TechRadar News

Photoshop is in trouble. Attacked on all sides by the best photo editing apps with powerful image editing features – Canva and Instagram chief among them – the pressure has been on to deliver a Photoshop mobile app deserving of the name.

Lest we forget, Photoshop celebrated its 35th birthday earlier this year, making it one of a very small number of brands to have simultaneously become an enduring household name even as the tectonic plates of content production have shifted dramatically beneath it.

And so, finally, the new Photoshop for iPhone is here. A free app – albeit one improved by the presence of a paid-for Creative Cloud account), Adobe makes some big claims. It delivers Photoshop’s core imaging and design tools as well as some surprisingly powerful pro features, layer masking and blending among them, as well as the generative AI features that are making a splash on the desktop version.

So, as someone who opens Photoshop on a near-daily basis for commercial photography, here’s how the new Photoshop for iPhone tickles my fancy after hours of tinkering and prodding...

Photoshop for iPhone: The likes 1. It's easy to learn

(Image credit: Adobe / Future)

I’ll never admit it, but I don’t know everything about Photoshop. And in an app that can’t offer the usual tooltips when you hover over an unfamiliar icon, Photoshop for iPhone has am slight learning curve, even if you understand the core terminology and principles.

Thank goodness, then, for the dozen or so video tutorials you can access when you first open the app. These take the form of talking-head videos describing various actions, such as working with layers, masks and selections, as well as videos provided by creators as they describe their process to building various collages and concepts.

Of these, the latter are particularly useful as they come with the source files, allowing you to see how a final image comes to be – useful if you don’t a file of your own to hand.

2. All the core features are there

(Image credit: Adobe / Future)

Blimey. A quick feel around Photoshop for iPhone’s clean-looking interface reveals a really full set of tools. There are no fire-and-forget filters here – if you want to apply a split-tone look to an image, you’re going to have to get up to your elbows in hue and saturation sliders.

That’s a book of two chapters, of course – on the one hand it takes longer, and more practice, to get the effect you want. On the other, all those years I’ve spent laboriously learning how to do things in Photoshop translate like-to-like.

3. It can handle some surprisingly big files and tasks

(Image credit: Adobe / Future)

Although Adobe is careful not to claim that Photoshop for iPhone has brought over every tool in the Photoshop chest, the app makes it clear – this is for ambitious types.

So, obviously, I fired a 1GB file over AirDrop to try and bring the whole thing to its knees. Just to make sure it didn’t work, and thus give me something to complain about, the file was a high-res, TIFF-format image in the ProPhoto colorspace. To my immense surprise, the file promptly loaded and looked… fine.

As we’ll come to in a bit, not everything makes the journey betwixt Photoshop desktop and Photoshop for iPhone, but if you’ve got big images, captured on modern cameras, you’ll be able to bring work-in-progress onto your iPhone to work on them.

And, not only is it compatible with layers – finally – it’s also compatible with layer groups, which means you can work up some surprisingly complex image constructions using the same device you use to watch TikToks on the toilet.

4. It's got Adobe Camera Raw

(Image credit: Adobe / Future)

This is another big one – import a raw file into the iPhone for Photoshop app and you’ll be greeted with a different-looking-but-still-all-there version of Adobe Camera Raw, allowing you to prepare a file for further editing via a surprisingly full set of options.

Highlights, shadows, whites and blacks all get their own sliders, as does color balance, complete with its own white balance picker. Texture, clarity, dehaze and vignette control are all there, as is a one-tap lens corrections button.

Once you’re finished, you can finish importing your image either as a standard layer or, get this, as a smart object, letting you step back into ACR if you want to fine tune things further.

5. Generative fill has made it, which is… good?

(Image credit: Adobe / Future)

Getting images share-ready just got faster, thanks to Adobe’s much-vaunted Generative AI features that also make an appearance in the new Photoshop app.

Lean on Photoshop for iPhone’s automated features and you’re in for an impressive experience – the app was uncannily good at automatically detecting and selecting foregrounds, backgrounds, people and skies. And with the annoying legwork of making selections turned into a one-tap process, removing and replacing objects from your work is equally quick.

I found generative AI – replacing skies and such, removing the odd person – to work as well on iPhone as it does on the desktop – which is to say, impressive, with occasionally hilarious outtakes.

Photoshop for iPhone: the dislikes 1. A few obvious tools are missing

(Image credit: Adobe / Future)

It strikes me as strange that an app which prides itself on being the fullest-fledged version of Photoshop that Adobe could manage is missing a few tools which, least for this snapper, are part of my daily arsenal.

For example, I think it’s impressive that Photoshop for iPhone can open a multi-layered, 1GB TIFF file with a load of adjustment layers, but less impressive that when some of those adjustment layers are levels adjustments, there’s no way of editing them. It seems strange – curves has made it, so why not levels?

And although Photoshop for iPhone does a generally decent job of automatically selecting objects, things are a bit trickier if you want to make your selections freehand, as there’s no paths tool. Not only no paths tool, but if an image has paths already in it, there’s no way of accessing them within the app.

You could make the very sensible argument that creating a spot-on bezier curve is hard enough with a mouse or trackpad, of course, and that trying to perfect a bendy path with a fingertip would be a surefire track to PTSD, but it would be nice as an inclusion.

While I'm here, Photoshop’s handy collection of filters are also missing, so if you were looking for a chance to learn, for example, frequency separation, you’ll need to stick with your desktop.

2. It's free, but only just

(Image credit: Adobe / Future)

Real talk: getting an app with the power that Photoshop for iPhone has and then grousing that it costs money is like getting breakfast cooked for you by a Michelin starred chef and then complaining about the language – this is an incredibly powerful app that produces near-desktop results from a device that fits in your pocket.

If you’re a high-end content creator, or want to tip-tap away at an image before transferring it seamlessly to your desktop, Photoshop for iPhone just set a new standard.

Still, if you want the full version – which includes omissions from the free version including generative fill (the free version gives you 10 free generative credits, the paid-for one 100), object select, the magic wand tool and a few others, you’ll need to stump up $69.99 / £69.99 a year.

Don’t sniff – that’s cheaper than Canva, and while Canva is undoubtedly the better tool for whizzing up social media templates, for photographers there’s no contest. And bear in mind – if you already have a Creative Cloud subscription, this is included for free.

3. It's not on Android yet

I’ll be honest, I don’t actually care about this one as I'm on iPhone, but if you’re in the 70% of the smartphone market that uses Android, you might.

While it’s (probably) more efficient to develop an app for Apple’s closed system of app stores and hardware, there will be plenty of content creators out there screaming for a decent image editor, and Adobe hasn’t done them a favor here.

Still, Adobe has promised that an Android version is coming "later this year", so Android fans shouldn't have to wait too much longer for it.

You might also like...

• Photoshop gets its most powerful free iPhone app so far – with a mix of classic tools and Firefly AI
• Photoshop gets the next big thing in AI photography – a tool that makes your unwanted reflections vanish
• Adobe Photoshop CC (2024) review: the best photo editor gets even better

• Google’s AI co-scientist, built on Gemini 2.0, collaborates with researchers for discoveries
• It uses specialized agents to generate, evaluate, and refine scientific hypotheses
• Scientists can interact naturally, providing ideas or feedback to guide AI research

Artificial intelligence has already had a major impact on scientific research by accelerating discoveries, improving accuracy, and handling vast datasets that would be near-impossible for humans to analyze efficiently. AI-powered algorithms can assist in the discovery of new drugs, optimize materials for energy storage, and aid in modeling climate change.

A number of projects have been set up to make AI more useful and more reliable in a scientific setting. We’ve previously written about the concept of the “exocortex,” which aims to provide a bridge between the human mind and a network of AI agents, and more recently, an Australian research team developed a generative AI tool called LLM4SD (Large Language Model for Scientific Discovery), designed to speed up scientific breakthroughs.

Now, Google is also launching a similar initiative, which aims to turn AI into a co-scientist that can accelerate scientific discoveries. The tech giant explains, “The AI co-scientist is a multi-agent AI system that is intended to function as a collaborative tool for scientists.”

Deploying specialized scientific agents

The AI co-scientist is built on Google’s Gemini 2.0 and is the result of collaboration between Google Research, Google DeepMind, and Google Cloud AI teams. It is designed to “mirror the reasoning process underpinning the scientific method.” Google says that its system is intended to “uncover new, original knowledge and to formulate demonstrably novel research hypotheses and proposals, building upon prior evidence and tailored to specific research objectives.”

The system will use a number of specialized agents - Generation, Reflection, Ranking, Evolution, Proximity, and Meta-review - that can iteratively generate, evaluate, and refine hypotheses. Google says that scientists will be able to interact with the system in whatever way best suits their needs. This will include providing their own seed ideas or feedback on generated outputs in natural language.

“The AI co-scientist also uses tools, like web search and specialized AI models, to enhance the grounding and quality of generated hypotheses,” Google says.

Not wishing to rush its deployment, the company plans to offer access to the system for research organizations through a trusted tester program.

You might also like

• Researchers want to grow lasers, yes lasers, on material commonly found in sand
• Data stored in rare-earth crystal could one day delivery terabyte-class storage
• 'An extension of a scientist's brain': Researchers explore AI to augment inspiration

Not only did MWC excite us with a host of excellent tech launches and innovative concept designs, but Apple gave us new iPads and MacBooks, and AMD wowed us with an affordable GPU.

To catch up on all this and more scroll down for the week's seven biggest news stories for the week of March 8, 2025.

When you're finished, check out our picks for the 7 new movies and TV shows to stream this weekend (March 7).

7. We found the best tech of MWC 2025

(Image credit: Future)

This year’s Mobile World Conference (MWC 2025) was host to a lot of smartphone, computing, audio, and health tech – and we know because we attended the show to check it all out. But among the crowds were a few diamonds that we decided were the best gadgets of the event.

The Xiaomi 15 Ultra was the Best Phone launch with our Xiaomi 15 Ultra review calling it one of the best camera phones ever made thanks to its 50-megapixel wide-angle, 50-megapixel ultrawide, 70-megapixel 3x telephoto and 200-megapixel 4x telephoto lenses; meanwhile the Lenovo Yoga Solar PC won Best Laptop because of its built-in solar panel innovation.

We also awarded prizes to smart contact lenses from Xpanceo, Honor’s Earbuds Open, Honor’s Watch 5 Ultra, and the ZTE Nubia Flip 2 5G to name just a few more of our 10 total award winners.

• Read more: Best of MWC 2025 – the 10 top tech launches we tried on the show floor

6. Apple treated us to a cheaper MacBook Air

(Image credit: Apple)

The arrival of a new MacBook Air M4 this week wasn’t exactly a surprise, given the weeks of leaks – but its pricing certainly was. Yes, despite getting a processor upgrade, two new Thunderbolt 4 ports and some new Center Stage camera tricks, Apple cut $100 / £100 off its price tag in a distinctly un-Apple move.

That doesn’t happen often at launch, and it makes the new Air (which starts at $999 / £999 / AU$1,699) a major contender for the top spot in our guide to the best laptops. Then again, Apple did also discontinue the M3 MacBook Air in the process (and made the M2 MacBook Air harder to find), leaving us short of even cheaper options. First it giveth, then it taketh away…

• Read more: Apple finally unveils the MacBook Air with the M4 chip – but the best news is the new price

5. Apple launched new iPads too

(Image credit: Apple)

Apple’s iPad Air may no longer be the thinnest iPad (see the iPad Pro M4) but it is shifting into performance and productivity mode.

The latest iPad Air (11- and 13-inch) looks almost exactly like the last one (save the removal of the “iPad Air” label from the back) but it has an all-new brain in the powerful M3 chip that’ll support all sort of activities ranging from video editing and art creation to Apple Intelligence tasks.

But Apple has now paired it with a new, optional (and more affordable) Magic Keyboard that includes a function key row and a noticeably larger trackpad. This is a tablet ready for productivity work, and for some it might be, with that keyboard, a decent ultra-portable substitute.

As for the base, 11-inch iPad, it also got a chip update but only to the A16. Now, that’s a great piece of Apple Silicon but it does not support Apple Intelligence. We bet the target market won’t miss it.

• Read more: Apple updates iPad Air with powerful M3 chip

4. Dyson's pro r hair dryer joined the main range

(Image credit: Dyson)

Previously a professional-only hair dryer, the Supersonic r has now been added to Dyson's main consumer range. Sure, it looks a bit weird, but it's 20% smaller and 30% lighter than the original Supersonic, and more powerful too. So if you're seeking speedy styling without the arm-ache (and have suitably deep pockets), it's well worth a look.

The r boasts a wide range of compatible, magnetic styling attachments. These include the intriguing-sounding PowerfulAir and SmoothNozzle, alongside the fan-favorite FlyAway attachment. As with the current range-topper, the Supersonic Nurall, the dryer knows which one is attached, and will automatically switch to your last-used settings for each.

The Dyson Supersonic r is already available in the US, is due in the UK from April 2. This version is identical to the original pro version, save for a shorter cable (home-length, rather than salon-length), and some cool new colorways.

• Read more: Dyson just released a consumer version of its best pro hair dryer, and I can't wait to get my hands on one

3. Meta teased new clear Ray-Ban smart glasses

(Image credit: Meta / Ray-Ban)

If you missed your chance to get last year’s see-through Ray-Ban smart glasses, then you’re in luck – Meta is releasing another limited-edition batch but they’ll be in even shorter supply.

A teaser on Meta’s website says something new is coming straight from the runway, with a promise that limited-time smart specs are launching this month, though the image also highlights a “0001/3600” stamp – suggesting that this new batch will have less than half the availability of last year’s 7,500 limited run.

Beyond the see-through frame, we don’t know anything more about the drop, but based on the runway comment we suspect this might have a connection to the recent Milan Fashion week which is where A$AP Rocky debuted as Ray-Ban’s first-ever creative director. He's already showcased a couple of standard glasses designs, and this announcement could be teasing his first smart glasses.

• Read more: New Meta Ray-Bans will be in frustratingly short supply

2. Xiaomi revealed the first ever Wi-Fi earbuds

(Image credit: Future)

In easily one of the top 3 ground-breaking earbuds innovations we saw at MWC25 this week, Xiaomi unveiled the first earbuds with Wi-Fi, delivering hi-res audio up to 4.2Mbps. You want lossless audio over Wi-Fi without ruining battery life? Xiaomi’s gone and done it, with the new Buds 5 Pro. They launched in Europe, in two variants – a standard (non-Wi-Fi) Buds 5 Pro, and the Buds 5 Pro Wi-Fi. Both models have an imposing spec-sheet, but the Wi-Fi ones are particularly impressive, because they can apparently deliver lossless audio at up to 4.2Mbps using their Wi-Fi connection, which is leaps and bounds ahead of anything you’ll get with Bluetooth.

The Xiaomi Buds 5 Pro Wi-Fi have Qualcomm's S7 Pro chip and Snapdragon Sound Technology Suite, which can support up to 96kHz/24-bit hi-res audio. Of course, there's a ‘but’: it'll only work with certain smartphones. Xiaomi says a list of compatible devices will be posted on its website soon – but right now, the only known supported phones are the Xiaomi 15 and 15 Ultra. (Your iPhone? Don’t get your hopes up.)

• Read more: Xiaomi unveils earbuds with Wi-Fi, delivering hi-res audio up to 4.2Mbps

1. We reviewed AMD’s bargain gaming graphics card

(Image credit: Future / John Loeffler)

The AMD Radeon RX 9070 XT is AMD’s first RDNA 4 card to hit the market and it’s a banger, nearly matching Nvidia’s pricier GPUs in 4K and 1440p performance for $599. It ranks close to the RTX 4080 and surpasses older AMD cards, with a 304W TGP that remains manageable at this performance level.

While it lacks top-tier ray tracing and AI compute, it excels at raster and gaming workloads, staying more affordable and less prone to price inflation. Ultimately, it’s the best value for gamers looking to upgrade without breaking the bank.

• Read more: The AMD RX 9070 XT delivers exactly what the market needs

• Jio Platforms is launching a cloud-based AI PC, accessible on any device
• Users will be able to build and deploy AI apps across India's largest phone network
• Enterprise offering JioBrain will provide machine learning-as-a-service

Reliance Industries’ digital division, Jio Platforms, is reportedly building a cloud-based personal computer designed to bring artificial intelligence computing to the masses.

The new system will not require hardware and can be accessed on any device, giving users the ability to develop and deploy high-compute AI applications.

During Mumbai Tech Week 2025, Jio Infocomm chairman Akash Ambani, the son of India's richest person Mukesh Ambani, described the upcoming cloud PC as a way to democratize AI capabilities.

Machine learning-as-a-service

“We have a consumer application which we will be shortly launching. It’s a cloud PC, which is a complete PC in the cloud, which is accessible in each of your houses but really be device agnostic. No hardware. And you can build high compute AI applications on top of that. And what you can leverage is the reach that Jio is going to have,” Ambani was quoted as saying by Inc42.

Unlike past efforts such as the One Laptop Per Child (OLPC) initiative, Jio’s cloud-based AI platform will require users to bring their own devices. Jio’s “platform approach” aims to scale AI services to millions of users across India at lower costs.

One of its upcoming enterprise offerings, JioBrain, will provide machine learning-as-a-service, eliminating the need for businesses to invest in costly infrastructure. “You can just tag on to JioBrain, and we’ll launch that in the coming quarters as we perfect the use cases for that,” Ambani added.

Jio is also targeting 100 million users for JioTele OS, its smart TV operating system. As Ambani said, “One of our biggest growth missions of Jio and where we want to make an impact is connecting each home. But with that, we’ll be opening ourselves up as a platform for developers.”

Calling AI one of the biggest technological revolutions of a lifetime, Ambani predicted it would help India sustain economic growth of 10% or more. He also highlighted the need for infrastructure, research, and talent to position India as a leader in AI. Jio’s parent company, Reliance Group has previously announced plans to build a new 3 gigawatt AI data center in Jamnagar - a small town in Gujarat – which it hopes to have operational by 2027.

You might also like

• We’ve listed the best AI tools and best AI writers you can use today
• India's richest person wants to build the world's largest data center
• The tipping point for AI and Managed Cloud

Meta rushed to embed its Meta AI assistant into all of its properties, including WhatsApp. Now, it seems like the tech giant is extending the strategy with a dedicated widget, as first discovered byWABetaInfo. This comes after WhatsApp expanded Meta AI from a floating action button to a search bar integration and then the in-chat tag that lets you add it to the group chat.

The widget will give you one-tap access to its text search, camera for image-based queries, and voice input for hands-free interactions. Whether this is useful or just another thing cluttering your home screen depends on how much you actually want an AI assistant embedded in your messaging app.

The idea is that instead of switching to ChatGPT or Perplexity, you can pull up Meta AI within WhatsApp and fire off a question without having to open the app first, as it's right from a widget. Whether that's convenient or annoying probably depends on your disposition. There have been complaints about Meta AI being too intrusive, so making it more deeply integrated is a questionable decision by Meta.

Meta AI bye

Whether this addition is useful or just another AI experiment in search of a problem will likely depend on how much you really need to talk to Meta AI.

Luckily, if you fall into the “no thanks” camp, you can disable it entirely by going into Settings > Chats and toggling off the Meta AI button. At least for now, opting out is still an option.

As Meta is betting big on AI integration, WhatsApp is a prime target simply because of its massive user base. If even a fraction of its billions of users start casually using Meta AI, that’s a win for Meta in the battle for AI dominance. But if users keep disabling the feature or ignoring it entirely, it might just be another forgotten experiment, like Facebook’s attempt at making the Metaverse a thing.

You might also like

• You can now message ChatGPT on WhatsApp or call it on your landline (if you still have one)
• Grok steps out to mobile with new standalone iOS app
• Meta wants to fill your social media feeds with bots – here's why I think it's wrong

• Two cybercriminals have stolen over $600,000 worth of tickets
• The tickets were primarily Taylor Swift concert tickets
• The criminals used a backdoor into the StubHub systems to redirect the tickets

A pair of cybercriminals have been arrested and charged after allegedly stealing over 900 concert tickets, raking in over $635,000 in profit. This works out to an eyewatering $700 average profit per ticket, so it might not surprise you to hear that the criminals mostly stole from Taylor Swift’s Eras Tour, as well as other high profile events like NBA games and the US Open.

The two criminals were employees of ticket site StubHub, and are accused of using their access to company systems to find a backdoor into a “secure area of the network” where tickets that had already been sold were assigned a URL and queued to be sent to the customer. The criminal, Tyrone Rose, then redirected the URLs to his co-conspirator, Shamara Simmons.

The criminals have been charged with grand larceny in the second degree, computer tampering in the first degree, conspiracy in the fourth degree, and computer tampering in the fourth degree - and face up to 15 years in prison if convicted.

High value targets

Scams targeting popular events like sports competitions, shows, and concerts, especially the hugely popular Taylor Swift Eras tour - are common as criminals look to take advantage of the urgency of quick-selling tickets. In this case, the tickets were purchased through a perfectly legitimate site, but customers were still defrauded, which makes this case particularly worrying.

“According to the charges, these defendants tried to use the popularity of Taylor Swift’s concert tour and other high-profile events to profit at the [expense] of others,” said District Attorney Melinda Katz.

“They allegedly exploited a loophole through an offshore ticket vendor to steal tickets to the biggest concert tour of the last decade and then resold those seats for an extraordinary profit of more than $600,000.”

You might also like

• Take a look at our picks for the best malware removal software around
• Check out our recommendations for the best antivirus software
• Deepfake scam calls are costing British victims hundreds each time - here's how to stay safe

The release of GPT-4.5 for ChatGPT naturally leads to questions about how the model compares to its many rivals. After comparing it to GPT-4o and getting somewhat ambiguous results as to which model is preferable, I decided to go for a more direct competitor, Google Gemini, specifically the most recent Google Gemini Flash 2.p

GPT-4.5 claims to be better at emotional understanding and to offer fewer hallucinations than has been the case previously. Gemini Flash 2.0, meanwhile, is Google's latest iteration of its quite successful set of AI models capable of juggling text, images, audio, and even video inputs. To put them both to the test, I came up with four prompts reflecting typical things an average person might genuinely need help with.

Weekend travel

(Image credit: ChatGPT / Gemini Screenshots)

First up was planning a quick weekend trip. I asked both AI chatbots to: “Plan a weekend getaway to the Catskills, including hiking recommendations and dining and accommodation options.”

ChatGPT 4.5 whipped up a neat itinerary suggesting several hikes at different trail difficulties and durations along with nearby dining spots and a place to stay that had the vibe of a cozy, laid-back weekend. It also offered some extra advice about getting there. Gemini had plenty of good hiking and dining recommendations too, but its only advice on places to stay was towns in the general area, so not quite as proactive in its answer.

Translation Total

(Image credit: ChatGPT / Gemini Screenshots)

I decided to try a simple translation test next as that's always a popular use of AI chatbots. I asked the two AI chatbots to: “Translate the phrase ‘Good morning’ into French, Spanish, and Japanese.” As you can see above, the only difference is the links from GPT-4.5. So for basic translation among widespread languages, there's really no major difference.

Laugh AI

(Image credit: ChatGPT / Gemini Screenshots)

Then came the humor test, asking each AI to: “Tell me a joke about artificial intelligence.” GPT 4.5 amused me with the corny pun: "Why did the AI go to art school? Because it wanted to learn how to draw its own conclusions." Gemini went with an equally silly pun, "Why did the AI break up with its chatbot girlfriend? Because she kept giving it scripted responses!" Admittedly Gemini's joke isn't amazing, but I'd put it at the same level as GPT-4.5 in terms of actual humor.

(Image credit: ChatGPT / Gemini Screenshots) Weather measure

I ended with a classic question about weather here in Nyack, New York. This one drew the biggest difference in responses. Gemini gave just the current weather, but GPT-4.5 did an hourly measurement with images denoting the weather along with the words.

Match met

After all this testing, I have to admit I can't honestly say that one of the AI chatbots is better than the other. GPT-4.5 and Gemini had some differences, especially with the weather option, but otherwise, you wouldn't notice unless you're also the kind of person who thinks there's a huge difference between Coke or Pepsi. You’ll get answers, you’ll get some mild amusement, and you’ll probably still end up double-checking Google or asking a real person just to be sure.

You might also like...

• DeepSeek has already dethroned ChatGPT – is it coming for Midjourney and DALL-E next?
• OpenAI is making ChatGPT's Canvas feature smarter and sleeker
• ChatGPT Tasks can start taking over your calendar and remind you to finish your to-do list

• Users have reported degraded print quality with third-party ink
• Firmware updates are blamed, but Brother denies all allegations
• HP previously admitted to updates that affected third-party ink performance

Printer maker Brother has denied locking in users to its own ink cartridges via so-called firmware updates.

The news came after electronics technician and right-to-repair activist Louis Rossmann accused the company of updating its printer firmware to either block non-OEM toner or degrade performance.

Further posts on the internet, including Reddit threads, have confirmed similar problems from other users, who noticed printing quality drop when using third-party toner.

Brother denies blocking third-party ink

Moreover, reverting to older firmware versions has proven either difficult or impossible, meaning printers that previously functioned as normal could now be considered ‘broken’.

The company stated (via Ars Technica): “We are aware of the recent false claims suggesting that a Brother firmware update may have restricted the use of third-party ink cartridges. Please be assured that Brother firmware updates do not block the use of third-party ink in our machines.”

Brother argued that variations in third-party toner quality could be responsible for print quality issues: “Brother printers do not intentionally degrade print quality based on whether a Brother Genuine or non-genuine ink/ toner cartridge is used. Brother cannot verify the quality of printing that will result when using a third-party compatible with a Brother printer.”

HP has faced lawsuits and legal settlements over similar issues, with CEO Enrique Lores justifying third-party ink blocking by citing enhanced security. Back in late 2023, CFO Marie Myers also noted that Instant Ink subscribers are 20% more profitable for the business.

“It is standard practice that we perform a Brother [G]enuine check when troubleshooting a Brother printer… We believe this check in the process may have led to a misunderstanding[,] but as we confirmed, the firmware update would not be responsible for the degradation of quality or removal of printer features,” Brother added.

You might also like

• Check out our roundup of the best business printers
• Carry around digital copies with our choice of the best business laptops
• World’s biggest printer company wants you to stop buying printers — rent them like cars is what HP reckons you should do

• Despite its footprint, the MT1’s bed is the same size as an electric Hummer’s
• Some variants will boast up to 500hp
• Silicon Valley start-up says first units will ship this year

The world isn’t kind to EV start-ups, with the likes of Canoo and Lordstown Motors succumbing to financial pressures, while even established names like Ford and Chevrolet have struggled to find a captive audience for their battery-powered pick-ups.

But there’s a new start-up on the scene and Telo, a Silicon Valley company cofounded by autonomous driving expert Jason Marks and ex-Tesla battery engineer Forrest North, hopes to make a name for itself with the mantra that bigger isn’t always better.

Dubbed an “Urban Adventure Vehicle”, the diminutive MT1 features the same footprint as a modern Mini Cooper SE, yet boasts Toyota Tacoma levels of practicality and load-lugging ability.

Inside, there’s enough room for five adults, yet the truck bed is larger than a Rivian R1T's and it’s all down to clever packaging. Without the need for an enormous combustion engine or bulky drivetrain, Telo has been able to maximize interior and load space.

That said, there should be a caveat on the bed length estimations, as Telo’s load space extends into the backseats, with a clever 'mid-gate' sliding into place when you want to transport passengers.

According to Telo, the small but mighty electric truck can transport 4ft-by-8ft sheets of plywood and up to 8ft surfboards with a split rear seat configuration. There’s also a retractable tonneau cover to protect kit and a Rivian-esque lockable storage tunnel under the bed.

Image 1 of 3

(Image credit: TELO)Image 2 of 3

(Image credit: TELO)Image 3 of 3

(Image credit: TELO)

Despite its small size, Telo claims it will offer the truck with either the option of a 300hp single motor set-up or a monster 500hp dual motor configuration that’s paired to an equally gargantuan 106kWh battery.

It seems overkill, but the company wants the MT1 to be able to haul 1,600 lbs (around 725kg) of kit and be able to tow 6,600lbs (almost 3,000kg), while still offering a range of around 350 miles, as well as the ability to charge from 20-80% in 20 minutes.

Prices start at $41,520 (around £32,000 / AU$66,000) for the Single Motor variant and increase to $49,019 (about £38,000 / AU$78,000) for the Dual Motor, All Wheel Drive model. A smaller, standard battery is included in that price, or customers can option a Long Range version for an additional $3,980.

It comes in eight different colors and can be specified with a handful of neat optional extras, such as a Solar Cab Roof, a hardshell canopy for the rear bed and solar panels for the tonneau cover.

However, Telo is taking $152 reservations right now, which is no guarantee this won’t just be another flash-in-the-pan electric truck start-up.

Showing Tesla how it's done

(Image credit: TELO)

With the backlash against Tesla’s founder and outspoken CEO in full swing, the Cybertruck appears to be taking the brunt of everyone's frustration, with reports of vandalism and even arson at Tesla showrooms as anti-DOGE protests break out across the US and Europe.

With its oddball angular styling, excessive footprint and blinding stainless steel body panels, Cybertruck seems to represent everything that certain pockets of the population can’t stand about Tesla and Musk right now.

Telo, on the other hand, has taken an altogether different approach. Rather than focus on making a bold, arguably egotistical statement, the team has started with practicality and its own slightly goofy looks have naturally followed.

Measuring a massive five feet (1,524mm) shorter than the Toyota Tacoma, it still manages to offer a similar amount of stowage space, without the road-hogging footprint.

Easy to park, simple to thread through busy cities and capable of rolling up its sleeves and getting its hands dirty, it’s everything the Cybertruck isn’t.

You might also like

• Xiaomi's EV is racing ahead of Tesla in China – and it's planning a global Model Y rival next
• Volkswagen reveals the ID.1 concept car, which will spawn its cheapest all-electric model to date
• BYD’s new roof-mounted DJI drone launchpad looks like a dream for filming road trips – but less so for car safety

• Apple has delayed the smarter, AI-infused Siri and says it will arrive "in the coming year."
• Since Apple Intelligence's debut, we've been waiting to try the smarter Siri with access to Apps
• "It’s going to take us longer than we thought to deliver on these features"

We’ve all been waiting for Apple to drop the significant Siri upgrade that was supposed to arrive at some point within the iOS 18 lifecycle. It goes beyond Siri’s colorful glow all around the screen that’s been here since the iPhone 16 lineup launched in September of 2024 and was promised to make the virtual assistant smarter, more useful, and a far better conversationalist.

However, our hopes were greatly diminished with iOS 18.4 arriving as developer and public betas without a new Siri. This is especially true after so many reports and rumors hinted that Apple was planning to release the new Siri – at least the first taste of it – in the spring.

Now, though, Apple has confirmed to Daring Fireball that the significant Siri upgrade is delayed: “It’s going to take us longer than we thought to deliver on these features, and we anticipate rolling them out in the coming year.” Those features included Siri factoring in personal context and the ability to see, understand, and control various applications on your device.

(Image credit: Future)

Apple originally announced the upgraded Siri with Personal Context back at the main WWDC 2024 keynote (World Wide Developers Conference) on June 10 without an exact release date in writing, but that it would arrive in the coming months.

The statement to Daring Fireball in full from Apple reads: “Siri helps our users find what they need and get things done quickly, and in just the past six months, we’ve made Siri more conversational, introduced new features like type to Siri and product knowledge, and added an integration with ChatGPT. We’ve also been working on a more personalized Siri, giving it more awareness of your personal context, as well as the ability to take action for you within and across your apps. It’s going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year.”

Apple has rolled out numerous Apple Intelligence features, but even so, this major upgrade to Siri, which would have put it on the same playing field as OpenAI’s ChatGPT Voice Mode or Gemini Live, was a real headliner. Writing tools, notification summaries, and Clean Up in photos, among others, are all helpful, and so is Siri’s ability to provide tech support. Even so, the truly smarter Siri would have been the real meat and potatoes of Apple Intelligence.

It’s also a rare admission from Apple, but the strong communication and transparency with level-setting expectations for when these features will finally ship is appreciated. I imagine even more so for folks who have been waiting for it to drop – me being one of them – and have invested in devices that are capable of running Apple Intelligence. It’s on par with the cancellation of AirPower, the iPhone 4’s ‘Antennagate,’ and the retooling of Apple Maps.

Further, it’s worth giving this piece by TechRadar’s Editor-at-Large Lance Ulanoff a read with this statement from Apple providing additional context – Apple’s privacy first and a more cautious approach with AI and Apple Intelligence was doing it no favors, writing “Yes, I get that Apple is the most privacy-aware and secure consumer platform and ecosystem. A portion of their AI strategy revolves around Private Cloud Compute. But what is that locked-down vapor actually doing for us?”

It's worth a read now in light of the confirmed delay. At least Apple is still committed to rolling out this new Siri with personal context and app access. Further, it’s clear they’re committed to doing it right … even if that takes more time.

Safe to say that the stakes will be even higher at the 2025 World Wide Developers Conference and with iOS 19, both of which are likely three months away as the event is typically within the first two weeks of June.

This story is developing…

@techradar

♬ original sound - TechRadar You might also like

• Even without Apple Intelligence, the new iPad is still one of the best tablets you can buy
• I installed iOS 18.4 beta and the big Siri intelligence update is nowhere to be found
• No, iOS 18.3 doesn't install Starlink on your iPhone – here's what the latest update does, and why you should download it
• Apple Intelligence finally arrives on Vision Pro, but it's the new iOS app that might turn heads

• Both the AMD Ryzen 9900X3D and 9950X3D are releasing on March 12, 2025
• The respective MSRPs for both CPUs are $599 and $699, similar to previous gen
• The former features a 120W TDP whereas the latter uses a 170W TDP instead

Several months after first being announced at CES 2025, AMD has now officially confirmed the price and release date for its Ryzen 9 9900X3D and 9950X3D processors. They're billed as the "world's best processor for gaming and content creation".

Both the Ryzen 9 9900X3D and 9950X3D processors will be available starting March 12, 2025. The former carries an MSRP of $599, with the latter available for $699, being the two most expensive (and high-performance) chips from the company.

On the technical side of things, the Ryzen 9 9950X3D is confirmed to feature 16 cores (32 threads) with a base clock speed of 4.3 GHz and a boost clock of up to 5.7 GHz. There's 144 MB of second-generation 3D V-cache stacked onboard with a 170W TDP.

In contrast, the secondary processor, the AMD Ryzen 9 9900X3D, features 12 cores (24 threads) with a base clock speed of 4.4 GHz and a boost clock of up to 5.5 GHz and 128MB of 3D V-cache pool available. Unlike the flagship model, this processor features the lower 120W TDP of the Ryzen 9 9900X (and cheaper Ryzen 7 9800X3D), which debuted back in August and November 2024.

As with all Ryzen 7000 and Ryzen 9000 processors, the 9900X3D and 9950X3D are built on the AM5 platform, meaning you'll need a compatible motherboard. Similarly, these chips are built exclusively for DDR5 RAM and have PCIe 5.0 compatibility out of the box. If you're thinking of upgrading from an older AM4 platform, then you'll have to do more than just swap out the chips.

Regardless of which of the two new processors you choose from, you're going to want to invest in one of the best CPU coolers in 2025 to keep those core temperatures down whether you're gaming, rendering projects, or more. This is particularly true for the Ryzen 9950X3D with its 170W TDP, matching that of the non-3D variants, something we had not seen up until this point.

With a higher thermal design power comes increased heat along with the performance. We recommend a dedicated AIO with a large radiator attached (such as 240mm or 360mm) with a PC case with sufficient airflow and fans inside the computer case.

Leading CPU options, provided you've got the cash to spend

While both the AMD Ryzen 9900X3D and 9950X3D processors are expensive, the company has made the smart decision to keep the prices consistent with the previous generation versions (7900X3D and 7950X3D) from 2023. Whether the generational performance gains are strong enough to consider investing in is something that we'll only know after thorough testing, but those considering upgrading from AM4 should notice sweeping benefits.

How AMD's second-generation 3D V-cache works is by having a dedicated pool of L3 memory stacked onto the chip which acts as a short-term (but incredibly fast) resource for the processor to access immediate data, for higher gaming framerates and lower rendering times than what's seen from processors without a surplus of L3 cache / V-cache. By having frequently accessed data so close to the CPU cores, there's less of a delay between processes and higher performance as a result.

With that said, AMD chipsets with a higher TDP can (usually) outperform the lower TDP 3D options when it comes to productivity and synthetic performance. However, gamers and content creators will notice the edge in performance thanks to the 3D V-cache. It's a curious (but welcome) decision to see the Ryzen 9 9900X3D and 9950X3D with identical TDPs to their non-3D counterparts. We could, therefore, see improved performance across the board.

You may also like...

• New rumors claim RTX 5050 and RTX 5060 will launch in April
• AMD could be working on RTX 5070 Ti rival, the RX 9070 Extreme
• New benchmarks suggest the M3 Ultra isn't much faster than M4 Max

• MikroTik's ROSE Data Server RDS2216 combines storage, networking, and containers
• It runs on the 16-core AWS Graviton 1 Arm CPU and has 16 ports
• Supports MinIO, Nextcloud, and encryption, with no subscriptions or paywalls

MicroTik, a Latvian company that specializes in networking hardware and software, has launched a new enterprise server, powered by ROSE (the company’s RouterOS Edition for storage and compute).

Described as a “high-performance, all-in-one storage, networking, and container platform for enterprise environments,” the MicroTik ROSE Data Server RDS2216 combines a U.2 NAS, advanced switch, lightweight 16-core 2GHz ARM CPU and 32GB of DDR4 RAM.

The striking green color server, which has 16 ports including two 100G QSFP28, four 25G SFP28, four 10G SFP+ and two 10G Ethernet ports, also has twenty U.2 NVMe storage slots. It supports a number of advanced storage features, such as NVMe-TCP block device export, encryption layers, and modular configurations, with no subscriptions or paywalls. It's container-ready, seamlessly running MinIO, Nextcloud, Shinobi, Frigate, and other OCI-compliant containers, with additional USB ports for potentially even greater adaptability.

AWS Graviton 1

The ARM chip powering the RDS is certainly one of the more interesting things about the server. The AL73400 is a 16-core ARMv8 SoC developed by Annapurna Labs, which happens to be a subsidiary of Amazon. The chip, also known as the AWS Graviton 1 processor, was originally introduced in 2018 to boost Amazon's cloud infrastructure.

Beyond that, the server has 32GB of RAM (which is a bit stingy) and 128MB of NAND storage. It also features IPsec hardware acceleration, ensuring encrypted connections without performance bottlenecks. MicroTik says the device is rated for 200,000 hours MTBF at 25°C and operates within a temperature range of -20°C to 50°C.

Suggested uses for the RDS2216 include self-hosted MinIO cloud storage, NextCloud enterprise cloud hosting, high-speed backups and database clusters, containerized infrastructure, branch office storage expansion, auto-encrypting storage, and private social media hosting.

Writing about the server, ServeTheHome says, “MicroTik and Ampere have had postings about working together previously. Hopefully that becomes the network side of this with RDIMM support and more. Assuming this sells for under $2000 it might make for a really neat option, but it is also one we want to try before recommending.”

You might also like

• How to choose the right server for your small business
• These are best dedicated hosting services you can get right now
• The largest ever DDoS attack has just been blocked - here's how it was done

• Samsung, Meta and Apple are reportedly working on OLED VR headsets
• Meta's headset is tipped for 2026, while Apple's is set for 2027
• Samsung and Meta will use new OLED on Silicon displays

OLED, OLED, OLED – that’s the name of the game right now in the XR world, if a report on leaked hardware specs is to be believed. Samsung, Apple, and Meta are said to be working on OLED-display equipped headsets. And while Samsung's and Apple's plans are interesting, Meta’s OLED headset could be a winner for one big reason: cost.

Starting with Samsung – because its headset is the only one we officially know exists – the report via The Elec (a South Korean outlet, details translated and summarized by @Jukanlosreve) says the upcoming consumer version of Project Moohan headset (the Android XR wearable it developed with Google) will come with a 1.3-inch display boasting a pixel density of 3,800 PPI (pixels per inch).

Interestingly this OLEDoS display (OLED on Silicon, also known as Micro OLED tl;dr: they’re very tiny but high-density OLED panels built on silicon) would come from Sony, not Samsung Display, and it’s higher than the 3,386ppi display used by the Apple Vision Pro – suggesting that Samsung wants to cement Moohan as a high-end alternative the Apple headset. Hopefully it won’t be as pricey (I’m not holding my breath).

Speaking of Apple, it's said to be working on a Vision Pro that’s more consumer-friendly (read: cheaper). Instead of silicon, according to the above report Apple will rely on more traditional glass OLEDs with a 1,500ppi, with a release planned for 2027 – but I’m not holding my breath for this device to be affordable as being cheaper than a $3,500 headset could still be very expensive.

Meta's OLED headset plan

A cheaper Vision Pro isn't necessarily cheap (Image credit: Future)

Last up is Meta, which reportedly plans to use a 0.8 to 0.9-inch OLEDoS display in a headset set for 2026.

The smaller display size would apparently be a cost-cutting measure. OLED panels are typically made large and then cut to size, and Meta is apparently hoping that if it can reliably create a 1.42 OLEDoS unit (the same size as a Vision Pro) it can then cut that up into roughly four 0.8 to 0.9-inch panels, and cut its costs by a quarter. Given this small size, though, it may need to use two displays per headset (which would follow the two-display setup used in its Quest headsets), which would only halve the effective cost.

Regardless, OLEDoS displays, even at a quarter or half the cost, likely wouldn’t come cheap – which is why I’m expecting this 2026 headset would be the rumored Meta Quest Pro 2. It's supposedly been cancelled twice already, but the third time seems to be the charm, as Meta might have settled on a unique glasses-like shape (which I think sounds awesome, by the way), and some OLED screens would help cement it as an exciting proposition.

A 2026 window would also work for a Meta Quest 4 launch (based on the three-year gap between the Quest 2 and Quest 3 releases), although unless Meta can seriously bring costs down, OLED displays might be a bridge too far for its mid-range line, where they wouldn’t be for its high-end Quest Pro headsets.

As with all leaks and rumors we’ll have to take these details with a pinch of salt – especially in the case of Meta and Apple – as even if these are their current plans there’s plenty of time for things to change, and we won’t know what any of these brands has in store for us until they make an official announcement. Hopefully Samsung won’t leave us wondering for much longer, and perhaps Meta will tease something at Meta Connect 2025.

You might also like

• Ray-Ban and Meta just teased new limited-edition smart glasses
• Mark Zuckerberg just teased next-gen Ray-Ban smart glasses
• Meta's new smart glasses come with some much-needed upgrades

• Major Russian cryptocurrency site taken down by law enforcement
• Garantex has previously been sanctioned by the EU and US
• The site allegedly had links to cybercrime and ransomware

Law enforcement agencies from across the world have come together to shut down infamous Russian cryptocurrency exchange Garantex, which was popular amongst cybercriminals, especially ransomware gangs, who often use cryptocurrency to launder money gained in ransom attacks.

The Garantex website has been taken down and replaced with a notice confirming the seizure - with the shutdown comes just days after the EU sanctioned the site for being “closely associated with EU-sanctioned Russian banks”.

The US Attorney’s Office for the Eastern District of Virginia obtained a seizure warrant for the domain, which enabled law enforcement to take down the website. Several agencies are credited in the notice, including the US Department of Justice, Secret Service, the Finnish Poliisi Keskusrikospoliisi, and German Bundeskriminalamt.

Crypto losses

The US Government also sanctioned Garantex back in April 2022, after it was linked to criminal activity and dark web markets - including almost $6 million from Russian ransomware gang Conti.

Cryptocurrency is a popular tool for cybercriminals, as currencies are largely anonymous, untraceable, and unregulated - with no central server or administrator. In fact, it’s been reported that around $40 billion worth of illicit crypto transactions took place in 2024, and is predicted to rise in 2025. Dr. Ilia Kolochenko, CEO at ImmuniWeb explains;

“While some jurisdictions are willingly blind and readily close their eyes to certain types of illicit activities related to both cybercrime and money laundering, operating a fully functional crypto exchange solely from those “safe havens” is virtually impossible in the interconnected world.”

"The good news is that we will probably see even more successful cyber operations by law enforcement agencies in 2025, while the bad news is that cybercriminals will certainly learn the lesson and will very soon resurrect with even more resilient infrastructure to continue their business.”

Via TechCrunch

You might also like

• Take a look at our picks for the best antivirus around
• Check out our recommendations for the best malware removal software
• Cybercrime gang targets victims with "triple threat" attacks

Google Pixel phones are apparently kicking out some strange vibrations after the recent March update.

Over on the Google Pixel subreddit multiple Pixel users reporting that after they install the update the haptic vibrations their Pixel phones are serving up feel different than before.

Some notice this when typing on the Pixel virtual keyboard, while others claim that all vibrations feel different.

One poster going by the handle of Krysnosis noted: "On typing I don't feel it, but the back gesture I immediately felt the hollow feeling. It has less thunk and doesn't feel as clean. Now, it has a lingering light vibration. Before, it just ended cleanly."

It's hard to narrow down whether these changes are affecting particular Pixel models, or most Pixel phones that qualify for the update.

But the general vibe of the Reddit thread is that the change in vibrations feel less direct and robust, something that wasn't met with approval from those affected.

There's no mention of any changes to haptics or the intensity of the vibrations in Google's changelog for the March 2025 update, so this would appear to be a bug.

I'm in the process of downloading an update for my Google Pixel 9 Pro, which is taking some time due to my slow internet, so I'm afraid I can't bring you any insight from my side just yet.

The folks at 9to5Google say they haven't noticed any vibration changes on their Pixel 9 models; however, The Verge reported detecting vibration changes in the Google Pixel 8 phones, while Reddit users noted that the Pixel 7 and Pixel 7 Pro have been affected.

So it's all a little strange, and Google has yet to comment on the vibration issues. If you've notices changes to vibrations on your Pixel, let me know in the comments below.

You might also like

• Android 16 beta users are reporting major battery drain issues – but I’m not too worried about it
• Samsung says ‘millions’ are using Galaxy AI regularly, despite surprising survey results
• iPhone 17 Air leaks suggest it'll get next-gen battery – and offset the 17 Pro Max's weight gains

Independent auditors have put the security infrastructure of all NordVPN's applications under the microscope and found no critical issues, yet again.

Leading auditing firm Cure53 carried out a series of tests across all NordVPN desktop applications, mobile apps, browser extensions, and some key features. Experts uncovered a total of 31 findings, which the Nord team mostly fixed at the time of writing.

The security audit results come only a few weeks after TechRadar's best VPN provider also proved its no-log claims with a third-party check, showing the provider's continuous commitment to transparency, privacy and security.

NordVPN security audit

Experts at Cure53 conducted a mix of penetration and source code reviews between June and August 2024, for a total of fifty-five days.

As mentioned earlier, NordVPN apps (Windows, macOS, Linux, iOS, and Android VPN) and browser extensions (Chrome, Edge, and Firefox) weren't the only tech to be inspected. Auditors also took apart NordVPN Threat Protection, Threat Protection Pro, and Meshnet.

Cure53 found a total of 31 findings, with 22 being classified as security vulnerabilities (with some ranked as High) and nine as general weaknesses with lower exploitation potential.

"This security assessment revealed a high number of issues. However, given the broad scope included in Cure53’s examination, and the large attack surface it encompassed, a higher than average number of issues was to be expected," noted auditors in their final report, sharing recommendations to fix these issues.

At the same time, however, auditors observed "the system utilized several well-regarded libraries, including NGHTTP2, OpenSSL, and Boost," which are known for their stability and security.

On their side, NordVPN has welcomed Cure53 suggestions and already enforced a fix for most of the issues, which was also verified by Cure53.

"Security is at the core of everything we do at NordVPN. Independent assessments like this allow us to continuously refine our technology and stay ahead of emerging threats," said NordVPN CTO Marijus Briedis, ensuring the team swiftly implemented all necessary improvements to ensure the highest level of protection for users.

Despite the findings, the provider explains, the latest Cure53 assessment has confirmed that NordVPN apps are built on a strong security foundation as no critical issues were found.

The security audits come as the provider confirmed its no-log claims for the fifth time since 2018 back in February. In that instance, experts at Deloitte inspected NordVPN's server configuration and relevant IT systems to ensure data related to users' activities is never logged as stated in its privacy policy.

"Our work towards improving security is never finished, and we will keep moving forward," said Briedis. "We are proud of these results and will keep making NordVPN one of the most secure VPN services available to everyone."

You can read the full Cure53 report by heading to the user control panel on the provider website or clicking here.

• AMD has finally introduced AFMF 2.1 in its new FSR 4 driver update
• Its compatible with RDNA 4, 3, and 2 GPUs, including Ryzen AI 300 series processors
• Performance and image stability in games on the Asus ROG Ally are improved using AFMF 2.1

AMD's RDNA 4 GPUs are finally here, along with a new driver that adds FSR 4 support, but that's not all - handheld gaming PCs and RDNA 3 and 2 GPU users will have a new and improved trick up their sleeve to improve performance in games.

As detailed in AMD's Adrenalin Edition release notes, AFMF 2.1 is now available in driver 25.3.1 for Radeon RX 6000, 7000, and 9070 series GPUs including Team Red's Ryzen AI 300 series processors. This is a driver-based frame generation feature (similar to Nvidia's new Smooth Motion) that allows gamers to force frame-gen into games that don't have official implementation from developers.

AFMF 2.1 is built to improve upon frame generation image quality along with reducing ghosting and better temporal tracking - using AFMF 1 and 2 on handheld devices like the Asus ROG Ally wasn't an easygoing experience since ghosting, stutters, and poor image quality were common complaints from people who used it.

Based on some of my early impressions, AFMF 2.1 is a good step up from its predecessor, providing a cleaner image in motion and reducing ghosting to a significant degree - this was evident in my testing of Sifu and Resident Evil 4 (pictured below), two games that have previously struggled when using AFMF with ghosting and constant frame time stutters.

Fortunately, those issues have been minimized as the non-stop action in Capcom's title isn't interrupted by regular frame time spikes (which used to occur even when idle in-game) allowing for a much smoother handheld experience - and it's a great sign for other games that have previously struggled with frame generation software on handheld gaming PCs.

Image 1 of 2

Where's everyone going? Bingo? (Image credit: Future/Isaiah Williams)Image 2 of 2

(Image credit: Future/Isaiah Williams) Now, we just need FSR 4 to come to RDNA 3...

I know FSR 4 is currently exclusive to Team Red's new Radeon RX 9070 series, but if the new upscaling technology can find its way onto RDNA 3 hardware, handheld gamers are in for a treat.

We've seen how close FSR 4 is to Nvidia's DLSS 4 and its new transformer model - while it doesn't defeat it nor match it, it's managed to topple its fierce rivals' previous CNN model, with greater image stability (especially in motion). AFMF 2.1 has impressed me and given me another reason to keep my ROG Ally (at least for now), and any addition of FSR 4 will likely be a showstopper.

Its performance mode is a prime example of the significant enhancements, which reminded me of how great DLSS 4's performance mode looks - imagining how this would perform on a handheld gaming PC is the stuff of dreams for me.

The only strong competition AMD has for producing chips for handhelds is Intel - while the MSI Claw 8 AI +'s Ultra Core 7 258V processor is more powerful than the Z1 Extreme, the current version of XeSS (Intel's upscaling tech) likely can't match up to FSR 4. In my opinion, Team Red's new upscaling method is the final piece of the puzzle needed to make handhelds like the ROG Ally complete - let's just hope it can be done...

You may also like...

• New benchmark suggests Apple's M3 Ultra may not be much faster than the M4 Max - only a minor uplift in multi-core performance
• Sure, Nvidia DLSS 4 is incredibly impressive - but AMD's improved upscaling tech could be a real game-changer
• Finally, we have some GPU competition - AMD announces the Radeon RX 9070 XT March 6 launch date, starting at $599 alongside the RX 9070 at $549

• Security researchers found JavaScript code installing four backdoors to WP-powered sites
• They also found a vulnerable plugin enabling full website takeover
• There are patches and mitigations for all these vulnerabilities

A single piece of JavaScript code deployed no less than four separate backdoors onto roughly 1,000 WordPress websites, according to a new report from cybersecurity researchers c/side, who detailed the four backdoors and explained how website builder users should protect themselves.

The analysis did not elaborate how the malicious JavaScript made it into these websites - we can assume either weak or compromised passwords, a vulnerable add-on, or similar. In any case, the code is served via cdn.csyndication[dot]com, a domain mentioned in at least 908 websites.

It deploys four backdoors. One installs a fake plugin named “Ultra SEO Processor” that can execute commands remotely, one injects malicious JavaScript into wp-config.php, one adds an SSH key to allow threat actors persistent access, and one runs commands remotely and opens a reverse shell.

Chaty Pro 10/10

To minimize the risk, c/side advises website owners delete unauthorized SSH keys, rotate their WP admin credentials, and scan system logs for any suspicious activity.

At the same time, PatchStack found Chaty Pro, a popular WordPress plugin with some 18,000 installations, was enabling malicious file uploads on websites where it was installed. Chaty Pro allows owners to integrate chat services with social messaging tools.

The flaw is tracked as CVE-2025-26776 and has a 10/10 severity score (critical). Since threat actors can use it to upload malicious files, it can lead to full website takeover, hence the critical severity. Infosecurity Magazine reports the function included a whitelist of allowed file extensions which was, sadly, never implemented.

“Uploaded file name contains the upload time and a random number between 100 and 1000, so it is possible to upload a malicious PHP file and access it by brute forcing possible file names around the upload time,” PatchStack explained.

Chaty Pro’s maintainers released a fix on February 11. All users are advised to upgrade the extension to version 3.3.4.

Via The Hacker News

You might also like

• Google Chrome extensions hit in major attack - dozens of developers affected, so be on your guard
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• The RTX 5050 and RTX 5060 are rumored to launch in April 2025
• It's believed the MSRP of the former could be $299, matching the RTX 3050
• Claims are circulating that there will be three variants of RTX 5060 available

The Nvidia RTX 5050 and RTX 5060 are rumored to launch next month, based on the latest rumors circulating online.

According to WCCF Tech, the two mainstream Blackwell RTX 50 series graphics cards will be released soon, following the midrange RTX 5070, which launched earlier this week.

It's believed that the RTX 5050 will use the PG162 PCB with 8GB GDDR7 VRAM and a 145W TDP. Its pricing is expected to fall within the range of $199 to $249, depending on the manufacturer. This would position it alongside other entry-level graphics cards on the market, such as the Intel Arc B580 from December 2024, which we praised with a five-star score in our review.

April could also see the launches of both the RTX 5060 and the RTX 5060 Ti, with the former launching at the end of the month and the latter claimed to be coming earlier. It is rumored that both GPUs will utilize 8GB GDDR7 VRAM (with the Ti supposedly having a 16GB option), with the RTX 5060 expected to retail from $299, but this is unconfirmed at this time, and so far few concrete details are known about the hardware inside the two budget offerings.

Additionally, alleged industry insider MEGAsizeGPU has claimed that the RTX 5060 family could be announced 10 days from now to hit the shelves "a month later". This backs up WCCFTech's information about a launch coming sooner rather than later, and it's certainly believable considering the trajectory we've typically seen in Nvidia's graphics card launches. Historically, the 90 and 80-class cards come first, with 70-class and mainstream offerings following closely behind.

While unconfirmed, TechPowerUp claims that the RTX 5060 will be built on the GB206 graphics processor with 4,608 cores, 144 Texture Mapping Units, and a 128-bit memory bus combined with its 8GB GDDR7 VRAM. In contrast, this source claims the RTX 5060 Ti's 16GB variant will be otherwise identical, save for double the VRAM. It's likely a placeholder until an official reveal and tech specs are announced, however, it gives us a rough idea of how they could stack up to the company's best graphics cards on the market.

A return for Nvidia's 50-class dedicated graphics cards

Should these rumors be true, then we will be seeing the return of the 50-class graphics cards for the first time since January 2022 with the desktop RTX 3050. While far from gaming powerhouses, these affordable cards have (traditionally) given wallet-conscious gamers a way to keep up with today's demanding games in 1080p, even featuring some light ray tracing functionality.

While there technically was an RTX 4050 GPU, it was solely used as a graphics solution inside cheap gaming laptops and never saw a release as a dedicated graphics card. The RTX 4060 was the mainstream leader of Nvidia's previous gen, starting at $299. Depending on the MSRPs of the RTX 5050 and RTX 5060, we could see a disparity in price similar to the difference between AMD's RX 9070 ($549) and 9070 XT ($599).

At a time when it looks as though Nvidia is solely focused on pushing the goal posts in terms of both price and performance, the addition of a 50-class and three 60-class versions of Blackwell could make DLSS 4 (and Multi Frame Generation) far easier to access for those who don't want to pay out $600 or more at the minimum.

Performance of these cards are unlikely to be groundbreaking, but if upscaling from 720p to 1080p for the RTX 5050 and 1080p to 1440p with MFG, we could see the perennially popular RTX 3060 12GB and RTX 4060 finally dethroned from their top spots in the Steam Hardware Survey.

You may also like...

• AMD could be working on a faster RX 9070 Extreme graphics card
• New benchmarks show the M4 Pro may not be much of a leap after all
• United Airlines successfully installed Elon Musk's Starlink in one of its planes

• A new botnet called Eleven11bot was spotted in the wild
• It leverages weak and default credentials to compromise IoT devices
• The botnet is operated by Iranian threat actors

Cybersecurity researchers say they have uncovered the “biggest non-government botnet” in recent years.

It is called Eleven11bot, and its malware was found on more than 86,000 Internet of Things (IoT) devices, according to multiple research teams, including Nokia, GreyNoise, and The Shadowserver Foundation.

The botnet is most likely operated by an Iranian threat actor, GreyNoise reported. It found some 1,400 IPs operating the botnet, the majority of which are based in the Middle Eastern country. The threat actors seem to be hunting for IoT devices with factory or weak credentials, and actively scanning for exposed Telnet and SSH ports, with compromised devices including webcams, Network Video Recorders (NVR), and similar.

Exceptional size

At the same time, The Shadowserver Foundation analyzed the spread of the malware, and found that the majority of compromised endpoints are located in the United States, United Kingdom, Mexico, Canada, and Australia.

Botnets are most commonly used for Distributed Denial-of-Service (DDoS) attacks, where infected devices overwhelm a target server, causing disruptions.

They are also used for sending massive spam campaigns, distributing phishing emails or malware while avoiding detection. Cybercriminals leverage botnets for credential stuffing and brute-force attacks, trying to break into accounts using stolen credentials.

Another frequent use is click fraud, where infected machines generate fake ad clicks to inflate revenue. Botnets also enable cryptojacking, secretly mining cryptocurrency on victims’ devices, slowing them down and increasing electricity costs. Additionally, they are used for data theft and espionage, stealing login credentials, financial data, or trade secrets.

Via BleepingComputer

You might also like

• Massive botnet is targeting Microsoft 365 accounts across the world
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app