, developers can log into the Google Play Console and navigate to the Releases or the App Bundle Explorer page. Here, they can choose the app versions for which they want to trigger the update prompts. Additionally, the Play recovery tools are accessible through the Play Developer API, and an upcoming extension will enable developers to target multiple app versions simultaneously. Note that developers must build the designated version as an app bundle to ensure a seamless and effective update process.
Developers can further refine the targeting criteria for the update prompts by factors such as country or Android version, offering developers flexibility in tailoring their approach. Remarkably, over 50% of users have responded positively to these prompts, noted Google. It leads to a substantial increase in users transitioning to the latest versions of apps. Following the update prompts, developers can leverage Play Console’s recovery tools to modify their update configuration, monitor its progress, or cancel the recovery action if needed.
This initiative by Google Play aligns with the platform’s commitment to ensuring users have access to the most optimized and secure versions of applications. By providing developers with these recovery tools, Google aims to create a more cohesive app experience for users. Google encourages developers to explore and implement this tool to enhance user engagement and satisfaction.
Today, Google is introducing a new way for Google Maps users to find places using generative AI. Google’s large-language models (LLMs) will analyze the detailed information that Maps has for over 250 million places and reviews and recommendations from more than 300 million contributors. With this huge data bank, Google Maps will be able to tell you where to go.
Google’s Local Guides will get first crack at the new feature this week. These are, in the words of Miriam Daniel, VP & GM, Google Maps, “…some of the most active and passionate members of the Maps community. Their insights and valuable feedback will help us shape this feature so we can bring it to everyone over time.”
So how can generative AI help you find places using Google Maps? An example that the company gives has you in San Francisco browsing around the city for some “unique vintage finds.” So you would ask Maps to find “places with a vintage vibe in SF.” Maps will then analyze information about businesses nearby augmented with the use of photos, ratings, and reviews from the Maps community to deliver suggestions that users can trust.
Using AI in Google Maps to find places with a vintage vibe in San Francisco
The results could be divided into different categories such as clothing stores, vinyl shops (what you might know as a ‘Record store’) and flea markets. Review summaries and photo carousels would show you why a specific store might be exactly what you’re looking for. You can follow up your original query by asking “How about lunch?” and Maps will look for dining experiences with a vintage vibe which, according to Google, would include an “old-school diner.”
All of the places recommended by Maps can be put into a list to keep you organized, or to share with friends, or to look at when you return to the city at a future date.
If the weather turns bad, you can quick find appropriate nearby activities
The new Google Maps feature can come in handy if an unexpected change in the weather takes place. If it starts raining, you can ask Maps to show you “activities for a rainy day,” and be directed to nearby movie theaters, bowling alleys, museums, and more with pictures and reviews from contributors who have been to these locations.
Google says, “This experimental capability introduces a whole new way for people to more easily discover places and explore the world with Maps. This is just the beginning of how we’re supercharging Maps with generative AI, and we’re excited to start with our passionate community of Local Guides as we shape the future of Maps together.”
The Internal Revenue Service has announced that the 2024 tax filing season has officially begun, with an expected 146 million individual tax returns to be filed. While it is costly and complex for the IRS to process so many digital and paper documents, it can also be a headache for many Americans.
Unsurprisingly, this is also the time of year where we see an increase in tax-related scams. From unsolicited phone calls claiming you owe money to the IRS to bogus tech support with your accounting software, fraudsters are just one step away from robbing you.
In this blog post, we will focus on the dangers of looking for assistance online and the numerous malicious ads trying to reel you in.
Calling the IRS
People will often go on Google to search for a phone or contact number for a business they are trying to get in touch with. Scammers are very well aware of this and will purchase ads to appear at the top of the search results.
Sometimes, somewhere in the fine prints such advertisers will say they are merely a third-party company providing services. But most people don’t read the fine print, and besides, the tactics used here are generally high deceiving.
A search for ‘IRS support’ returns the following ad claiming to be the “IRS Support Line”:
The website tied to this ad is shown with the corresponding advertiser (left), while a different website from a different advertiser is seen (right). Both have the same template. Which one is the real one? Are any of these identities even real?
A testimonial mentioning the aforementioned website seen in the ad claims that it existed back in 2016 while whois records show the domain name was only created in 2023. To add insult to injury, the same testimonial was also used to promote a different domain. Did the same person prepare their taxes on two different sites in 2016, when neither of them actually existed yet?
Tax software
There are a number of programs to help you file your taxes. As intuitive as they can be, it is quite common for people to have a problem with the software itself or a specific question related to taxes.
Scammers buy ads on search engines that often show before the product’s official website. Those ads are extremely misleading for the average user who’s looking for assistance.
One technique they use to bypass ad validation checks is using special or similar-looking characters in the particular brand they are impersonating.
Victims end up on the phone with someone in a large call center, typically located in a foreign country. While the pitch varies, a common scam consists of selling expensive support plans, such as in this example below which is for the first advert seen earlier:
AI-assisted taxes
AI has been all the rage in the past year or so. We are seeing a number of dubious services advertised online that claim to give quick answers after paying a small fee.
With AI comes a lot of potential for abuse and scammers are already leveraging this tool to automate a lot of interactions that seem human-like. It becomes less human when two different sites with a different expert assistant answer the exact same way.
We did not investigate these services fully, so we won’t be showing their full name. However, please beware that they are highly suspicious.
Avoiding scams
These days it has become increasingly difficult to navigate online without being exposed to a scam. People have become accustomed to trust their search engine and naturally follow the different paths laid in front of them.
While some websites look obviously fake to someone, they may fool someone else. At the same time, the tools to build convincing schemes are readily available to anyone for free.
Before calling a number, ensure that it is legitimate by visiting the official site directly.
Beware of unsolicited phone calls or emails, especially those that ask you to act immediately.
Beware of impersonators which may hide behind sponsored results and instead click on organic search results.
Always check the website you visit by looking at the address bar. If in doubt, close the page and open a new one.
If a website asks you for a small fee upfront it likely is trying to get your credit card to sell you more expensive services.
Use security software that blocks phishing domains and other scam sites. Malwarebytes Premium does this, leaving your computer and financial assets protected.
We don’t just report on threats—we remove them
Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.
Apple managed to ship most phones in 2023, it trumped even Samsung
Also, based on the released info, Apple has the same market share as Samsung in 2023, though Samsung was crushed in the fourth quarter of the year. That information got released too. So let’s take a look at Q4 first, and then we’ll switch over to 2023 as a whole.
Apple shipped out 78.1 million smartphones in Q4 last year, while Samsung managed to push 53.5 million out the door. Their market shares were 24% and 17%, respectively. Apple grew 7% YoY, while Samsung’s shipments declined by 8%.
Xiaomi managed to take third place with 41 million smartphones shipped. The company actually grew 23% YoY in Q4 2023. Transsion took fourth place (the company behind TECNO), with 28.5 million smartphones shipped, and a growth of 65%. Vivo is fifth-placed with 23.9 million sold smartphones. All the other OEMs were holding 30% of the market in Q4 2023.
Both Apple & Samsung have seen a decline compared to 2022
Now, things are a bit different when we look at 2023 as a whole. Apple sits on top with 229.2 million smartphones sold, compared to 225.4 million Samsung managed to rake in. Both companies are holding 20% of the market. Both of them declined compared to 2022, though Samsung more than Apple (13% decline compared to 1%).
Xiaomi is third-placed with 146.4 million smartphones sold, and 13% of the market. The company declined 4% compared to 2022. OPPO is fourth-placed with 100.7 million sales and 9% of the market. OPPO also declined YoY, by 11%.
Transsion is fifth-placed for the entire 2023, as the company sold 92.6 million smartphones, and grabbed 8% of the market. Transsion actually managed to grow by 27% YoY. All the other smartphone OEMs sold 347.9 million smartphones together and grabbed 30% of the market.
Right now, everyone’s a little iffy on the fate of Google Bard. On the same day that we got the news that Google could rename Bard to Gemini, we also got the news that Bard can now generate images. What’s also interesting about this is the fact that it’s free to use for everyone.
Image generation is nothing new, as we’ve already been exposed to generators such as DALL-E, Stable Diffusion, MidJourney, and about 25 million more across the internet. However, Google has been pretty behind the times when it comes to unveiling an image generator. We heard about the image-generating capabilities coming to Gemini Ultra, but that’s still unreleased. Also, it will require a monthly fee to use. So, this is good news for people looking to see Google’s flavor of generative AI.
Google Bard can now generate images.
This shouldn’t come as much of a surprise, as we’ve known about Google’s plans to launch an image generator for well over a year. The company first unveiled Imagen back in May of 2022. Back then, image generation was much less advanced than nowadays with tons of uncanny features, inaccurate lighting, missing reflections, and we can’t forget those messed up hands. Modern image generators still mess up on hands, but they’ve evolved to a point where artists are at risk of losing work.
So, in order to add more fuel to the fire, Google announced that Bard can now generate images. This came as a bit of a surprise. We know that Bard is currently being powered by Gemini Pro, and that model was not announced to have image generation capabilities.
Using this feature
So, one thing to note is that the Bard image generator is available for a limited selection of markets at the moment. There’s a chance that you will not have access to the feature in your area.
To use it, all you have to do is tell Bard to make an image for you and then describe the image. It’s as simple as that. Just like with any other image generator, there’s a limit to what you can generate. While trying to generate an image of a warrior, we were told that Bard could not create it. So, anything dealing with violence, sexual content, child abuse, etc., we’ll be swiftly blocked.
The images generated look really high-quality compared to other image generators. They look realistic with that classic “AI-generated look” to them. Also, the hands look pretty decent.
Bard generates two images at a time. Under the images, you’ll see a Generate more button. This will generate two more images. After that, however, you’ll have to enter the prompt again to get more variations on it.
You can use this on your desktop browser, and you can also use it on your phone’s mobile browser. When you tap on the image, you will have the option to download a full-resolution version of the picture. The images download at a 1536 x 1536 resolution.
A week after rolling out Android 14 QPR2 Beta 3.1 to Pixel devices, Google has released Beta 3.2. The latest build is exclusive to the Pixel 5a. It fixes issues “that caused the device to boot slowly and the user interface to lag or freeze.” There is no new release for other models.
Google releases a new Android 14 QPR2 beta update for the Pixel 5a
QPR2 is the second Android 14-based quarterly platform release, aka feature drop, for Pixels. It is expected to arrive in March (QPR1 rolled out in December 2023). Ahead of that, Google is testing the new features among interested users through a public beta program. It is available for the Pixel 5a, Pixel 6, Pixel 6 Pro, Pixel 6a, Pixel 7, Pixel 7 Pro, Pixel 7a, Pixel Fold, Pixel Tablet, Pixel 8, and Pixel 8 Pro.
The first Android 14 QPR2 beta came in mid-November last year. Google has since released four more beta updates for the aforementioned Pixels (Beta 1.1, Beta 2, Beta 3, and Beta 3.1). The last one came a week ago, fixing several bugs present in the previous builds. The company has now followed up with another beta update specifically for the Pixel 5a to fix a few more bugs.
The Android 14 QPR2 Beta 3.2 for the Pixel 5a comes with the firmware build number AP11.231215.010. The Google Play Services version and the security patch level remain unchanged—23.45.23 and January 2024, respectively. Google says all eligible Pixel 5a units enrolled in the beta program will get the update over the air (OTA). You can also manually check for updates from the Settings app.
Note that this update only fixes the bugs stated above. Other known issues from Beta 3.1 remain unfixed. Those include the Assistant At a Glance widget occasionally displaying the weather information for the wrong location. There may be some unknown issues too. Google may release another beta build before rolling out the stable QPR2 update to the Pixel family.
This might be the penultimate feature update for the Pixel 5a
The Pixel 5a is the oldest phone that Google supports currently. Released in August 2021, it is also the last phone to be powered by a third-party processor (Snapdragon 765G). All newer models use Google’s custom Tensor chips. The company guarantees feature updates for the Pixel 5a until August this year. This essentially means it won’t get Android 15. However, we expect the device to receive two more feature drops, i.e. QPR2 and QPR3. Stay tuned for the stable QPR2 rollout.
Nothing has officially confirmed that the Nothing Phone (2a) is on the way. The company hasn’t revealed the launch date but says the device is “coming” to let you “see the world through fresh eyes.” Rumors suggest it will debut at MWC 2024 in Barcelona at the end of this month. You can already download its official wallpapers in high resolution.
Nothing Phone (2a) will be a mid-range phone
Rumors about the Nothing Phone (2a) first surfaced at the end of November 2023. Over the past couple of months, the device has been subject to numerous leaks, revealing some of its key specs. Based on the available information, it appears to be a mid-range handset, more like a sequel to the Phone (1) than the Phone (2). Akis Evangelidis, co-founder and head of marketing at Nothing, has just confirmed that.
Evangelidis briefly talked about the Nothing Phone (2a) during the company’s first Nothing Community Update of 2024. He said last year’s Phone (2) was aimed at the flagship segment. While the device delivered to the firm’s expectations, its asking price may have been a bit too high for some people. With the Phone (2a), it aims to strike a balance of performance and affordability.
According to Evangelidis, the upcoming handset may not give you the latest and greatest specs but you will get improved “core smartphone functionalities” along with Nothing’s “unique and innovative hardware and software.” He compared the Phone (2a) directly with the Phone (1). The Nothing executive said the new model will double down on the core user needs such as performance and camera.
“It delivers on a clear upgrade compared to the Phone (1) on every single front,” Evangelidis assured. At the same time, you will also get some of the most loved features of the Phone (2). All of this hints at a premium mid-range smartphone with no compromise on Nothing’s specialties, i.e. the Glyph Interface, transparent back, and design. We should learn more about it soon.
Leaked specs and wallpapers
While Nothing hasn’t officially announced the specs of the Phone (2a), leaks have revealed that it will feature a 6.7-inch AMOLED display with an FHD+ resolution (1084×2412 pixels) and a 120Hz refresh rate. MediaTek’s Dimensity 7200 chipset will reportedly power the device, paired with 8GB/12GB RAM and 128GB/256GB storage. Its battery capacity is not known but the phone is said to boast 45W charging speed.
Nothing reportedly doesn’t plan to change the camera hardware from the previous two models. The Phone (2a) might get the same 50MP+50MP dual camera setup at the back and a 32MP selfie camera. The handset will run Android 14-based Nothing OS 2.5 out of the box and could get three years of Android OS upgrades. The device could come in White and Black colors with a price tag of around $400.
With Nothig officially starting to tease the Phone (2a), it might not be long before the device breaks cover. In the meantime, you can download the stock wallpapers for the phone. The company will ship the device with seven stock wallpapers named Ambra, Azunim White, Nexul, Orbique, Rubrane Black, Ruxe, and Virmar. You can download them all in the original quality from the button at the end of this article (images in the gallery below are compressed).