Feed aggregator

Yesterday morning, I logged into Facebook and saw an image of the Colosseum turned into a water park. On LinkedIn, everyone was busy transforming their headshots into Studio Ghibli characters, courtesy of ChatGPT’s latest update. Threads showed me a video reimagining the cast of Severance crawling around Lumon as babies. And X kindly served up a Grok-generated image of Elon Musk and Donald Trump in a pose I wish I could unsee.

It’s not just me, right? Social media has become a swirling mess of synthetic content. We’ve written about the rise of AI slop before – part cringe-inducing art gallery, part uncanny valley fever dream. And sure, the algorithm is feeding me stuff it thinks I care about – Severance babies, travel inspo, a healthy dose of Musk mockery. That checks out.

But I keep thinking: has AI officially ruined social media? And is this just harmless chaos or is it quietly rewiring how we see truth, trust, and reality?

The four horsemen of the AI slopocalypse

One of the biggest problems with AI content is the sheer volume – and it’s coming from all sides. AI evangelists. Your coworkers. Your friends. Your gran (who probably doesn’t even realize what she’s reposting). And brands that absolutely should know better.

To make sense of the mess, I spoke to Joe Goulcher, a creative director and social media expert. He works with brands on this stuff daily and has a front-row seat to the AI slop flood. According to Joe, AI content tends to fall into four distinct “strains” – “like a virus,” he tells me.

Functional AI slop
“Crudely, badly made stuff that is basically stock imagery used to fill a hole where an image should be,” Goulcher explains. “It's bottom of the barrel, and barely any conscious thought has gone into why or what it is.” This is the lowest-effort tier. Bland visuals slapped onto posts just to have something there. Placeholder content that somehow became the content.

Clickbait slop
“Stuff that makes us stop scrolling and think ‘god this is disgusting and bad,’ but it’s by design to generate conversation,” Goulcher says. This one’s the most insidious. It’s not trying to be good, it’s trying to be just bad enough to go viral.

The ‘look what I made!’ post
“This is where people use AI to create something that looks like LOTR, Star Wars or another behemoth IP because it’s trained on it,” Goulcher tells me. “They say things like ‘look what I did in ten mins!! this is going to change the industry and you should be scared and using it now instead of spending thousands with artists.’ It usually goes down like a sack of bricks.” This is the hype-fuelled, tech-bro theatre of generative content. The tone is always breathless, and the results are always underwhelming.

The genuinely good stuff
“There are actually good AI campaigns, with purpose, permission, and laced with incredible VFX, handcrafted where AI couldn't do the job,” Goulcher says. Yes, some brands are doing it well, with thought, care, and actual artistry. But it’s rare. And usually buried under a steaming pile of junk.

Breaking it down like this might feel bleak – like we’ve gone full epidemiology on the most cursed content – but it’s actually useful. Categorizing the chaos helps explain why AI content feels inescapable, and why it hits so many different shades of dystopian.

AI for the sake of AI

We’re in an era of AI content being made simply because it can be. People, brands, entire businesses are churning it out. Not necessarily because they have something to say, and not because it’s better than the alternatives, but because the tools exist, they're easy to use, and the pressure to use them is enormous.

This speaks to a deeper issue in tech. Just look at Apple’s recent AI missteps. Despite all the hype, AI isn’t delivering the magic it was sold on. It’s being shoved into products not because users need it, but because shareholders want to hear “AI” on earnings calls.

And right now? It’s not revolutionizing much of anything. In fact, in most applications, it’s starting to look like a very expensive gimmick.

“It reminds me of the early days of torrenting or music streaming – a bit of a Wild West,” Goulcher explains. “Even though brands whacked a logo on them, it didn't make it ethically good in any way. They were just trying to ride the waves of legality and dosh until legislation kicked in far too late.”

Like crypto or NFTs, we see there’s innovation, hype, overuse, and then fatigue. “I think we can apply the same tech hype graph to AI content,” Goulcher says. “There's always something in my gut that’s like, ‘this bubble will burst.’ And I still think that will happen. When AI data sets start eating themselves, and the innovative wow factor wears off – what’s left?”

And he’s asking the question more of us probably should be: “How are any of these billion-dollar tools actually making our lives better? Because the novelty is wearing off, the ethics (or lack of them) are becoming clearer, and the shock value is beginning to wane.”

AI hasn’t just broken social media, it’s broken the truth

It’s not like social media was perfect before this. AI didn’t start the rot, things were already slipping. But this latest wave of generative content has pushed it straight into uncanny, derivative, brain-melting chaos.

And if AI is flooding our feeds with pointless slop, it’s also doing something more dangerous: weaponizing it. It’s easy to laugh at AI-generated celeb babies or cringe at a brand ad that crawled out of the uncanny valley. But that reaction misses the bigger, scarier picture.

Because AI can fuel misinformation at scale. For example, across Europe, far-right groups are using AI-generated images to provoke outrage, spread conspiracy theories, and stoke division. And these aren’t just fringe trolls – they’re coordinated campaigns, designed to manipulate public opinion.

That’s just one example. Election misinformation. Deepfake porn. Fake war footage. Yes, it's the kind of content that has always existed online. Only now, the rules have changed. You don’t need skills. You don’t need a team. You don’t even need a budget. Just a narrative and a willingness to abandon reality. And on social media, the rest takes care of itself.

That’s the real horror story. Not just that we’re drowning in junk, or that brands think we want AI-generated ads, but that we’re slipping into a world where what’s fake moves faster than what’s true. And the algorithm doesn’t care if it’s real – just that it’s getting your clicks, your likes, your attention.

You might also like

• I used ChatGPT to engage in a debate about zombies but did its arguments leave me for dead?
• Pi is a friendlier, cosier alternative to ChatGPT – whether that's a good thing depends on you
• ‘Our GPUs are melting’ – OpenAI puts limits on image creation and delays rollout to free accounts

• Microsoft has reportedly canceled data center leases amid shifting infrastructure priorities
• TD Cowen analysts suggest lease deferrals reflect oversupply vs demand
• Microsoft doesn't deny the report but says it's investing $80 billion in infrastructure

We recently reported Microsoft had cancelled leases with at least two private data center operators in the US, totaling “a couple hundred megawatts,” and how the company was also not converting “so-called statements of qualifications into leases,” according to claims from TD Cowen analysts.

Shortly after that news broke, Microsoft pulled out of a $12 billion deal with CoreWeave (the "WeWork of AI"), passing on buying more data center capacity from the AI hyperscaler. That option was snapped up by OpenAI, but as it counts Microsoft as its biggest backer, it was essentially paying CoreWeave with Microsoft money!

Microsoft appears to be taking a more tactical approach to AI spending, a move that is echoed in a new Bloomberg report quoting TD Cowen analysts saying Microsoft has walked away from additional data center projects in the US and Europe.

On track to spend $80 billion

Bloomberg writes, “Microsoft’s retrenchment in the last six months included lease cancellations and deferrals, the TD Cowen analysts said in their latest research note. Alphabet Inc.’s Google had stepped in to grab some leases Microsoft abandoned in Europe, the analysts wrote, while Meta Platforms Inc. had scooped up some of the freed capacity in Europe.”

Responding to the article, Microsoft pointed out that it was still on track to spend about $80 billion investing in growing infrastructure projects.

“Thanks to the significant investments we have made up to this point, we are well positioned to meet our current and increasing customer demand,” a Microsoft spokesperson said in a statement.

“While we may strategically pace or adjust our infrastructure in some areas, we will continue to grow strongly in all regions. This allows us to invest and allocate resources to growth areas for our future.”

Bloomberg adds TD Cowen analysts Michael Elias, Cooper Belanger, and Gregory Williams said, “We continue to believe the lease cancellations and deferrals of capacity points to data center oversupply relative to its current demand forecast.”

You might also like

• Is Microsoft hesitating on AI? Tech giant has cancelled data center leases
• Microsoft is building wooden data centers in its latest push to cut emissions
• Investments, action plans, and the shifting AI landscape

I sat down with Khazan's voice actor Ben Starr and Neople CEO Myeongjin Yun to chat about their new game in the Dungeon & Fighter universe.

• A Hyundai Ioniq 5 has driven over 500,000km on its original battery
• Battery health degradation is considerably lower than many predicted
• Models from Nissan, Tesla and Chevrolet can be found with high mileage

News has surfaced this week of a Hyundai Ioniq 5 owner in South Korea who clocked up a staggering 360,000 miles (579,363 km) on their EV's original battery.

Rewind the clock to 2011 and the introduction of the first-generation Nissan Leaf and industry experts were worried that EV batteries wouldn’t last five years before they need replacing… at a huge cost.

However, the Hyundai Ioniq 5 owner, who regularly posts on a Facebook Group called “Mileage Impossible”, claims he covers around 10,000 miles, reaching 360,000 miles before Hyundai offered to replace the battery for free, despite it being well beyond its warranty period.

Perhaps most amazing of all is that the original battery pack was still showing 87% of its original health, despite the owner reportedly only using DC fast charging to top up the battery packs, according to Electrek. This was likely so Hyundai could run tests on the battery pack for its own data records.

Still going strong, the Hyundai Ioniq 5 has now hit 666,255km (413,991 miles) and counting, although it hasn’t been completely free of issues. The owner reported that the Integrated Charging Control Unit (ICCU) broke, meaning it can’t be charged on Level 1 and Level 2 low-powered AC outlets, while standard maintenance and general wear and tear will have likely thrown up some bills.

But the high-mileage Hyundai is a shining example of how robust and reliable a modern EV can be if it is taken care of properly. And it isn't alone...

Electric vehicles keep rolling

(Image credit: GearBrain)

For data nerds, the high-mileage Hyundai example is an anomaly and it is much better to look at numbers sourced from the wider EV industry to get a more reliable picture on battery and electric motor health after vehicles reach very high mileage.

That said, the news isn’t awash with reports of EV batteries failing after just a few years (you know it would be if that was the case). And there are numerous reports of owners maxing out their mileage in a number of electric vehicles with very little issue.

Here are just a handful of examples...

1. Tesla Model S 70D

(Image credit: Tesla)

Inside EVs ran a story about how one Tesla owner, who used the vehicle as a taxi service, covered a mammoth 70,000 miles (112,650km) a year. At the last count, the car showed a staggering 424,000-miles on the clock (682,361km).

According to the report, the battery pack was replaced under warranty at 250,000 miles (402,335 km), as well as a new front motor 380,000 miles (611,550 km).

Amazingly, it still uses its original rear motor unit but, like any combustion engine counterpart, it has also received a number of new parts to keep it roadworthy.

Rear axles, suspension components and the AC compressor all required replacing, which is understandable on a vehicle that has gone that sort of distance. We bet this isn't the only example of a Model S managing to out-last combustion engine counterparts.

2. Tesla Model 3

(Image credit: Tesla)

Although it lacks the video evidence of the previously mentioned Model S, one Model 3 owner took to Twitter in 2022 to post an image of their Model 3 Long Range Dual Motor hitting 500,000km (or 310,685).

According to the owner, it was still using the original battery pack, motors and even brakes, which is slightly concerning. Although the owner admitted that they drove 90% of the time at highway speeds.

The only maintenance, aside from tires and other commonly replaced parts, was a new drivetrain oil pump at 286,000 miles (460,000 km), which reportedly cost just 250 Canadian Dollars ($194).

3. BMW i3

(Image credit: BMW)

BMW put out a press release in 2019 to celebrate Shaun Maidment's achievements, an i3 owner in South Africa who had clocked up 200,000km in a first generation 60Ah model in just three years of ownership.

According to an update in 2022, the vehicle had amassed 187,000 miles (301,998 km). However, it is worth noting that Shaun’s i3 was the range extender variant, so the smaller battery pack is capable of 126 miles on electricity alone, with petrol used to power the range extending system to around 200 miles.

Regardless, it's another great example of a car that wasn't designed to hit such lofty mileage proving the naysayers wrong.

4. Chevrolet Bolt

(Image credit: Chevrolet)

The Chevy Bolt forums are awash with members trying to outdo each other on mileage, but one video surfaced that showed an owner managing a staggering 138,000 miles (222,089km) on its original battery, with the second still going strong at a further 122,000 miles (196,339km).

The vehicle is still cruising happily at 300,000 miles and counting, with the content owner documenting the life of his 2019 model in various videos on YouTube.

5. Nissan Leaf

(Image credit: Nissan)

A driver in Washington managed to rack up 150,000 miles in his first-generation Nissan Leaf, with the original battery pack managing to go the distance, despite the owner reporting that the battery fade has been considerable.

But considering the battery technology is now 16 years old, it is amazing that the packs still offer 60% capacity, which is good for around 35 miles of range in warm weather.

Owner Steve Marsh told Inside EVs that he now has a newer 2014 Leaf for everyday use, with the the last few clicks to hit 150,000 miles on his original model more a "proof of capability" than a conscious decision to use the car.

You might also like

• 5 questions Severance season 3 needs to answer when the Apple TV+ hit returns
• I hate to say it Apple TV+, but Severance season 2 has a pacing problem
• The best mechanical keyboards of 2025
• Severance season 2 episode 10 ending explained: what does Mark do, who dies, will there be a season 3, and more big questions answered

• AMD's CEO Lisa Su hints at more Ryzen AI Max APUs coming to desktop PCs
• Some of the APUs have been used in Framework Desktops
• It could be the start of discrete GPUs being subsided

While Nvidia and AMD's new GPU lineups are at the forefront of PC gaming news, Team Red's processors are on the verge of taking the spotlight: the Ryzen AI Max APUs have mostly been reserved for laptops, providing enough power to run demanding triple-A games, but that could change very soon.

In an interview with DeepSeek Exploded on Bilibili, AMD's CEO Lisa Su hinted at more Ryzen AI Max processors making their way to the desktop space. We've seen the top-end processor (Ryzen AI 395+) in some mini-PCs and laptops, which provide exceptional performance results in games - and we could see the same results for even more desktop PCs.

Thanks to the Ryzen AI Max+ 395's Radeon 8060S iGPU with 40 RDNA 3.5 GPU cores and 16 CPU cores with 32 threads, gaming works great above 1080p. It's evident in ETA Prime's Framework Desktop PC tests on YouTube: a prime example is Cyberpunk 2077, with its frame rate hovering between 55 and 60fps with ultra graphics settings at 1440p.

It's also worth noting that FSR 4 could be making its way to RDNA 3 - it's currently reserved for RDNA 4 - and this could be incredibly beneficial to systems powered by the APUs.

The important thing is that all of this is possible without a discrete GPU, and with a desktop PC, this could work wonders for efficiency. While it won't match high-end desktop GPUs, it still packs a punch and could be the start of something new - or perhaps, the beginning of the end for discrete GPUs.

(Image credit: AMD) If we can get stronger APUs sooner, it may be time up for some discrete GPUs

For a long time, discrete GPUs for both laptops and desktop PCs have been the one-way ticket to great gaming experiences. Now, that's gradually beginning to change with APUs becoming more powerful: it's evident with handheld gaming PCs like the Asus ROG Ally, Lenovo Legion Go, and the recent MSI Claw 8 AI+ using Intel's Core Ultra 7 258V processor.

While the high-end Ryzen AI Max chips haven't been used in handheld gaming PCs yet, they're the ideal chips for mini-PC setups. Discrete GPUs (specifically desktop ones) provide fantastic performance across the board but use plenty of power - however, that won't be the case with an AI Max+ 395.

Tech is moving faster than ever, and APUs are the perfect example. If we manage to get more powerful APUs in the foreseeable future, we might just see a slow fade away of mid-range discrete GPUs - and perhaps a strong rival to current consoles (if the prices for the PCs are reasonable, anyway).

You may also like...

• AMD in 2024: year in review
• What is AMD FSR? The upscaling technology explained
• The best AMD processor in 2025: top AMD CPUs for gaming, creating, and more

Xbox fans: College Football 25 Standard is now a massive $60 off. But act fast -- this deal is only available until the end of today.

Black Mirror season 7 looks very promising so far, and I'm excited to dive back into one of the best Netflix shows when it returns on April 10. Like all anthologies, there have been some duds along the way, but personally, I've enjoyed most of the twisted tales.

Charlie Brooker's dystopian series has changed a lot since its Channel 4 days, achieving global success on Netflix, one of the best streaming services. Huge things are coming in season 7, too, like the anthology's first-ever sequel episode.

Take a look at the announcement teaser below for a proper look at season 7.

What will each episode in Black Mirror season 7 be about?

Will Poulter and Asim Chaudhry return for Black Mirror season 7. (Image credit: Netflix)

First off, we have six new episode titles giving us little clues of what to expect. These are 'Common People', 'Bête Noire', 'Hotel Reverie', 'Plaything', 'Eulogy', and 'USS Callister: Into Infinity'.

We do have brief synopses for each too, if you're keen to learn more. For 'Common People', which stars Sunny's Rashida Jones, Netflix teases: "When a medical emergency leaves schoolteacher Amanda fighting for her life, desperate husband Mike signs her up for Rivermind, a high-tech system that will keep her alive – but at a cost…"

Then there's 'Bête Noire', which reads: "Maria is a high-flying development executive at a chocolate company. Everything is going well for her until someone she hasn’t seen since school – a woman named Verity – shows up for a focus group tasting session.

"It could be the chance for a heartfelt reunion – except there’s something very odd about Verity, and Maria seems to be the only person to notice."

Emma Corrin, Issa Rae, Awkwafina, and Harriet Walter star in 'Hotel Reverie', a black and white one, much like how they filmed season 4's 'Metalhead'. The synopsis reads: "Hollywood A-list actress Brandy Friday is thrown into an unusually immersive high-tech remake of a vintage romantic movie, and must stick to the script if she ever wants to make it home…"

Episode 4, 'Plaything', stars Peter Capaldi alongside Will Poulter and Asim Chaudhry, who are reprising their roles from the interactive movie Bandersnatch, something that came as a shock to me in my Black Mirror season 7 trailer reaction.

The synopsis is: "An eccentric loner named Cameron who harbours an intense obsession with a mysterious '90s video game is arrested in connection with a grisly cold case – and his interrogation soon goes to places the police weren’t expecting."

'Eulogy' is the fifth episode, and its synopsis is: "An innovative system that enables users to literally step into photographic memories of the past leads a lonely man to re-examine a heartbreaking period from his past."

Finally, the anticipated 'USS Callister' sequel sees the old crew reprising their roles and reads: "Robert Daly is dead, but the crew of the USS Callister – led by Captain Nanette Cole – find their problems are just beginning."

You might also like

• I love Netflix but I'm sad to see these 5 movies with over 89% on Rotten Tomatoes leave in April 2025
• Everything leaving Netflix in April 2025 – from the scariest movie ever made to a beloved DreamWorks animation with 99% on Rotten Tomatoes
• Netflix is firing up the Mystery Machine with a live action Scooby-Doo series

New court filings give more details about a small number of DOGE staffers granted sweeping access to sensitive government data systems.

(Image credit: Scott Olson)

Apple devices rarely go on sale, so it's wise to check out this last-minute Amazon Big Spring Sale deal on iPads.

• Sony consoles have a VRR problem
• Tests have revealed it's not an issue with a specifc game or monitor
• There's no fix yet, but hopefully Sony will release one soon

If you’ve been experiencing stuttering while playing games on your PS5 or PS5 Pro then you aren’t alone. And finally we might know what’s causing it, but on that front there’s good and bad news.

Many players have noticed that when playing games that support 120Hz VRR (variable refersh rate) there’s an occasional stutter. It usually takes about 20 minutes to begin stuttering, but once it does happen, the stutter recurs about every 8 seconds – and naturally, this regular cadence adds to the distraction it causes.

Digital Foundry decided to investigate and has found that the issues appear to be with the console itself rather than any specific game or other hardware.

As detailed in its very in-depth YouTube explanation (and shared in written form via Eurogamer) Digital Foundry discovered this visual hiccup occurs on several different titles – including The Last of Us Part 2, Hogwarts Legacy, Elden Ring, and Kingdom Come Deliverance 2.

It also found the stuttering occurred across different displays including LG OLED TVs, a Samsung mini-LED display, and an Eve Spectrum monitor. Digital Foundry even tested the same games across both Xbox and Sony consoles and found only the Sony hardware had issues.

However, while Sony’s tech has been pinned as the source of the issue, what hasn’t yet been discovered is why the games are stuttering.

We do know that the stutter itself is because games will be running within a smooth frame rate window (say 60 to 70 fps) then every 8 seconds or so, spike momentarily to 120Hz. It’s why this spike happens which we don’t know.

With all signs pointing Sony’s tech being the issue it should hopefully be fixable with a simple software update in the near future. For now, your best immediate solution would be to disable VRR in your system’s settings and play with a fixed refresh rate.

It’s not ideal, but it should tide you over until Sony can put out a patch.

You might also like

• The PS5 Pro has made me realize I hate choice
• Best PS5 games 2025 - top PlayStation 5 titles to play right now
• The best PS5 deals for March 2025

Don't be surprised if you have to turn your VPN off at some point.

• A Google Photos redesign could be on the way
• A survey shows a new design with a focus on search features
• It also makes Memories larger and rounds the edges of your snaps

According to a leak from a recent survey, Google Photos could be about to get a design overhaul, which could make managing your library of snaps a little easier.

Details of a supposed Google survey were shared with Android Authority by a user on Telegram called @Arfus_UwU, who sent a screenshot of two Google Photos layouts. One is the current layout, and the other a redesign, which the user is asked to judge based on how modern or outdated it feels out of 50.

At first glance, the designs seem very similar, with only a couple of minor changes, such as making image borders more round and making the Google Photos Memories action buttons a little larger – the latter of which you’ll either appreciate or hate based on how frequently you use Memories.

However, a few subtler changes could make a world of difference to the Google Photos experience.

Big changes coming?

The current Google Photos design, which could soon be changed (Image credit: Future)

Firstly, we can see that the Today heading no longer has a checkmark next to it and instead shows what appears to be a filter icon. This means instead of being able to quickly select every photo you took that day you’d be able to trim down the snaps you see based on criteria like where they were taken and who’s in the shot.

At the bottom, we can also see that the Photos Collections and Search tabs have been replaced by a floating search bar and an icon that looks like it leads to your collections pages.

All of these changes look to streamline the process of finding a specific photo in as few taps and swipes as possible.

Whether you love or hate the supposed redesign, remember to take this leak with a pinch of salt. Even if Google is conducting surveys, there’s a chance it won’t copy the survey’s design exactly in its eventual rollout – it’s probably testing for a bunch of different factors when asking for opinions.

Even so, we, for one, love the new look. Memories could be smaller but the easy access search bar and filter options look super useful, especially as Google’s AI gets better at smart search so it can locate the precise picture you’re describing but can’t spot in your camera roll no matter how much you scroll.

You might also like

• Google Photos is getting a handy Quick Edit tool for sharing images, and I don’t know why nobody thought of this sooner
• Google Photos gets new Recap feature – and it’s like Spotify Wrapped for your life in 2024
• Look for the AI disclaimer from Google on photos that look a little too good to be true

• Surfshark is looking to improve the privacy of current end-to-end encryption with a new patent
• The plan is currently based on distributed trust-based communication frameworks
• Despite this, Surfshark says its approach differs from the likes of NymVPN and ObscuraVPN

Surfshark has just registered a new patent looking to improve the privacy of current end-to-end encryption (E2E) systems.

Based on a distributed trust-based communication infrastructure, Surfshark's proposed method seeks to reduce the amount of visible metadata – meaning all the data that isn't the content – by splitting the encryption process between two separate VPN providers.

Better metadata privacy

Encryption refers to the scrambling of data into an unreadable form and is the technical solution born to protect the network-based communications we all rely on daily – whether that's a text message, a document, or a photo shared with another user over the internet.

End-to-end encryption (E2E) is the tech used by virtual private networks (VPNs), some secure email services, and messaging platforms to ensure the content of these online activities remains private between the sender and the receiver.

A lot of metadata is still visible to the provider

Karolis Kaciulis, Surfshark

Yet, Surfshark's Lead System Engineer, Karolis Kaciulis believes that it's time to go beyond E2E.

He told TechRadar: "After the emergence of E2E encryption, we feel that the topic of user anonymity and security while using various messaging systems and technologies has stagnated. We believe there is still room for improvement."

The main issue with today's encrypted messages, Kaciulis explains, is that while these messages cannot be accessed by unwanted third parties, "a lot of metadata is still visible to the provider."

Metadata includes details such as who sent a message to whom, when the message was sent, the size of the message, and many others.

At the time of writing, Surfshark reroute your internet traffic via its own servers only. The provider is now looking if this new tech could fit in its product design roadmap. (Image credit: Surfshark)

This is where Surfshark's new patent comes in. Based on a distributed trust-based communication framework, it seeks to introduce a new way for VPN providers to handle encryption and de-centralize the ownership of the message.

Such a framework would involve two different VPN companies handling the encryption process so that no single entity has all the information in its entirety.

"The patented method would ensure that the information is split," said the patent inventor. "Thus, the metadata seen by the provider companies (as well as governments where they reside) is reduced."

Don't call it decentralized VPN

It's worth mentioning that some providers already offer decentralized VPN solutions that split users' information between several entities without having a single point of governance.

For example, the newly launched NymVPN is built on a decentralized server network run by anonymous users across the world. Obscura VPN employs a two-party VPN structure, using Mullvad's WireGuard VPN as an exit hop.

Kaciulis, however, thinks that Surfshark's patent goes beyond what a decentralized VPN stands for.

"Personally, I believe that today 'decentralized VPN' is a little bit of a buzzword used to convince users that it’s a better solution than the status quo. The Internet is built on trust and authority, and losing said authority only makes it less safe," Kaciulis told TechRadar.

Therefore, this solution comes as a way to emphasise the importance of having even more authority. "It’s just that the authority is shared between multiple actors instead of one."

You might also like

• Mullvad VPN brings Multihop to Android – promising to make online tracking even harder
• 4 reasons why Surfshark is still the best cheap VPN
• Metadata privacy matters – and this VPN promises to help

This $15 USB-C charger even comes with a cable in the box, too.

• Oracle Health has potentially suffered a second data breach
• Any such incident could affect sensitive patient data
• Oracle has not yet confirmed the scale of the breach

Oracle Health has denied having had sensitive patient data stolen by threat actors in two separate data breaches, leaving millions of customers potentially at risk.

The company had previously denied any breach after a hacker claimed to hold six million records belonging to the company but now a second incident appears to have led to a separate breach.

The company hasn’t yet commented on the compromises, but BleepingComputer has now reportedly seen private communications sent to impacted customers which confirm patient data was stolen.

Sensitive stolen data

The attack used compromised customer credentials to breach servers, and the legacy Cerner data migration servers sometime after January 22 2025, and the firm was made aware of the breach on February 20, 2025.

Reports confirmed patient information was included in the information stolen in the attack, and that the company will help identify the affected users. It’s not clear if this was the result of a ransomware attack, or if this was just data exfiltration, and it's also as yet unknown how the customer credentials were obtained.

The attacker, going by the name “Andrew”, has not claimed affiliation with any ransomware or hacking groups, and is demanding millions of dollars in cryptocurrency to stop the sale or leak of the exfiltrated information.

Healthcare organizations are increasingly at risk from cyberattackers, especially given the sensitive nature of the data they collect, and the often limited budgets for cybersecurity.

In fact, a 2024 breach of insurance firm United Healthcare impacted almost 200 million patients.

Since a data breach containing personally identifiable information such as this would put those exposed at serious risk of identity theft or fraud, Oracle Health has apparently offered to pay for credit monitoring services for those impacted.

"As cybersecurity leaders, we’re responsible for strong cyber hygiene: continuously monitoring our environments for unusual activity, leveraging cyber threat intelligence to stay ahead of emerging risks, and empowering employees to be our human firewall," commented Pierre Noel, Field CISO EMEA at Expel.

"No system is completely impenetrable, but understanding our risk landscape and layering defenses can make it much harder for attackers to succeed. Cyber resilience starts with us."

You might also like

• Take a look at our picks for the best malware removal software around
• Check out our choice for best antivirus software
• One of the most powerful ransomware hacks around has been cracked using some serious GPU power

Here are some hints and the answers for the NYT Connections: Sports Edition puzzle, No. 189, for March 31.

To say that Garmin has upset its users with its new subscription service would be putting it mildly. The company recently announced Garmin Connect+, a new paid premium tier of its Connect app that features some shiny new paywalled features, including AI-powered insights and a performance dashboard for all of the best Garmin watches.

It hasn't gone down well. Thousands of furious Garmin fans have taken to Reddit over the news, demanding customers take a stand.

The uproar has been so intense that you might have actually missed what the subscription entails, so if you're at least a little bit curious or you'd happily give the free trial a go, here's what you get from Garmin Connect Plus for $6.99 / £6.99 / AU$12 per month or $69.99 / £69.99 / AU$120 annually.

1. Active Intelligence

(Image credit: Garmin)

Garmin Connect+ gets you AI-powered training insights. Garmin says you'll get personalized suggestions based on health and activity data. Previews showcase friendly summaries about your sleep, battery, and more. As you'd expect from AI, the insights will reportedly get better over time. Here's an example:

You fell 1 hour and 23 minutes short of your sleep need, but you still managed to charge your battery 57 points. You're ready for exercise today, but balance it with rest to avoid bottoming out.

2. Performance Dashboard

(Image credit: Garmin)

Connect+ features a comprehensive view of training data, available on the website, that lets you compare your fitness and health data over time. Showcased examples include your running duration comparison, time in power zones, average pace, and distance over time.

3. Live Activity

(Image credit: Garmin)

Your smartphone will now give you real-time heart rate, pace, reps and videos for your indoor workouts (such as strength and yoga) when paired with a compatible smartwatch.

4. Exclusive Coaching Guidance

(Image credit: Garmin)

As you use Garmin Run Coach or Garmin Cycling Coach, you'll get expert guidance from coaches, including educational content and videos.

5. Enhanced Livetrack

(Image credit: Garmin)

Starting an activity on a compatible device will now notify your friends and family via text, and you can create a personalized LiveTrack profile page for people to follow.

This seems to have changed the existing text feature for Garmin users, and we've got some concerns that this might have pushed some of the existing LiveTrack features behind a paywall. We've reached out to Garmin to confirm.

6. Social features

(Image credit: Garmin)

Finally, social features include exclusive badges, some of them worth an increased amount of points, and badge challenges that you can use to personalize your profile.

Not a lot?

The Garmin Connect+ subscription isn't enormously expensive, but it's not cheap either. As you can see, you don't get an enormous amount for your money, and Garmin hasn't exactly been forthcoming on the details about these features – this is really all the information we have right now.

As such, it's possible we'll get more detailed insight about some of these features over the coming days and weeks.

On paper, Garmin Connect+ looks pretty thin, and while that's not the main reason people are objecting to its rollout, I have to imagine that it's a factor.

You may also like...

• Furious Garmin users revolt over new subscription service – "We need to take a firm stand"
• I love my Garmin watch, but there's one health feature I've always wished it had
• Garmin adds premium Garmin Connect+ tier with AI features – but promises your free experience ‘is not going away’

• IDC says AI demand drives hyperscalers to accelerate server infrastructure investments
• Super Micro challenges Dell with rapid revenue growth in 2024
• Nvidia leads GPU servers as AI adoption continues expanding rapidly

The global server market saw a record-breaking surge in the final quarter of 2024, with revenue nearly doubling compared to the same period in 2023.

New figures from IDC claim total server sales reached $77.3 billion in Q4 2024, marking a 91% year-over-year rise.

This rapid expansion, fueled by hyperscalers, cloud storage service providers (CSPs), and enterprises investing in high-performance computing, is expected to push the market past an annual run rate of $300 billion by 2025.

Record growth fuels server market expansion

A significant portion of this increase came from x86 servers, which saw a 59.9% revenue jump, while non-x86 servers experienced an unprecedented 262.1% year-over-year surge.

The growing adoption of AI-driven workloads has also contributed to the rise of servers equipped with embedded GPUs, with Nvidia maintaining a dominant 90% market share in this segment.

As AI integration accelerates, demand for high-performance servers, and the best dedicated server hosting providers, continues to reshape the industry.

Super Micro has also emerged as one of the fastest-growing server vendors, with its revenue surging by 55% in Q4 2024. With $5 billion in sales, it is now in a statistical tie with Dell Technologies, which saw a more modest 20.6% increase in revenue during the same period.

The company’s surge is largely fueled by demand for AI-optimized infrastructure, and best small business server solutions also favors other key players, including Hewlett Packard Enterprise, Lenovo, and IEIT Systems, all of which reported revenue growth exceeding 50% year-over-year.

“IDC expects AI adoption to continue growing at a remarkable pace as hyperscalers, CSPs, private companies, and governments around the world are increasingly prioritizing those investments,” said Lidice Fernandez, group vice president, Worldwide Enterprise Infrastructure Trackers.

“Growing concerns around energy consumption for server infrastructure will become a factor in datacenters looking for alternatives to optimize their architectures and minimize energy use”

Cloud service providers and hyperscalers accounted for nearly half of all server sales in Q4 2024, while ODM Direct category, comprising manufacturers that sell directly to large-scale operators like Amazon Web Services, Microsoft Azure, and Google Cloud generated a staggering $36.57 billion in revenue during the quarter, marking a 155.5% increase year-over-year.

This surge in spending reflects the increasing demand for the best web hosting and cloud computing solutions as organizations migrate more workloads to the cloud and invest in AI-powered infrastructure.

You may also like

• These are the best firewalls around today
• Take a look at the best business VPNs
• Thousands of websites have now been hijacked by this devious, and growing, malicious scheme

• DTS support seems to have been removed from 2025 LG TVs
• Almost all streaming services are Dolby, not DTS
• Your AV streamer or soundbar may have a DTS decoder

History appears to be repeating: LG has removed DTS support from its 2025 OLED TVs, including the LG C5, just like it did in 2020.

Noticed by TV reviewer Vincent Teoh (via FlatpanelsHD) reports, the EDID information on the LG G5 – the information that tells audio and video sources what formats a TV supports – doesn't include DTS audio.

When LG brought DTS back in 2023, we said that it was good news for home theater lovers, as it would give them support for content that didn't use Dolby Atmos. But it seems that according to LG's internal data, there weren't many such users or much of that content being played on LG TVs: Dolby was by far the most-used surround sound format.

That's largely because Dolby and Atmos is the preferred format of the best streaming services. The one notable DTS exception, the IMAX Enhanced streamed by Disney+, wasn't supported on LG TVs due to its particular variation on the format.

The 2025 models will be Dolby-only – but that doesn't mean you can't still enjoy DTS soundtracks.

If DTS is a dealbreaker, there are tons of C4 and G4 OLEDs still available – and for amazing prices now. (Image credit: Future) How to get around the lack of DTS on 2025 LG TVs

The issue here is all about decoding. DTS is a compressed audio format, and that means you need a device that can uncompress it and send it to one of the best soundbars, or your home theater speakers.

If your TV is removed from the equation, then it doesn't matter if it's supports DTS, though.

If your TV can't decode DTS, you may have other devices that can. So if your source device is connected to an AV receiver that's DTS-compatible, then you don't need to worry whether your TV has DTS: the decoding happens on your receiver, not on your TV.

With soundbars it's a little more complex, because you'll need a soundbar that supports HDMI passthrough as well as DTS so that the soundbar can handle the audio and send the video to the TV.

Many models don't offer that connectivity, though most high-end or mid-range models from Sony, Samsung and LG do. You can can connect one of the best 4K Blu-ray players or other DTS source directly to the soundbar's HDMI input and then connect the TV to the other port, and the soundbar will take and decode the DTS audio.

If that's not an option, there's one more thing you can do: if your LG TV (or any other TV) doesn't have the ability to decode DTS, but your 4K Blu-ray players does, then you can usually set the player to decode the DTS and to output uncompressed audio.

Try setting your player's audio output format to LPCM, often labelled Linear PCM. This format can be passed through the TV to your soundbar in surround sound without any concerns about format at all, though you'd lose the spatial sound of DTS:X.

It's a shame that LG has dropped this format, and we're not sure that LG TVs likely adding Eclipsa Audio in the future will really make up for it.

However, for the next few months, LG's 2024 TVs will remain on-sale, including the LG C4 and LG G4, and are arguably better value than its 2025 models – doubly so if you want DTS support…

You might also like

• Samsung promises to repair soundbars bricked by its disastrous software update for free – but it'll probably involve shipping
• Sony launches new version of the best cheap 4K Blu-ray player that drops the streaming tech – but the price looks odd
• Why AV receivers are still the heart of any good home theater

• A newly redesigned MacBook Pro is rumored to arrive in 2026
• If leaks are accurate, it will feature an OLED display and a thinner chassis
• This redesign will also be powered by Apple's M6 silicon

It could finally be happening, folks: the long-awaited OLED MacBook is finally on the way, if new claims from renowned Apple leaker Mark Gurman are to be believed.

Gurman has claimed in his latest Power On newsletter for Bloomberg that Apple is planning a major overhaul to the MacBook Pro in 2026, which will deliver several noteworthy changes – and chief among them is the introduction of a Tandem OLED display designed by Samsung, the same type seen on last year's new iPad Pro 13-inch.

It may also ditch the 'notch' in the screen for the webcam in favor of a floating 'hole cut' camera – perhaps adopting something more similar to the Dynamic Island seen on modern iPhones.

That's not all, though: Gurman's sources also state that the new MacBook Pro will feature a thinner design than previous models, stating that Apple "had once hoped to release this new version in 2025".

In other words, next year's MacBook Pro could look very different from the 14-inch M4 MacBook Pro we reviewed back in November. But how does Apple plan to make its pro-grade laptop even more compact without sacrificing power? The answer could lie with the M6 chip.

Powered up

Yes, Gurman also believes that these next-gen MacBook Pros will come featuring the M6 chip (presumably also with M6 Pro and M6 Max versions) - implying that we can expect M5 MacBooks this year, without the OLED upgrade.

The M6 chips will be the first laptop processors from Apple that utilize chip fabricator TSMC's new 2nm process (alongside the expected A20 chips for iPhones). The 2nm process is expected to bring new developments in both performance and power efficiency, which could explain how Apple expects the M6 chips to run at a high performance level even in a thinner MacBook.

As for the MacBook Air, there's no news on any potential M5 or M6 variants, but we have to assume that they'll be in the works; after all, the new M4 MacBook Air didn't arrive on the scene until several months after the M4 chip landed in MacBook Pro and iPad Pro models. However, a previous report we covered claims that an OLED MacBook Air is still on the way, though it won't use the same Tandem OLED panel as the Pro and will likely arrive in 2027 or later - something that would align with Apple's existing release cadence, if the OLED MacBook Pro does indeed land in late 2026.

You might also like...