Feed aggregator

A good night's sleep is different for everyone, and Sleep Number's smart technology lets you and your partner optimize each side for your own comfort.

• Lee Enterprises filed a new report with the SEC
• It confirmed suffering a ransomware attack and having files encrypted
• As outage drags on, the investigation continues

The recent cyberattack on newspaper firm Lee Enterprises has turned out to be a ransomware attack, after all.

The company confirmed the news in a recent 8-K report filed with the US Securities and Exchange Commission (SEC), sharing more details about the attack, confirming that it was a ransomware strike.

“Preliminary investigations indicate that threat actors unlawfully accessed the company’s network, encrypted critical applications, and exfiltrated certain files,” it was said in the filing. “The company is actively conducting forensic analysis to determine whether sensitive data or personally identifiable information (PII) was compromised. At this time, no conclusive evidence has been identified, but the investigation remains ongoing.”

Advanced evasion techniques

The news comes roughly a week after it filed a 10-Q form with the SEC saying it suffered a cyberattack which forced it to pull parts of its IT infrastructure offline.

“On February 3, 2025, the company experienced a technology outage due to a cyber incident affecting certain business applications, resulting in an operational disruption,” it was said in the filing. “The company is actively investigating the incident, implementing recovery measures, and assessing the potential impact on its operations, financial condition, and internal controls.”

The incident impacted Lee’s operations, including distribution of products, billing, collections, and vendor payments, the company further stressed.

Distribution of print publications across its portfolio of products experienced delays, and online operations are partially limited. It still said that it now distributes all core products in “normal cadence”, although weekly and ancillary products have not yet been restored.

Some of the affected publications include the Winston-Salem Journal, Albany Democrat-Herald, Corvallis Gazette-Times, and others. A full list of affected outlets can be found on this link.

Via TechCrunch

You might also like

• This new keylogger uses Microsoft PowerShell scripts to quietly steal sensitive information — here’s how to protect your data
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

The cost of data breaches is at an all-time high. According to IBM, the global average cost of a security breach currently sits at a staggering $4.9 million, spiking by 10% in 2024. These numbers are only expected to worsen, with USAID forecasting that cybercrime will cost the global economy $24 trillion by 2027.

This upward trajectory is driven in part by the emergence of a new threat: AI-powered malware. While AI has become an integral tool for business workflows and innovation, this technology has also proven particularly beneficial for malicious actors, allowing them to develop more stealthy and sophisticated tactics than ever before. In a world where everything is increasingly connected, it’s critical for businesses to anticipate the incoming wave of new threats and develop surefire strategies to not only mitigate incidents but stop them in their tracks.

AI’s double-edged sword: today’s threat landscape

In recent years, the world witnessed incredible advancements in the power of AI. While AI has helped strengthen cyber resilience, with technologies like automated phishing detection and predictive analysis, it has also become a severe threat to business networks. Stepping up from traditional malware that relies on static attack patterns, AI-powered malware is capable of adapting, evading, and learning, constantly evaluating ways to overcome obstacles and successfully inject itself into protected systems. Building on the knowledge from failed attempts to develop new attack strategies, the ever-changing nature of these threats makes detection increasingly difficult, and in turn, weakens defense mechanisms.

BlackMatter ransomware is an infamous yet prime example of how AI has dramatically changed the cyber threat landscape. As an evolution of DarkSide, BlackMatter has quickly become one of the most formidable ransomware threats, bringing a new level of sophistication to the space. Its AI algorithms are built to refine encryption strategies and analyze victims’ defenses in real time, allowing it to circumvent endpoint detection and response (EDR) tools that would typically neutralize ransomware threats. This evasion tactic ultimately renders conventional defences ineffective.

This dangerous shift in gear poses significant challenges to cybersecurity defenses. To effectively protect against these threats, businesses must first assess the biggest threats that lie ahead.

Autonomous attacks

With self-directed learning for real-time evasion, AI-powered malware can essentially ‘think for itself’, independently altering its behavior to bypass existing cybersecurity measures. In other words, no human input necessary. This leads to more frequent attempts in a shortened time frame, making it harder for defenders to counter new attack vectors, and increasing the likelihood of a successful breach.

Furthermore, AI-enhanced malware can spread across networks or systems without instruction. Once it’s infiltrated one machine, the malware self-propagates and infects other connected systems and devices within minutes.

Intelligent attack methods

AI-powered malware can also carry out more destructive ransomware attacks by identifying the most valuable files and systems to target. Instead of simply encrypting, AI algorithms can identify critical databases, financial records, or intellectual property to maximize disruption and increase the likelihood of a ransom payment.

Moreover, AI-powered malware can apply machine learning to mimic the behaviors of manual systems, making it more difficult for traditional intrusion detection systems to identify. It can even avoid detection by only executing malicious actions during off-peak periods.

Sophisticated targeting

AI can also be used to conduct highly targeted attacks by analyzing large amounts of data, such as social media profiles or network behaviors. This facilitates the weaponization of social engineering tactics, generating personalized phishing emails, which are harder to detect. For instance, an AI-powered phishing email might reference a known contact, a recent purchase, or even mimic the writing style of a trusted colleague. This level of personalization makes it much easier coerce individuals into clicking on malicious links, downloading infected attachments, or revealing sensitive information.

Combatting suspicious activity

Attackers aren’t the only ones that can apply AI to their cyber strategy. In fact, it’s important that defenders follow suit and employ AI-based threat intelligence solutions for effective mitigation. According to IBM, organizations that used AI and automation extensively in attack prevention averaged $2.2 million in cost savings than those that didn’t.

How can AI be implemented? This could manifest in AI-powered anomaly detection software, with continuous monitoring and behavior analysis to flag real-time threats, such as unusually high levels of entropy in software code.

Physical network segmentation

While software-based security measures are a key element of any cybersecurity strategy, data and systems remain vulnerable to AI-powered attacks that exploit vulnerabilities. The constantly self-developing nature of these threats calls for a different approach, which has led to a renewed focus on physical segmentation and isolation as a means of protecting networks.

Moving away from the ‘always on’, interconnected model of today’s systems, physical network segmentation is the practice of isolating different parts of a network using physical hardware, such as routers and switches, enabling organizations to establish secure zones with restricted access to critical systems and data.

In prevention, physically disconnecting digital assets from the internet when they’re not in use will significantly reduce the attack surface and provide a much higher level of protection for sensitive data and systems. This is particularly crucial for critical infrastructure, operational technology, and sensitive research data, which may not require constant internet connectivity.

In crisis, if one segment is compromised, the damage is contained within its isolated network, preventing the rapid spread of malware and cutting off connection before the problem worsens. This segmentation technique acts as a powerful defense-in-depth strategy that makes it notably harder for cyber threats to traverse the entire network and target sensitive areas.

Cyber resilience is in your hands

AI-powered malware presents a myriad of dynamic and unpredictable cyber threats. With its enhanced ability to learn, adapt, and autonomously exploit weaknesses, attackers can use AI to execute more sophisticated, evasive, and damaging attacks. Traditional cybersecurity measures are often ill-equipped to handle these advanced threats.

To truly combat cyber threats, there’s a need for a more tangible solution. By isolating different data sets and physically disconnecting from networks, businesses can rest assured they are optimizing their cyber resilience and mitigating the impact of successful attacks.

We've featured the best firewall software.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

The Toy Story maker's first-ever original series hits streaming this week.

• Canva and HubSpot unveiled a new integration
• The new offering is designed for small businesses and solopreneurs
• It's available globally right now

Canva and HubSpot have just announced a partnership that will see Canva’s design tools integrated into HubSpot’s customer relationship platform.

In a press release shared with TechRadar Pro earlier this week, the two companies said the new integration will give growing businesses, small and medium-sized organizations, and solopreneurs, a creative edge that will level the playing field against large enterprises and other big budget organizations.

The embedded experience, which includes Canva’s complete design offering, as well as Magic Studio AI tools, will be accessible through a native “Design with Canva” button, now available in the HubSpot portal.

The two organizations said the integration is available globally for free across all HubSpot hubs and tiers with a Canva subscription.

How the integration works

As the announcement explains, teams will now be able to add Canva Pro Elements to their designs, comment and share Canva designs with their teams, and find designs created in HubSpot on their Canva homepage. The two organizations claim businesses will be able to create on-brand content “in minutes”, including email templates, social posts, and landing pages.

For example, to add an image to a social post, ad, page, or email, users can simply click the Upload icon and select “Design with Canva”. From there, they can select a format, design the image in the Canva editor, and press Save.

“Visual storytelling is one of the most powerful ways to show, not tell, your story to customers - one image can do the work of a thousand words,” said Karen Ng, Senior Vice President of Product & Partnerships at HubSpot. “By bringing together Canva's creative platform with HubSpot's customer platform, we're making it fast and easy for teams to create professional designs at scale. Whether you're a team of one or hundreds, this integration empowers you to meet the growing content demands of today's market.”

You might also like

• Canva unveils new image generator powered by Leonardo.ai
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

It may sound like hyperbole to say that machine learning operations (MLOps) have become the backbone of our digital future, but it’s actually true. Similar to how we view energy grids or transportation systems as part of the critical infrastructure that powers society, AI/ML software and capabilities is quickly becoming essential technology for a wide range of companies, industries, and citizen services.

With artificial intelligence (AI) and machine learning (ML) rapidly transform industries, we’ve also seen the rise of a new age of “Shadow IT” now referred to as “Shadow ML.” Employees are increasingly deploying AI agents and ML models without the knowledge or approval of IT departments, often circumventing security protocols, data governance policies, and compliance frameworks.

This unchecked proliferation of unauthorized AI tools introduces significant risks, from data leakage to model bias and vulnerabilities that threat actors could exploit. CISOs and IT leaders are now tasked with shining a light into the shadows– ensuring that AI-driven decisions are explainable, secure, and aligned with enterprise policies. Understanding the evolving role of MLOps in managing and securing the rapidly expanding AI/ML IT landscape is essential to safeguarding the interconnected systems that define our era.

Software is critical infrastructure

Software is an omnipresent component of our day-to-day lives, operating quietly but indispensably behind the scenes. For that reason, failures in these systems are often hard to detect, can happen at any moment, and spread quickly across the globe, disrupting businesses, upsetting economies, undermining governments or even endangering lives.

The stakes are even more significant as AI and ML technologies increasingly take center stage when it comes to software development and management. Traditional software operations are giving way to AI-driven systems capable of decision-making, prediction, and automation at unprecedented scale. However, like any technology that ushers in new but immense potential, AI and ML also introduce new complexities and risks, elevating the importance and need for strong MLOps security. As reliance on AI/ML grows, the robustness of MLOps security becomes foundational to fending off evolving cyber threats.

Understanding the risks of the MLOps lifecycle

The lifecycle of building and deploying ML models is filled with both complexity and opportunity. At its core, these processes include:

• Selecting an appropriate ML algorithm, such as a support vector machine (SVM) or decision tree.
• Feeding a dataset into the algorithm to train the model.
• Producing a pre-trained model that can be queried for predictions.
• Registering the pre-trained model in a model registry.
• Deploying the pre-trained model into production by either embedding it in an app or hosting it on an inference server.

It’s a structured approach but one with significant vulnerabilities that threaten stability and security. These vulnerabilities, broadly categorized as inherent and implementation-related, include:

• Inherent Vulnerabilities: The complexity of ML environments, including cloud services and open-source tools, can create security gaps that may be exploited.
• Malicious ML models: Pre-trained models can be weaponized or intentionally crafted to produce biased or harmful outputs, causing trickle-down damage across dependent systems.
• Malicious datasets: Training data can be poisoned to inject subtle yet dangerous behaviors that undermine a model’s integrity and reliability.
• Jupyter “sandbox escapes”: In another example of “Shadow ML,” many data scientists today rely on Jupyter Notebook, which can serve as a path for malicious code execution and unauthorized access when not adequately secured.

Implementation vulnerabilities

• Authentication shortcomings: Poor access controls expose MLOps platforms to unauthorized users, enabling data theft or model tampering.
• Container escape: Containerized environments with improper configuration allow attackers to break isolation and access the host system and other containers.
• MLOps platform immaturity: The rapid pace of innovation in AI/ML often outpaces the development of secure tooling, creating gaps in resilience and reliability.

While AI and ML can offer enormous benefits for organizations, it’s crucial not to prioritize rapid development over security. Doing so could compromise ML models and put organizations at risk. Furthermore, developers must exercise caution when loading models from public repositories, ensuring they validate the source and potential risks associated with the model files. Robust input validation, restricted access, and continuous vulnerability assessments are critical to mitigating risks and ensuring the secure deployment of machine learning solutions.

MLOps hygiene best practices

There are many other vulnerabilities across the MLOps pipeline, underscoring the importance of vigilance among teams. Many separate elements within a model serve as potential attack vectors, which organizations typically manage and secure. Therefore, implementing standard APIs for artifact access and ensuring seamless integration of security tools across various ML platforms for data scientists, machine learning engineers, and core development teams is essential. Key security considerations for MLOps development should include:

• Dependencies and packages: Teams often use open-source frameworks and libraries like TensorFlow and PyTorch. Providing access to these dependencies from trusted sources—rather than directly from the internet—and conducting vulnerability scans to block malicious packages ensures the security of each component within the model.
• Source code: Models are typically developed in languages such as Python, C++, or R. Employing static application security testing (SAST) to scan source code can identify and alleviate errors that may compromise model security.
• Container images: Containers are used to deploy models for training and facilitate their use by other developers or applications. Performing comprehensive scans of container images before deployment helps prevent introducing risks into the operational environment.
• Artifact signing: Signing all new service components early in the MLOps lifecycle and treating them as immutable units throughout different stages ensures that the application remains unchanged as it advances toward release.
• Promotion/release blocking: Automatically rescanning the application or service at each stage of the MLOps pipeline allows for early detection of issues, which in turn helps with swift resolution and maintaining the integrity of the deployment process.

By adhering to these best practices, organizations can effectively safeguard MLOps pipelines and ensure that security measures enhance rather than impede the development and deployment of ML models. As we move further into an AI-driven future, the resilience of the MLOps infrastructure will become an increasingly key component to maintaining the trust, reliability, and security of the digital systems that power the world.

We've featured the best online cybersecurity course.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

South Korea's data protection regulator has accused DeepSeek of sending citizens' data using its AI chatbot to TikTok's parent company ByteDance. This accusation comes a day after ordering a temporary block on new DeepSeek downloads over privacy concerns.

"We confirmed DeepSeek communicating with ByteDance," an official at the country's Personal Information Protection Commission (PIPC) told Yonhap News Agency, adding that it is "yet to confirm what data was transferred and to what extent."

While not directly commenting on such allegations, DeepSeek reportedly acknowledged "shortcomings in considering local protection laws" and expressed the commitment to cooperate with the regulator on the matter.

Chinese tech under increased scrutiny

South Korea may be the first to formally accuse DeepSeek of illegally sharing users' data with the controversial TikTok owner, but it's not the only country going after the new Chinese AI chatbot on privacy grounds.

DeepSeek has sent some serious shockwaves across the globe since January's release, in fact. With 12 million curious people worldwide downloading the new app within 48 hours of its launch, governments have been increasingly looking at DeepSeek's phenomenon with suspicion.

Italy, France, Germany, the US, Taiwan, Australia, and South Korea are only some of the countries targeting DeepSeek AI. Most are accusing the new Chinese AI chatbot of violating the country's data protection laws.

DeepSeek's issues are mainly linked to the fact the chatbot stored all users' data on its Chinese servers and, as per the platform's privacy policy, will be used to "comply with our legal obligations, or as necessary to perform tasks in the public interest, or to protect the vital interests of our users and other people."

Now, South Korea's PIPC allegations are reinforcing growing concerns among Western governments about how their citizens' data is processed by such companies.

While the popular video-sharing app is back in the US app stores, the TikTok ban is set to be enforced again on April 5 unless President Trump can find a solution before that. (Image credit: Shutterstock / Daniel Constante)

Concerns about Americans' data ending up in the Chinese government's hands are one of the reasons behind the push for a TikTok ban in the US.

Outside the US, many countries including the UK and some European nations have also banned TikTok from officials' devices over security concerns.

ByteDance, however, has long dismissed these risks – and it did so also for the new accusation brought toward by South Korea.

As The Independent reported, TikTok's parent company believes its link with DeepSeek is simply because the AI chatbot uses the cloud services provided by Volcano Engine, which is owned by ByteDance. Yet, this "does not mean it has access to DeepSeek data," the paper notes.

South Korea's data regulator isn't the only one suggesting the two Chinese companies are communicating. A recent report published by cybersecurity firm SecuirtyScorecard has also found "integration with ByteDance‘s services" within DeepSeek's code.

What's certain now is that the provider operating Deepseek will have to respond to South Korean officials' clarifications over the matter now that an official investigation into its data handling practices has been launched.

Cambridge Audio Evo One: two-minute review

The Cambridge Audio Evo One is a high-end wireless speaker that's capable of playing audio at up to 32-bit/192kHz. But that’s really only part of the picture: it’s also a multi-room speaker, an internet radio, a soundbar, and a turntable phono stage. It’s almost quicker to list the things it can’t do. The Evo One packs a lot of skills into its substantial frame, making it a natural choice for anyone who wants to cover every base but doesn’t have the space or the cash to splash on buying a whole system of hi-fi separates.

Once the Cambridge Audio Evo One was optimized for my listening space – you can manually adjust its sound for the reflectivity of your room and its distance from the wall – it rewarded me with lush, well-rounded sound. The low end is warm and yet tight – there’s none of the unfortunate swampiness you get with cheaper, bass-heavy speakers. Meanwhile, trebles are well handled, neither succumbing to timidity nor becoming so strident that they grate. This means that rather than facing off aggressive overtures from above and below, the mids are allowed the proper space to express themselves, letting the harmonics of vocals and instruments shine.

The Evo One’s separation is right up there with the best wireless speakers on the market. Listening to Wasting My Young Years by London Grammar was a revelation – the strings are beautifully expressed, feeling vibrant despite their low volume, while the reverb on those gorgeous vocals is given plenty of room to breathe. Its soundstage isn’t quite as impressive though: it offers decent stereo spread without extending far beyond the bounds of the unit, but any single-unit speaker is never going to beat a true pair of the best stereo speakers.

Pitch-perfect streaming is just one of this speaker’s talents though: it offers a wide range of additional sources, from internet radio to TV audio via eARC. To test its mettle when it comes to movie audio, I hooked it up via HDMI and watched Ari Aster’s Midsommar – it expertly handled the swelling drone of its forbidding bass, while giving dialogue sufficient clarity and presence. Using its built-in phono stage, I also hooked up my turntable and span a selection of records – it did an excellent job with the likes of Black Sands by Bonobo and Kingdoms In Colour by Maribou State, expressing both with plenty of dynamism and clarity.

When it comes to design, the Cambridge Audio Evo One is hardly the kind of speaker you’ll want to hide away – its charcoal speaker mesh and smooth walnut surface make it an item of furniture you’ll want to show off, and I certainly felt that it earned pride of place in my living room.

Meanwhile, its front display gives you a stronger connection to the artist you’re listening to, although it’s not quite as bright as I’m used to in this era of ubiquitous 1,000-nit phone screens. However, let's be clear that you probably should think of it as an item furniture – it's big.

With its wide array of connectivity options, polished, poised audio and sophisticated looks, the Cambridge Audio Evo One is a worthwhile purchase, even taking into account its not inconsiderable price. Personally, I’m happy to give the Cambridge Audio Evo One a full-throated endorsement, and I doubt you’ll be disappointed with its performance or capabilities. If you want something that's also wonderful for sound but less expensive, the Bowers & Wilkins Zeppelin Pro is also a highly capable wireless speaker for nearly half the price.

Cambridge Audio Evo One review: price & release date

• Released on August 9, 2024
• List price is $1,499 / £1,299 / AU$2,495

The Cambridge Audio Evo One was released on August 9, 2024. It retails for $1,499 / £1,299 / AU$2,495 for this single unit – that's not exactly the cheapest price you’ll find a wireless speaker going for, but once you scratch the surface of the things this device can do, that asking price becomes much more reasonable. When a speaker is this capable, you really are paying for what you get.

(Image credit: Future) Cambridge Audio Evo One review: specs

(Image credit: Future) Cambridge Audio Evo One review: features

• Multi-talented feature set
• Crammed with well-amplified drivers
• Manual room optimization

Given its bulk, you won’t be surprised to hear that the Cambridge Audio Evo One is absolutely crammed full of drivers. It has 14 in total: four one-inch silk dome tweeters, four 2.25-inch aluminum cone mid-range drivers, and an utterly absurd six 2.75-inch long-throw woofers. Each of those drivers has 50W of amplification behind it, so it’s reasonable to expect that this beast is going to have gargantuan sound.

Such a well-furnished setup makes sense when you see how Cambridge Audio has positioned the Evo One. While it’s nominally a wireless speaker, it has plenty more tricks up its sleeve than just cueing up tunes from your Roon. Designed to sit at the heart of your home audio ecosystem, it really is a renaissance speaker, able to tackle the vast majority of listening tasks you throw at it.

First off, as well as offering Bluetooth 5.1 connectivity, it’s capable of streaming hi-res audio at up to 24-bit, 192kHz from streaming services, offering you significantly better quality than even CDs can muster. You can do this via built-in streaming services like Tidal and Spotify, or using its multi-room functionality to stream over Google Home, Apple AirPlay 2 and Roon. You can also stream radio over the internet, should you fancy letting someone else take over DJ responsibilities, or play direct via USB at the frankly ludicrous high quality of 32-bit/384kHz.

But its talents don’t stop there. Hook it up to your TV using an eARC cable and you can use it as a substitute soundbar – it won’t give you Dolby Atmos or height channels, but it will give your TV’s flimsy inbuilt sound a colossal boost. You can also hook up an external source through its RCA line-in, or plug one of the best turntables into its integrated phono stage and spin your records through it. It really does offer an astounding level of functionality that puts less gifted wireless speakers to shame.

Beyond sources, the Cambridge Audio Evo One offers some excellent functionality to help you polish the sound you’re getting. It offers a seven-band EQ to tailor its audio profile to your tastes – while this isn’t quite as versatile as the nine-band that some speakers offer, realistically it’s going to be plenty granular enough to tweak the sound balance.

Even more helpful are the Evo One’s room-calibration settings, which allow you to adjust its sound profile to account for how reflective your room is and how close the speaker is to a wall. I actually prefer this to the auto-calibration some speakers offer – it really allowed me to use my own judgement and knowledge of my space to get the sound just right, which I’m sure fellow audiophiles will appreciate.

All told, the Cambridge Audio Evo One is an incredibly well-specified speaker, offering a ton of functionality that even similarly high-end wireless speakers aren’t necessarily always capable of. If you’re just looking for a setup that allows you to stream hi-res audio, then maybe that will prove unnecessary window dressing, but if you’re looking for a Swiss Army Knife of a speaker that will sit at the heart of your home audio system, then it’s hard to argue with what’s on offer here.

• Features score: 5/5

(Image credit: Future) Cambridge Audio Evo One review: sound quality

• Once calibrated, it sounds beautifully well balanced
• Excellent separation
• Not the widest soundstage

Hand on heart, I was a bit underwhelmed by the Cambridge Audio Evo One’s sound at first. Listening to Black Eye by Allie X, the saw-edged bass synth sounded full but a little swampy, and there was a slightly echoey vibe to the kick, while a lot of the mix sounded muddier in the middle. That’s when I realised I hadn’t adjusted its room-optimization settings…

And boy, did tweaking those settings make a difference. Lowering the setting for room reflectivity and increasing the slider for how far the speaker was from the wall tamed the boom back and upped the transients. The bassline was tight while still being deliciously warm, the detail of the percussion became much more crisp, and the vocal rose above everything without obscuring anything. Suddenly it was like hearing my favorite songs for the first time, and it rapidly transformed my opinion of what this speaker is capable of.

Once properly calibrated, it’s not just the bass that sings; the rest of the frequency range sounds great too. When I listened to I Want You by Moloko, everything was so tightly integrated in the mids; that tight funky guitar lick and the timbre of the strings rang through clearly, while Róisín Murphy’s voice glued the whole mix together, from the nice warm sub end of the funky bass to the transient bite of the snare hits. And it masterfully handles songs that can be make or break in terms of treble – it deftly handles the crisp harmonics of the strummed guitars and snare and tom hits on You Can Go Your Own Way by Fleetwood Mac, while giving them sufficient body to stop the whole mix feeling aggressively bright.

But the Evo One isn’t just great at balancing different frequencies; it has real poise when it comes to separation as well. I stuck on Claire de Lune by Kamasi Washington and the results were exquisite. Those honey-sweet horns feel so richly textured that you could almost take a bite out of them, while the pluck of the double bass feels like it’s almost it’s in the room with you. Even the piano, oft overshadowed when listening to this mix on weaker speakers, cascades out when appropriate – everything has a palpable sense of space that makes everything feel much more real and evocative.

The Evo One doesn’t just handle digital audio sources well. I hooked up my Audio-Technica AT-LP60XBT turntable to the speaker via its built-in phono stage and cued up Kingdoms In Colour by Maribou State. The overall sound was faithful and gave everything a sizable presence and, while it wasn’t quite as remarkable as its hi-res audio output, I put this down more to my budget deck than the Evo One’s innate capabilities.

One area where this wireless speaker doesn’t quite trounce the competition though is its stereo soundstage. Don’t get me wrong: firing up Daydreaming by Radiohead I could absolutely appreciate how the snatches of Thom Yorke’s vocal echoed to the left and right, while the guttural reversed strings that end the track swam convincingly from side to side. However, unlike some speakers, including the Sonos Era 300, the sound doesn’t extend much beyond the confines of the cabinet, and there’s no attempt to convince you there’s real height here. I’d argue that’s not really its whole vibe though – the Evo One aims for hi-res sound that’s very well reproduced, rather than high-tech trickery.

To test how the Evo One handled TV audio, I wired it up to my TV via ARC and settled in to watch Ari Aster’s hallucinogenic horror Midsommar. The Evo One delivered an appreciable improvement over the TV’s built-in speakers, giving the movie’s ominous soundtrack much more warmth and gravitas; Voice Clarity mode in particular helped the speech in the movie ring clear, while letting the droning bass continue to create suspense. While not a home theater soundbar replacement – you’re not going to get Dolby Atmos height channels or significant width here – it certainly represents an upgrade over TV sound alone.

Ultimately, the sound on offer here is impressive, particularly given how many tools it has in its box. If you’re wanting a premium, single-unit speaker that will cover all the bases, this is likely your best bet at this price range.

• Sound quality: 4.5 / 5

(Image credit: Future) Cambridge Audio Evo One review: design

• Elegant, high-end design
• It’s a big fella
• Display is a little dim

In light of its price and the brand behind it, I’m sure I’m not going to shock anyone by saying that the Cambridge Audio Evo One offers smart, premium design. Everything about it is designed to form a focal point in your home, from its FSC-certified walnut panelling to its charcoal-colored fabric speaker grille. It definitely looks suitably high-end – although an alternate color would have been a great option to include, as beige cloth and a lighter wood such as birch or beech would look amazing alongside airier interiors.

Naturally, with all the drivers, amplification and other sundries inside its frame, this isn’t a small unit – in fact, at 26.6 x 5.1 x 11.4 inches / 675 x 129 x 290mm and 32lbs / 14.5kg, it’s a bit of a beast. But there’s something dependable about its big frame – it makes it feel like a genuine item of furniture that you’re pleased to show off. The only caveat here is if you want to place it in front of your TV to really take advantage of its eARC connectivity: it's no small fry height-wise, and will likely block the bottom of your screen, so you’ll want an entertainment unit that can accommodate both.

The main focal point of its front is the 6.8-inch, 1280 x 480 LCD display. Offering a range of options from track information and artwork to a stereo pair of VU meters, the display offers an attractive way of connecting more closely to the tunes you’re listening to. It’s not a very bright display, though, but expecting the kind of 1,000-nit HDR displays that have become commonplace in phones and tablets was always going to be a stretch. It’s plenty bright enough to carry out its intended function, even if you’re not going to be reveling in sumptuous, vibrant colors.

The Evo One’s remote control is well designed, nicely balancing functionality with simplicity. Standard controls are all idiot-proof – you can play, pause, skip forwards and backwards, raise and lower the volume, and power the unit on and off – but there are also some convenient presets that allow you to fire up your favorite sources, stations or playlists at the press of a button. You can also adjust the display’s brightness and switch between its various screens. You’ll need to reach for your phone to do anything particularly complex, but it's nice to be able to use it like a hi-fi separate when you want some screen-free time.

All in all, the Evo One oozes premium style, without too many unnecessary bells and whistles piled on top. Sure, the display could be brighter, and you’re out of luck if you're not into the dark wood and charcoal upholstery Cambridge Audio has opted for. But make no mistake: this is a luxury speaker that feels like an item of furniture in its own right.

• Design score: 4.5 / 5

(Image credit: Future) Cambridge Audio Evo One review: value

On the one hand, the Cambridge Audio Evo One is unarguably expensive for a single-unit speaker – $1,499 / £1,299 / AU$2,495 is not pocket change for anyone. On the other hand, this isn’t just a single-unit speaker. When you view the Evo One in terms of all of the gadgets it potentially replaces – wireless speaker, streamer, soundbar, phono stage – the value equation shifts significantly in its favor. For the price, it’s incredibly versatile, and is an excellent all-in-one option for someone who finds a whole range of hi-fi separates too extravagant an investment in terms of money and/or space.

And there are few rivals that can beat it for value. For example, while the Naim Mu-so 2nd Gen offers a premium, single-unit speaker for $1,299 / £899 / AU$1,299, it doesn’t have a screen, and nor does it offer eARC connectivity or a phono stage. Conversely, the gorgeous Ruark R410 does offer everything you can get here – wireless streaming, multi-room, Bluetooth, DAB+ radio, eARC input, built-in phono-stage – but it’s also just as pricey, coming in at $1,799 / £1,299 / AU$2,599.

• Value score: 4/5

(Image credit: Future) Should I buy the Cambridge Audio Evo One? Buy it if…

You want a feature-packed single-unit speaker
The Evo One is an impressive all-in-one package, taking the place of four or five home audio separates. So if you want a set-it-and-forget-it sound solution that minimizes the cable chow mein choking up your living room, this is a very good option.

You want lush, confident sound
When properly optimized and fed some hi-res material, the Evo One offers exquisite balance and impressive separation of different elements. It’s a polished and impressive performer that will happily sit as the jewel in the crown of your multi-room speaker setup.

Don’t buy it if…

You actually want a soundbar
Stop the presses: the Evo One isn’t one. It’ll add some much-needed poise and gravitas to soundtracks, but if you’re expecting height channels and a wide soundstage, you really do need to invest in an actual soundbar.

You’re looking to save some cash
No two ways about it: the Evo One is an expensive speaker. At $1,499 / £1,299 / AU$2,495, it’s more comparable to the Ruark R410 or the Bang & Olufsen Beosound A5 – you’re paying a premium price for a premium speaker.

Cambridge Audio Evo One review: also consider

Naim Mu-so (2019)
If you love the form factor and are yearning for superlative-sounding, hi-res streaming without extra features such as a phono stage or eARC, but for a lower price in exchange, the Naim Mu-so 2nd Gen is an excellent choice. At $1,299 / £899 / AU$1,299, it's a powerful and impressive speakers, as our full Naim Mu-so (2019) review explains.

Bowers & Wilkins Zeppelin Pro
Following in the footsteps of the classic airship-shaped iPod dock from the 2000s, the Zeppelin Pro is a powerful wireless speaker capable of rich, hi-res sound. Admittedly it’s less talented than the Evo One – again, there’s no connectivity options like phono in or eARC and it has considerably fewer drivers. But it is much more affordable, coming in at just $799 / £699 / AU$1,349, and still sounds exquisite for the price, as we dig into in our full Bowers & Wilkins Zeppelin Pro review.

How I tested the Cambridge Audio Evo One

• I tested it for a month
• I used it for a wide range of audio, from hi-res streams to TV audio via eARC
• I have decades of experience using and reviewing audio gadgets

I tested the Cambridge Audio Evo One over the course of a month. First of all, I used the TechRadar testing playlist, as well as many of my own standard testing tracks, to see how its sound stacked up. I listened to a wide range of different sources, from hi-res streaming on Tidal to Spotify’s lowest-quality tier to see how it coped with different audio quality.

I also tested all of its additional features. I listened to several vinyl albums, both via line-in and through its built-in phono stage, using my Audio-Technica AT-LP60XBT. I also connected it to my TV via eARC and tested its AV sound by watching multiple TV shows and movies. Finally, I tweaked its various sound optimization settings, and played with the EQ settings to test out how they impacted the sound balance.

I’ve spent many years covering audio gadgets, from headphones to speakers, and use a wide variety of sound setups at home, from Bluetooth speakers to my system of hi-fi separates. On top of this, I’ve spent decades making and producing music, which has given me a great deal of insight into how to assess the balance of frequencies and breadth of soundstage.

• First reviewed: February 2025
• Read more about how we test

• Samsung server has teased the existence of an SM-F761 smartphone
• No device has this model number, and it’s a model number for a foldable
• The likely candidates are the Galaxy Z Flip FE or Flip 8

A new leak seemingly confirms the existence of the long-awaited Samsung Galaxy Z Flip FE, as Samsung itself is seemingly teasing the device on its over-the-air software distribution servers.

The details were shared on social media by tipster Erencan Yılmaz who spotted that Samsung is sending software updates to a device labeled “SM-F761” – a model number that hasn’t been assigned to any official Samsung devices. What’s more, because Samsung’s model numbers follow a pattern, we’re confident this is the Samsung Galaxy Z Flip FE.

They start SM then the string of letters and numbers after the dash is determined by what kind of device it is. Here that string is F761 with F usually indicating it’s a foldable, the 7 specifying it’s a Galaxy Z Flip (rather than a Z Fold), and the 61 indicating which edition it is.

The Z Flip 5 is SM-F731, the Z Flip 6 is the SM-F741, and so the Z Flip 7 should be the SM-F751. The SM-F761 could then be the Z Flip 8. However, it would be odd for Samsung to send out software builds for a device that's likely well over a year from launch given its usual timeline.

That would leave us with this needing to be some kind of new foldable, and the most likely candidate is the much rumored Samsung Galaxy Z Flip FE – though as we usually warn with leaks, while we’re feeling confident about our prediction until Samsung makes anything official we don’t know this for certain.

2025's most interetsing phone launch is coming

(Image credit: Philip Berne / Future)

Beyond teasing the existence of Samsung’s next foldable, the leak doesn’t tell us much else – thankfully other Galaxy Z Flip FE leaks could help us fill in the blanks.

For one it’s expected to sport the same display as the Z Flip 7, and another leak suggested it’ll boast the best Samsung-made chipset – the Exynos 2500 chipset – which should help with any Galaxy AI features it offers. Though leaks have yet to confirm the camera specs which could be where it takes the biggest hit.

Regardless of if it packs weaker cameras, I think the Z Flip FE could be Samsung’s most exciting foldable launch of 2025, and potentially the most exciting phone launch of the year – even compared to Samsung’s tri-fold launch which is expected in 2025 as well.

Even though I am a firm believer that the Fold design is superior to the Flip’s clamshell, the Flip FE should bring something to mainstream foldables we haven’t really seen yet: accessibility. More people getting to experience new and exciting technology will never be a bad thing and it could bring some much-needed enthusiasm for foldables which have struggled to find meaningful mainstream success.

This enthusiasm could in turn lead to more foldable phone designs – with brands like Apple entering the space with its own take on a folding phone, or more brands releasing affordable foldables (of which there aren’t many currently).

So while I expect the device itself will probably be a fine phone, I think it has the potential to ignite a foldable revolution if it gets the features to price trade-off right. We’ll have to wait and see what Samsung is cooking up, but of all of 2025’s phone launches, this is the one I’ll be watching.

You might also like

• The best foldable phones 2024: the top folding smartphones right now
• I switched to a folding phone – and I love it so much that I’m never going back
• Here's how big the two displays on the foldable iPhone could be

Rock your inner jock with a pair of sturdy, sweatproof, and tangle-proof headphones. Here are our favorites.

• Shanling has announced a new portable Hi-Res music player
• It runs on Android 13 with support for major music streaming services
• It has vacuum tube amplification, if you like that sort of thing

If you want to enjoy Hi-Res Audio on the go, but would love to mix up the all-digital music world, then this new music player from Shanling may be one to check out, thanks to its inclusion of vacuum tubes.

The M8T is the new flagship of Shanling’s portable Hi-Res music player product line, featuring two AKM DACs (digital-to-analogue converters) and multiple amplification methods for varied, high-quality listening up to 32bit/768kHz. It retails for a notable $1,499 (around £1,190 / AU$2,360).

It’s technically an update to 2020’s M8 model, designed to play nicely with the latest headphone DACs from AKM, and make the most of Shanling’s technological improvements over the past few years.

It’s slightly larger, and heavier than the M8, measuring 147 x 82 x 22 mm with a 6-inch HD (1080 x 2160) touchscreen, and weighing 383g, with rounded edges and a sleek design available in either black or red.

The Panasonic Z95B is coming this year, and it looks stunning. (Image credit: Shanling)

One big change is the addition of two JAN6418 vacuum tubes, with the M8T featuring both a tube amp and a transistor-based headphone amplifier.

Tube amps are often sought-after for their ‘natural’ sound recreation, louder output per watt, and limited distortion compared to transistor amps – though listeners will be able to opt in or out on the fly, depending on their preferences and what they’re currently playing. It definitely makes the Shanling stand out compared to the wider competition among the best Hi-Res Audio players.

A word of warning – taking tube amps mobile can introduce interference into the sound through a microphonic effect, but if they're well-isolated from the outside world touching them, this can be avoided. It looks like the Shanling has them well-secured, but we'd have to test to be sure.

In a blog post announcing the launch of the M8T, Shanling adds that “And because we understand that tubes might not be ideal for all types of headphones or music, we made sure that the M8T can switch quickly between its tube-enhanced output and the pure transistor mode.”

On-board storage has also been upgraded, doubling to 128GB with a MicroSD card slot that can accommodate a maximum of 2TB, meaning there’s plenty of room for all those massive Hi-Res Audio files.

And an improved 8,350mAh battery allows for 13-15.5 hours of Hi-Res audio output, and up to 96 hours over Bluetooth 5.0.

Tidal, Qobuz, and more

Whereas some Shanling products use its own in-house operating system, MTouch, the M8T runs on Android 13 in order to provide “many years of software updates and problem-free app compatibility” – meaning you can access any of the major music streaming services that support Hi-Res Audio, such as Apple Music, Qobuz, and Tidal.

With Hi-Res Audio expected to come to Spotify this year too, it’s a great time to start investing in Hi-Res audio hardware that really lets you benefit from the increased quality, detail, warmth and chutzpah available for audiophiles today.

If you like the look or the M8T, consider the excellent Shanling HW600, which made our list of the best over-ear headphones as an upmarket pair of wired, open-backed over-ear headphones with some genuine areas of sonic expertise”. Or you can check out this personally tested pairing of a Hi-Res player and headphones, recommended by TechRadar’s Audio Editor Becky Scarrott.

You might also like

• Here's why I ditched Tidal for Spotify after just five days
• Amazon Music Unlimited just got a price hike – giving you no reason to choose it over Spotify or Apple Music
• I changed this one small Spotify feature and it made my music sound dramatically better

• The first two episodes of Dope Thief air on March 14, 2025
• The series focuses on criminal pals from Philly who pose as DEA agents
• Academy Award-winning director Ridley Scott is an executive producer

As a 39-year-old dude with bad knees, I probably shouldn’t be using the following word, but screw it: Dope Thief looks… well, positively dope. The new Apple TV+ crime drama has just been revealed, and despite appearing to rock first class production values and superb cinematography in its debut trailer, that’s not the headline news.

No, the real limelight-grabber is the fact legendary filmmaker Sir Ridley Scott is not only producing the Dope Thief – due to air on March 14 – but the 87-year-old is also directing its first episode. There’s just no slowing down the old boy, is there? After all, this is the mythic Hollywood great behind Alien, Blade Runner and Gladiator, who not so long ago managed to film Gladiator 2 in a frankly absurd 51 days.

Based on Dennis Tafoya’s book, the eight-episode series tells the tale of two criminal chums, played by Brian Tyree Henry and Wagner Moura. What seemingly separates Dope Thief from your standard drama about crooks? These Philly pals aren’t just content to get into firefights and blow shizzle up, they also occasionally pose as DEA agents in order to rip off drug dealers.

Of course, as with many movies and shows involving that one seemingly perfect heist / burglary, which simply can’t go wrong, Henry and Moura’s felons eventually mess up royally. Lured into fleecing an unknown home in the countryside, their robbery quickly goes belly up when they discover they’ve accidentally stumbled upon “the biggest hidden narcotics corridor on the Eastern Seaboard”, according to Apple. As a certain cowardly cartoon doggo would say, “ruh-roh”.

Dope-a-dope

As you can see above, Apple has clearly thrown a whole heap of cash into making Dope Thief one of the best Apple TV+ shows (what a shocker). That trailer is so darkly lit in places, I briefly thought I was watching a David Fincher flick, yet the cinematography should still shine thanks to the involvement of Sir Ridley’s production company, Scott Free.

Dope Thief also has other serious heavyweight talent behind the camera. Apple’s latest big budget drama is created and written by Academy Award Nominee, Peter Craig. The man who’s probably best known for writing Ben Affleck’s masterful 2010 heist caper The Town. Oh, and you’ve not watched The Town, please, please do. It’s a brutal Boston beaut about increasingly risky robberies set in the Charlestown area of Beantown.

The Big A is no doubt hoping Dope Thief is going to go down as one of the best Apple TV Plus shows. With the trailer seemingly blurring the lines between cops and crims, it’s easy to see comparisons to The Wire.

Not that I’m banking on David Simon’s masterful Baltimore drama being outgunned by Dope Thief. After all, The Wire is arguably the best TV show of all time. It’s not just content to focus on cops busting drug corners, over the course of its five seasons, it tackles everything from the death of dockyards, broken school systems, corrupt mayoral races and a certain modern day Robin Hood with a big ol’ shotgun who just loves stealing from dealers. Omar Little, you’re the best.

Once Dope Thief makes its global debut on March 14, you’ll be able to catch new episodes of this potentially breakout drama on Apple TV+ every Friday until it finishes on April 25. If it’s even a third as good as The Wire, we’re all in for quite the treat.

You might also like...

• Apple TV Plus: how to sign up, price, TV shows, movies, devices, and more
• The 50 best shows of 2024: TechRadar's top TV picks of the year
• Severance season 2 is a big Apple TV+ hit – here are 5 reasons why I’m obsessed with it

I've always loved comics. There's something magical about writers and artists bringing a story to life with writing and drawing. It's one reason I've always been skeptical of the idea that AI-generated images could replace human artists, no matter how high quality. So I had my doubts when I opened Pocket Toons, an app full of comics made with the assistance of AI. But, the app exceeded all my expectations for lifeless, slightly eerie characters, dull, repetitive backgrounds, spiced with awkward and sometimes offensive stories.

Pocket Toons was created by Pocket Entertainment, best known for its audiobook platform Pocket FM. The app promises it can revolutionize the comic industry by using AI to pump out episodes 20 times faster and at a fraction of the cost of traditional methods. Pocket Toons uses an AI system called Blaze! to automate backgrounds, set up scenes, color the pages, and perform "quality control" to keep every character on-model.

It supposedly allows human artists to focus on storytelling and character development. That sounds impressive on paper. In reality, the results look like what happens when you ask an AI to "draw a cat," and it spits out a creature with five legs and no soul. Some of it positively burns the eye when you gaze upon it.

(Image credit: Screenshot from Pocket Toons) AI art fail

The art is dull, robotic, and aggressively uninspired. Every character has a few expressions they repeat over and over, much like the copy-pasted backgrounds. Entire chapters of the comics felt like an infinite loop of déjà vu. And let’s not even get into the occasional bizarre hands. Forget low effort; this was no effort. This could only appeal to people who have never seen comics or any drawing.

Artists have been sounding the alarm about AI-generated art for a while now, and honestly, they’re right to be mad. The reason great comic book art succeeds is because actual people put their creative stamp on it. AI, on the other hand, is all about mimicking patterns, which is why Pocket Toons' art feels like a parody of itself. Imagine someone feeding a million pages of manga into an algorithm and sucking out everything that made them enjoyable. It raises questions as basic as "What's going on?"

(Image credit: Screenshot from Pocket Toons) Writing from the asylum

If this were just bad art, that would be one thing. But the "writing," is just as bad if you can even call it that. If the spam links about weird tricks and aging celebrities at the bottom of many websites came to life, they would still have more solid plotting than some of these comics.

I’d love to be able to blame AI for the scripts, too, but as far as I can tell, AI is only responsible for the images. I hesitate to blame the named authors for them, either. They feel like someone blended Wikipedia summaries of popular books and comics with an SEO tool. Every so often, a comic tosses in a moment that almost makes sense, like a boy pining for a girl he likes who ignores him, and you think, “Oh, maybe this one is different.” Then, two panels later, she is assaulted by a gang of goons in front of him, and he just sadly goes home without doing anything about it.

That's bad enough, but I became genuinely angry when another comic opened with a woman discovering that she's pregnant and claiming she's never been intimate with a man despite being married to one for several months. That was strange, but I almost threw my phone across the room when the doctor proceeded to claim that because her health wasn't great, she couldn't have an abortion and would have to give birth. I'm not a doctor, but I checked with a couple just to confirm that this is quite literally the opposite of what a qualified doctor would tell someone with health issues early in a pregnancy. The awful art didn't help make the story better.

I really wonder, why make an app like this? If it was to prove that AI can replace human creativity, it has done the opposite. If it was to flood the internet with thousands of cheap, empty, nonsensical comics for the sake of pure content volume, mission accomplished. If someone told me it was designed to use AI to make people hate comics, art, and stories, I would believe it.

The worst part is that Pocket Toons keeps bragging about how much content it’s churning out. New episodes every day! A faster production pipeline than ever before! It’s like a version of that old joke: The food is terrible! But at least the portions are large!

(Image credit: Screenshot from Pocket Toons) Art attack

Art is not meant to be squeezed out of a machine at high speed like some kind of nightmare assembly line. Comics are supposed to feel alive and vibrant. Even bad comics usually have something valuable within them, even if it's only a single interesting idea or well-drawn panel. Pocket Toons is mostly empty when it's not full of garbage.

I'm not against using AI as a tool in art in any form. AI can assist in tedious tasks, allowing creators to focus more on imagination and craftsmanship. But Pocket Toons doesn't use AI to aid an artist; it just turns out mush in the vague shape of artwork. If aliens came to Earth and were told this is what we chose to do with some of our most sophisticated technology, I'd be hard-pressed to argue against them wiping us out.

The good news is, for now, AI comics aren’t taking over. Pocket Toons is proof that when you strip all the humanity out of the creative process, what you’re left with is nothing of value artistically or commercially. Marvel, DC, independent artists, and anyone who’s ever scribbled a stick figure on a napkin have to stay vigilant against those with more money than artistic sensibility deciding they're equally worthwhile, but I'd be as shocked at these becoming at all popular as the pregnant woman who never learned basic biology.

You might also like

• I replaced my to-do list with ChatGPT’s Tasks feature and it completely changed the way I plan my life
• Perplexity's AI search engine goes deep to get your answers
• I pitted Gemini 2.0 Flash against DeepSeek R1, and you might be surprised by the winner

• Insight Partners confirms being struck by a cyberattack
• The threat actors engaged in "sophisticated" social engineering
• The investigation is ongoing, company says

Global venture capital and private equity firm Insight Partners has disclosed suffering a data breach which resulted in sensitive company data being stolen.

The company published a statement in which it said it spotted, “unauthorized third-party access” on January 16, 2025. The unnamed threat actor apparently accessed certain information systems through a “sophisticated social engineering attack”.

“As soon as this incident was detected, we moved quickly to contain, remediate, and start an investigation within a matter of hours,” the announcement said. Insight added it notified the police and its stakeholders, warning them to be vigilant and to tighten security protocols “irrespective of having shared data compromised."

Investigation underway

The company claims to have successfully ousted the attackers, stating, “There has been no additional disruption to Insight’s operations as a result of the incident.”

At the moment, Insight is investigating the attack and its potential effects, so the details about compromised data are not yet available. The company said it hired third-party cybersecurity experts, a “leading forensic and eDiscovery expert”, and external legal counsel, whose work “will take several weeks.”

Right now, the company does not believe the attack will have a material impact on portfolio companies, Insight funds, or other stakeholders.

Insight Partners specializes in high-growth technology, software, and internet businesses. It currently manages over $90 billion in regulatory assets and has invested in more than 800 companies worldwide, with over 55 portfolio companies achieving an IPO.

Via BleepingComputer

You might also like

• Thousands of servers could be at risk due to major OpenSSH security flaw
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• The new flagship Signature S950 launch this March
• Brazilian wood used to improve sound
• Updated 52mm drivers promise elite audio

Not content unveiling with the stunning anniversary Grado Signature HP100 SE at the end of 2024, Grado is looking to outdo those cans with the Grado Signature S950, which not only feature some impressive speakers, but also some mighty alluring wood.

The most attractive design feature of the Signature S950 is surely their Brazilian walnut wooden shell, which gives these headphones a unique and elegant look. Said shell will also help improve sound quality, Grado claims – the company promises a more consistent listening experience, “thanks to the lack of expansion and contraction due to changes in temperature or humidity helping to maintain the acoustical characteristics." The wonders of wood, eh?

On the sound front, the Signature S950 uses an updated version of Grado’s 52mm S drivers, which were initially designed for its previous flagship model. The Brooklyn-based company set its sights on creating speakers that offered even more high-frequency clarity and detail than the drivers of the HP100 SE, and Grado promises the S version will deliver “a smooth midrange and powerful bass while maintaining minimal distortion”. Translation: your favorite tunes should sound the harmonic business on the Signature S950.

(Image credit: Grado) Making the Grado

Naturally, there are plenty more techy details to entice audiophiles looking for the best wired headphones. The S950’s driver uses a “lightweight copper-plated aluminum coil". Paired with a “high-flux magnetic circuit made from rare earth alloys" – which really does sound like something from Back to the Future – we have high hopes for the dynamism and power of these headphones.

The second set of cans to appear in Grade’s Signature range after the aforementioned HP 100 SE, the S950 headphones weigh 395 grams. The fact that they’re relatively lightweight (for serious, high-end headphones) should ensure they’re easy to wear for hours on end.

(Image credit: Grado)

The comfort factor should further be upped by a revised version of the PH 100 SE’s headband, which now includes 50% more padding. As is custom for Grado’s Signature series, the S950 also ship with detachable cables.

Of course, all this comfort and high-flux magnetic parts come at a price. Want to pick up the Grado Signature S950 when these cans launch in March? Prepare to cough up $2,195 / £2,495 (about AU$4,900). Hey, your audio-obsessed ears deserved nothing but the best right – especially with a mid-century-style walnut flourish.

You might also like...

• The best headphones 2025 tested by our experts, for all budget
• These luxe wireless headphones stick a tube-amp hi-fi on your head, and offer really hardcore Hi-Res Audio support
• The best wireless headphones 2025, chosen by our experts for all budgets

• A new leak shows off The Sony WF-C710N earbuds in huge detail
• Specs suggest a proximity sensor and updates to ANC and battery life
• The earbuds could launch as soon as May 2025

Two years after the Sony WF-C700N were released, and topped our list of the best budget wireless earbuds at the time, the electronics manufacturer looks set to do it again.

The Walkman Blog has obtained leaked photos of the WF-C710N earbuds, alongside a host of specifications and guidance on price and a release window, seemingly provided by “a well trusted source”.

This follows a previous design leak via a Taiwanese certification body in January, but there are plenty more details this time around.

The earbuds are shown in four different colors – black, white, pink, and a pleasingly retro transparent blue that shows the bud’s inner workings – alongside a companion charging case.

The WF-C710N earbuds share its predecessor’s 5mm drivers, IPX4 water resistance, and support for ANC, with a couple of notable changes.

For one, there’s now a dual microphone setup for more advanced noise cancellation, and a proximity sensor hopefully means they'll offer wear detection, which the previous model lacked. Bluetooth 5.3 is also included, up from Bluetooth 5.2 in the last model.

Battery life will be slightly changed too, with the buds reaching 8.5 hours with ANC activated and 12 hours total through the charging case, rather than the 7.5 hours from the buds and 15 hours total through the charging case of the previous model. That's an improvement in one direction, but a step back in the other – perhaps we'll be pleasantly surprised and will find out the total battery life is better than it looks.

The Walkman Blog predicts a late May release window, at an RRP of $110-120, around the same price as the previous model.

Sony's WF-C700N earbuds are some of our favorite budget headphones. (Image credit: Future) Improving on perfection

In our WF-C700N review, TechRadar’s Audio Editor Becky Scarrott wrote that “Sony has taken everything I loved about its more expensive earbuds and squirrelled it into smaller earbuds, at a lower price. You don't get lossless LDAC or the firm's top-tier DSEE upscaler, but for this money, the audio quality cannot currently be beaten.”

While there are better headphones out there, this line of earbuds are pretty much perfect for their near-$100 price, and it’s exciting to see Sony build on that success with a new model. It does seem like an iterative upgrade, rather than a transformative change, but if it keeps the WF-C710N earbuds affordable, we’ve got little to complain about.

Let’s just hope we get some of the more interesting colors of the WF-C700N, like lavender or sage green, down the line.

You might also like

• Spotify will 'double down' on music in 2025, but does that mean Hi-Fi or AI?
• Apple finally gives instructions on how to update AirPods… sort of
• More of the best wireless earbuds in 2025, all real-world tested

Find the fastest internet provider and cheapest internet plans in this Denver suburb, according to our CNET experts.

• Purple Rain will be back in movie theaters for one night only with Dolby Vision and Dolby Atmos.
• You’ll need to make it to a Dolby Cinema on March 5, 2025
• To up the visuals, Dolby performed an 8K scan on the 35mm Original Camera Negative, then color-graded it

Whether you love listening to the sweet, rocking guitar solo of Prince on Purple Rain or are a fan of the epic 1984 film, Dolby and Warner Bros. Discovery offer a downpour of good news.

The film Purple Rain will be re-released with better visuals and audio for a one-night-only event at Dolby Cinemas across the United States and the United Kingdom. On March 5, 2025, Purple will take center stage at select AMC theaters in the US or Odeon theaters in the UK with tickets on sale now. Suffice it to say there will be demand.

The film shown on March 5th at Dolby Cinema theaters will be up to the Dolby Vision and Dolby Atmos standards, which did require some heavy lifting on Dolby’s part. For starters, the film was completely digitally restored by running an 8K scan of the 35mm Original Camera Negative, and then that footage was color-graded for Dolby Vision.

On the audio side, Dolby restored the original stereo tracks at the same time as remixing the 20th anniversary 5.1 multi-channel mix – both from the original master files. The end result is promised to be “a faithfully remastered Dolby Atmos” experience.

Dolby also promises that the film will remain epic rock musical drama but that the restoration process is delivering much-improved visuals – likely with a higher peak brightness and wider contrast range – along with much improved audio quality. I wouldn’t mind hearing the roaring solos during the title track of When Doves Cry, assuming it’s a well-outfitted Dolby Cinema location.

Jed Harmsen, Head of Cinema and Group Entertainment at Dolby calls the re-release a “celebration of Prince’s genius and enduring legacy” and explains that “Devoted and new fans alike prepare to be exhilarated by an unforgettable and unparalleled performance from one of music’s greatest of all time.”

(Image credit: Dolby)

Purple Rain represents Prince’s acting debut, but the rockstar also produced the entire soundtrack, including original tracks. The film itself won a Best Original Song Score Oscar at the Academy Awards and the Grammy for Best Album of Original Score Written for a Motion Picture or Television Special.

If you’re sold and ready to experience the Purple Rain on March 5, tickets are up for presale now at Dolby Cinemas in the United States from AMC and Odeon theaters in the United Kingdom.

You might also like

• I experienced Dolby's reference 34-speaker Dolby Atmos theater – and it made me appreciate the importance of movie theaters
• The best Dolby Atmos soundbars 2025: home theater upgrades for all budgets
• I experienced Dolby Atmos music in a Cadillac OPTIQ, and now I want to drive in an immersive audio bubble

• Trend Micro has spotted Earth Preta dodging antivirus in new attack
• The malware deployment checks to see if ESET antivirus is installed
• Malware hijacks legitimate processes to inject malicious code

A Chinese hacking group tracked as Earth Preta and Mustang Panda has been spotted using the Microsoft Application Virtualization Injector to dodge antivirus software by injecting malicious code into legitimate processes.

New research from Trend Micro’s Threat Hunting team revealed how the group has also been using Setup Factory, a third-party Windows installer builder, to drop and executive malicious payloads.

Earth Preta’s region of focus mostly revolves around the Asia-Pacific region, with the group targeting Taiwan, Vietnam, and Malaysia in recent attacks.

Dodging antivirus software

The attack begins with Earth Preta spear-phishing a victim and depositing a mix of legitimate and malicious files into the ProgramData/session directory using IRSetup.exe. Contained within this mix of files is a legitimate Electronic Arts (EA) app (OriginLegacyCLI.exe) that is used to sideload a modified TONESHELL backdoor, EACore.dll.

While this is happening, a decoy PDF is loaded in the foreground to distract the users from the payload deployment. In the vector studied by the Trend Micro researchers, a PDF asking for the user’s cooperation in listing phone numbers to be added to an anti-crime platform supported by multiple law enforcement agencies was shown to the victim.

In the background, the EACore.dll file is checking to see if two files associated with ESET antivirus are running on the device - ekrn.exe and egui.exe. If either file is detected on the system, EACore.dll executes the DLLRegisterServer function by registering itself with regsevr32.exe.

In order to bypass the antivirus, the malware will then use MAVInject.exe to exploit waitfor.exe in order to inject malicious code into a running process. The waitfor.exe function is used to synchronize processes or trigger a specific action after a signal or command is received, and is therefore typically ignored by antivirus software as it is a legitimate and trusted system process.

If the files associated with ESET are not detected, an exception handler is triggered causing the waitfor.exe to directly inject malicious code using the WriteProcessMemory and CreateRemoteThreadEx APIs. Finally, the malware will establish connection to a threat actor controlled command and control (C2) server.

Due to the attack vector’s similarity to other campaigns observed by Trend Micro, and the observance of the same C2 server in another Earth Preta attack, the researchers attribute this attack to Earth Preta with medium confidence.

You might also like

• These are the best firewalls around
• Take a look at our guide to the best endpoint protection
• US military and defense contractors hit with Infostealer malware

There are some minor areas this gaming desktop could have improved, but overall, it's a comely beast ready for serious gaming.