TechRadar News

AI continues to spark debate and demonstrate remarkable value for businesses and consumers. As with many emerging technologies, the spotlight often falls on large-scale, infrastructure-heavy, and power-hungry applications. However, as the use of AI grows, there is a mounting pressure on the grid from large data centers, with intensive applications becoming much less sustainable and affordable.

As a result, there is a soaring demand for nimbler, product-centric AI tools. Edge AI is leading this new trend, by bringing data processing closer to (or embedded within) devices, on the tiny edge, meaning that basic inference tasks can be performed locally. By not sending raw data off to the cloud via data centers, we are seeing significant security improvements in industrial and consumer applications of AI, which also enhances the performance and efficiency of devices at a fraction of the cost compared to cloud.

But, as with any new opportunity, there are fresh challenges. Product developers must now consider how to build the right infrastructure and the required expertise to capitalize on the potential of edge.

The importance of local inference

Taking a step back, we can see that AI largely encompasses two fields: machine learning, where systems learn from data, and neural network computation, a specific model designed to think like a human brain. These are supplementary ways to program machines, training them to do a task by feeding it with relevant data to ensure outputs are accurate and reliable. These workloads are typically carried out at a huge scale, with comprehensive data center installations to make them function.

For smaller industrial use-cases and consumer industrial applications – whether this is a smart toaster in your kitchen or an autonomous robot on a factory floor – it is not economically (or environmentally) feasible to push the required data and analysis for AI inference to the cloud.

Instead, with edge AI presenting the opportunity of local inference, ultra-low latency, and smaller transmission loads, we can realize massive improvements to cost and power efficiency, while building new AI applications. We are already seeing edge AI contribute towards significant productivity improvements for smart buildings, asset tracking, and industrial applications. For example, industrial sensors can be accelerated with edge AI hardware for quicker fault detection, as well as predictive maintenance capabilities, to know when a device’s condition will change before a fault occurs.

Taking this further, the next generation of hardware products designed for edge AI will introduce specific adaptations for AI sub-systems to be part of the security architecture from the start. This is one area in which embedding the edge AI capability within systems comes to the fore.

Embedding intelligence into the product

The next stage in the evolution of embedded systems is introducing edge AI into the device architecture, and from there its “tiny edge”. This refers to tiny, resource-constrained devices that process AI and ML models directly on the edge, including microcontrollers, low-power processors and embedded sensors, enabling real-time data processing with minimal power consumption and low latency.

A new class of software and hardware is now emerging on the tiny edge, giving the possibility to execute AI operations in the device. By embedding this capability within the architecture from the start, we are making the ‘signal’ itself become the data’, rather than wasting resources transforming it. For example, tiny edge sensors can gather data from the environment that a device is in, leveraging an in-chip engine to produce a result. In the case of solar farms, sensors within a solar panel can specifically detect nearby arc faults across power management systems. When extreme voltages occur, it can automatically trigger a shutdown failsafe and avoid an electrical fire.

With applications like arc fault detection as well as battery management or on-device face or object recognition driving growth in this space, we will see the market for microcontrollers capable of supporting AI on the tiny edge grow at a CAGR of over 100% (according to ABI Research). To realize this potential, more work is needed to bridge the gap between the processing capabilities of cloud-based AI and targeted applications from devices that are capable of working on, or being, the edge.

However, like with any new technology: where there is a demand, there is a way.

We are already seeing meaningful R&D results focused on this challenge, and tiny AI is starting to become embedded in all types of different systems – in some cases, consumers are already taking this technology for granted, literally talking to devices without thinking ‘this is AI’.

Building edge AI infrastructure

To capitalize on this emerging opportunity, product developers must first consider the quality and type of data that goes into edge devices, as this determines the level of processing, and the software and hardware required to deal with the workload. This is the key difference between typical edge AI, operating on more powerful hardware, capable of handling complex algorithms and datasets, and tiny AI, which focuses on running lightweight models that can perform basic inference tasks.

For example, audio and visual information - especially visual - are extremely complex and need a deep neural architecture to analyze the data. On the other hand, it is less demanding to process data from vibrations or electric current measurements recorded over time, so developers can utilize tiny AI algorithms to do this within a resource-constrained or ultra-low power, low latency device.

It is important to consider the class of device and microcontroller unit needed in the development stage, based on the specific computational power requirements. In many cases, less is more, and running a lighter, tiny AI model improves the power efficiency and battery life of a device. With that said, whether dealing with text or audio-visual information, developers must still undertake pre-processing, feeding large quantities of sample data into learning algorithms to train the AI using them.

What’s on the horizon?

The development of devices that embed AI into the tiny edge is still in its infancy, meaning there’s scope for businesses to experiment, be creative, and figure out exactly what their success factors are. We are at the beginning of a massive wave, which will accelerate digitalization in every aspect of our life.

The use-cases are vast, from intelligent public infrastructure, such as the sensors required for smart, connected cities, to remote patient monitoring through non-invasive wearables in healthcare. Users are able to improve their lives, and ease daily tasks, without even realizing that AI is the key factor.

The demand is there, with edge AI and tiny AI already transforming product development, redefining what’s classified as a great piece of technology, enabling more personalized predictive features, security, and contextual awareness. In just a few years, this type of AI is going to become vital to the everyday utility of most technologies – without it, developers will quickly see their innovations become obsolete.

This is an important step forward, but it doesn’t come without challenges. Overcoming these challenges can only happen through a broader ecosystem of development tools, and software resources. It’s just a matter of time. The tiny edge is the lynchpin through which society will unlock far greater control and usefulness of its data and environment, leading to a smarter AI-driven future.

We feature the best Computerized Maintenance Management System software.

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

• The Golden Ticket is a new reality competition series coming to Netflix
• It's inspired by Roald Dahl's Charlie and the Chocolate Factory novel
• US residents can apply to be one of Wonka's guests and compete

There’s a magic in the very name of Wonka! And now Netflix has given a green light to The Golden Ticket, a reality competition series inspired by Willy Wonka, that legendary candy maker initially appearing in Roald Dahl’s 1964 children’s classic, Charlie and the Chocolate Factory.

As a precocious California child of the ‘70s back when the last Ice Age melted, celebrated author Roald Dahl’s imaginative kids’ book centered around an eccentric confectioner and a poor European lad who finds a Golden Ticket to tour Willy Wonka’s mysterious headquarters was always the first title I grabbed off my bedroom bookshelf whenever staying home from school.

Years later, I was enthralled to see my all-time favorite book adapted into a Hollywood feature film, Willy Wonka and the Chocolate Factory, starring Gene Wilder as the kooky cacao wizard. I even enjoyed (to a lessor degree) director Tim Burton’s Johnny Depp-led version from 2005 and Wonka, the most recent musical iteration by filmmaker Paul King starring Timothée Chalamet.

Netflix’s next appetizing new TV series based on Roald Dahl's story, aptly called The Golden Ticket, is best described as mixing the ingredients of logistical tactics and fun interactive gameplay, all while sugar-crazed contestants seek to achieve entry to a “retro-futuristic” candy-making factory using their Golden Ticket and negotiating through a number of chocolatey challenges to complete the various objectives.

“We are thrilled to bring the magic of The Chocolate Factory to life like never before,” said Jeff Gaspin, vp of unscripted material at Netflix. “This one-of-a-kind reality competition blends adventure, strategy and social dynamics, creating an experience that is as captivating as it is unpredictable. For the first time, a lucky few won’t just have to imagine the experience — they’ll get to step inside the factory and live it.”

With the profusion of popular cooking shows, obstacle challenges, and food-based reality programs scattered across the streaming landscape these days, a chocolate-coated project centered on Roald Dahl’s masterpiece seems destined for instant success. Eureka Productions (The Mole, Dating Around, TwentySomethings Austin) will serve as the series producers.

Netflix purchased the rights to Roald Dahl’s entire catalog of intellectual property back in 2021, which includes books such as The Fantastic Mr. Fox, Matilda, James and the Giant Peach, The BFG, The Witches, and the sequel to Charlie and the Chocolate Factory titled Charlie and the Great Glass Elevator. This endeavor will be Netflix’s first dip into the world of Willy Wonka.

Although there’s no official release date yet for Netflix’s The Golden Ticket, we’ll be sure to deliver the full scoop on any upcoming details and developments for what is sure to be one of the best Netflix shows.

You might also like

• Adolescence is the TV show of the year on Netflix
• 3 Netflix shows I stopped watching and wouldn't go back to
• Netflix has 8 new movies and shows with 100% on Rotten Tomatoes

• Mind Maps are rolling out to NotebookLM
• The new feature works in both the paid-for and free versions
• NotebookLM is shaping up to be one of our favorite learning tools

Hot on the heels of its announcement that NotebookLM's Audio Overviews are now available in Gemini, Google has revealed that a new feature, Mind Maps, will now be available as an option in NotebookLM.

Mind maps are great at helping you understand the big picture of a subject in an easy-to-understand visual way. They consist of a series of nodes, usually representing ideas, with lines that represent connections between them.

The beauty of mind maps is that they show you the connections between ideas in a way that helps make those connections more obvious.

Another string to its bow

NotebookLM is Google’s AI research helper. You feed it articles, documents, even YouTube videos and it produces a notebook summarizing the main points of the subject and you can chat to it and ask questions, as you would a normal AI chatbot.

Its best feature is that you can also generate an Audio Overview in NotebookLM, which is an AI-generated podcast between two AI hosts that discusses the subject, so you can listen to it and absorb the key points while doing something else at the same time. The Audio Overview can sound so natural it’s hard to believe you’re not listening to two humans talking!

Now Mind Maps have been added as another string to NotebookLM’s bow for helping you absorb information. They work in either the standard free version of NotebookLM or the paid-for Plus version.

(Image credit: Google/Apple) Better understanding

To generate a Mind Map you simply open one of your notebooks in NotebookLM, or create a new one, then click on the new Mind Map chip in the Chat window (the central panel).

Once you’re viewing your Mind Map (it appears in the Studio panel once it has been generated) you can zoom in or out, expand and collapse branches, and click on nodes to ask questions about specific topics.

NotebookLM is shaping up to be an essential tool for students who have a lot of information to digest, and don’t necessarily read very quickly. Using the power of AI you can get AI to do a lot of the leg work for you, then present you with the key bits of information, and Mind Maps is just another way for NotebookLM to help you on your path to better understanding.

You may also like

• From ChatGPT to NotebookLM, these are the 10 best free AI productivity tools to help you get things done
• I used NoteBookLM to help with productivity - here’s 5 top tips to get the most from Google’s AI audio tool
• I’ve found a new AI podcast creator, and it leaves Google’s NotebookLM in the dust

For years, the tech industry has been obsessed with the next ‘big thing’. Blockchain, AI voice assistants, the metaverse – each promise to revolutionize marketing as we know it. Venture Capitalists have poured billions into these dazzling innovations and marketers have scrambled to understand and implement the latest shiny platforms.

But as we look ahead, the real marketing revolution won't be found in the dazzling and disruptive. It will be found in the dependable, the practical, and yes, dare I say it, the ‘boring’.

Because in 2025, the biggest marketing trend won’t be about chasing fleeting novelty; it will be about mastering the essential, often overlooked, technologies that underpin true digital transformation. We're talking about the unglamorous heroes of modern marketing: composability, next-generation Content Management Systems (CMS), robust data strategies, sophisticated automation and laser-focused customer segmentation.

Why this shift? Because while the hype cycle churns, a critical mass of businesses are realizing that flashy gimmicks don’t deliver sustainable growth. They’re discovering that true market impact, the kind that drives revenue and builds lasting customer relationships, comes from a solid foundation of functional, reliable and adaptable technology.

The 93% Problem: Missing the Mark in Market Impact

Here’s a stark reality: 93% of companies are still missing out on significant opportunities to accelerate their market impact. Why? Because they’re hampered by outdated, inflexible tech stacks. They're trying to run a 2025 marketing strategy on systems built for 2015, or even 2005. Our recent survey of IT and marketing leaders found that traditional monolithic CMS platforms are falling short for teams today - nearly all expressed frustration with limited integration options with other services and tools in their tech stack, with 38% stating they constantly need a better integration experience.

Legacy CMS are like trying to run optimized code with outdated hardware - every move is held back by limitations. Developers are wrestling with monolithic platforms, spending countless hours on maintenance and workarounds instead of building innovative customer experiences. Last year alone, businesses shelled out nearly $3 million on tech upgrades, yet IT teams dedicated up to 25 hours per week simply maintaining these legacy systems. That’s a staggering waste of resources and potential.

These platforms simply cannot keep pace with modern demands. They struggle to integrate with AI tools, cross-platform solutions, or seamlessly connect with diverse digital channels. This isn't just an IT problem; it's a fundamental marketing bottleneck. It prevents marketers from executing agile campaigns, personalizing customer journeys, and delivering truly omnichannel experiences - the very things that drive competitive advantage in today’s market.

Startup Speed, Enterprise Security: The Achievable Ideal

Big enterprises often envy the speed and agility of startups. But the myth that large organizations must be slow and cumbersome is just that - a myth. By embracing ‘boring’ but brilliant technologies, even the largest enterprises can move with startup-like speed without compromising security or stability. The key lies in composable architectures and modern, headless CMS platforms. These are the ‘boring’ building blocks that unlock incredible agility.

Imagine a tech stack where you can pick and choose the best tools for each aspect of your marketing - from e-commerce and analytics to personalization and CRM -- and connect them seamlessly via APIs. This “composable architecture,” increasingly favored by 70% of retail decision-makers (up from just 44% two years ago), allows enterprises to break free from rigid, all-in-one legacy systems.

Speed-to-Market: The New Competitive Battlefield

In today's hyper-competitive landscape, speed-to-market isn't just an advantage; it's becoming the ultimate competitive edge. Customers expect instant gratification, personalized experiences and seamless interactions across every channel. If you can’t deliver at speed, you’ll be left behind.

Headless CMS is a game-changer in this regard. It fundamentally changes how websites, apps and digital experiences are built and managed. By decoupling the content repository from the presentation layer, headless CMS empowers developers to work with their preferred front-end frameworks like React, Vue, Laravel and Svelte. It’s about reducing friction in the stack, which means more time engineering features and less time managing legacy dependencies or maintenance tasks. This translates directly to faster campaign launches, quicker website updates and the ability to rapidly adapt to changing market demands.

Ending the Dev Bottleneck Marketers Despise

For too long, marketers have been locked in a frustrating dance with development teams. Every content update, every new campaign landing page, and every adjustment to the website often required developer intervention, creating bottlenecks and slowing everything down.

Headless CMS breaks down these silos. Its integrated workflows empower non-technical marketing teams to make real-time updates without disrupting site performance or requiring constant developer support. Marketing teams can make real-time updates while developers stay in control of the site’s performance. It’s like giving marketers their own sandbox to work in… so no one is stepping on each other’s toes.

This newfound autonomy is transformative. Marketers can be more agile, responsive, and creative, launching campaigns faster and optimizing content in real-time based on performance data. Developers, freed from constant content update requests, can focus on high-impact projects that truly drive innovation.

Omnichannel, Finally Realized

The promise of omnichannel has been around for years but for many businesses, it remains elusive. Legacy systems simply weren’t built for today’s fragmented digital landscape. With headless CMS though omnichannel delivery becomes a tangible reality.

Headless architecture allows you to create content in a central hub and seamlessly distribute it across websites, apps, digital displays, IoT devices and even voice-activated interfaces like Alexa. This centralized content management ensures consistency, optimized delivery and effortless scalability as new channels emerge.

The Smart Money is on ‘Boring’

The allure of the cutting-edge is undeniable. But in 2025, the savviest marketers will recognize that true competitive advantage isn't about chasing the flashiest new toy. It’s about building a robust, flexible and future-proof marketing infrastructure based on ‘boring’ but brilliant technologies like headless CMS and composable architecture.

It’s time to shift our focus from novelty to necessity, from hype to helpfulness. The future of marketing isn’t about the shiniest object; it’s about the strongest foundation. And in 2025, that foundation will be built on the power of ‘boring’ tech.

We feature the best online marketing service.

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

Full spoilers follow for Severance season 2 episode 10.

After 10 weeks of teases, never-ending fan theories, and shocking in-universe moments, Severance season 2 has ended.

And, boy oh boy, what a finale it was. The hit Apple TV show has been the talk of the town since its return in mid-January, and I don't expect discussions to end any time soon, especially after this episode.

As the dust settles, let's dissect what happened in Severance's season 2 finale, titled 'Cold Harbor'. Major spoilers immediately follow, so make sure you've watched it before reading on.

Is Gemma rescued by Mark in the Severance season 2 finale?

Devon and Harmony convince Mark S to help them rescue Gemma (Image credit: Apple TV+)

In short: yes! But, there's a large lead-up to that. So, strap in, folks, because this explanation is going to be a long one.

Picking up immediately after Severance season 2 episode 9, Mark's innie – Mark S – informs Harmony Cobel that he hasn't completed Project Cold Harbor (more on this later) yet.

Good, that means Gemma, i.e., the wife of Mark's outie (aka Mark Scout) is still alive and can be rescued! Not so fast, Mark S says. While Harmony, Mark Scout, and Mark's sister Devon have a plan to save Gemma, Mark S is extremely hesitant to play his part. If Gemma escapes and the world learns of Lumon's nefarious operations, the evil biotech corporation will be shut down.

That includes the severed floor, so the innies, Mark S included, won't exist anymore. Understandably, Mark S doesn't want to give his life and the lives of his fellow innies for a woman his outie loves, but he doesn't.

Handheld camcorders still come in *ahem* handy in the Severance universe! (Image credit: Apple TV+)

To convince Mark S that this is the right thing to do, Devon gives him a handheld camcorder, tells him to watch the video recording, and then record his response to it.

The message has been left by Mark Scout, who apologizes for using severance as an escape from Gemma's apparent death before asking Mark S to help break her out. The interaction is cordial at first, but things get intense as each persona becomes suspicious of the other's true motives.

Mark Scout makes the grave error of bringing up Helly R and calling her Heleny; a similar slip-up to one Helena made when she got Gemma's name wrong in Severance season 2 episode 6. That, alongside other issues, eventually leads Mark S to tell Mark Scout that he doesn't trust him and won't help save Gemma.

Mark S becomes increasingly emotional as he speaks to his outie throughout episode 10's first art (Image credit: Apple TV+)

Cue Harmony Cobel's intervention. Dropping some deep Severance lore by way of story exposition, she tells Mark S that the numbers he and the rest of the Macro Data Refinement (MDR) team have been working on are actually Gemma.

According to Harmony, they are a doorway into Gemma's mind, with each cluster correlating to one of the Four Tempers that Lumon founder Kier Eagan is said to have tamed decades earlier.

Anyway, Project Cold Harbor is the 25th and final file Mark S must finish to help Lumon complete their severance-based work. Oh, and once that's done, Lumon will dispose of the innies.

This particular revelation infuriates Mark S who, before storming out of the birthing cabin that he, Harmony, and Devon have been hiding at, informs the pair that, unless Mark Scout returns to Lumon, he won't see Gemma again.

Well, that's new... (Image credit: Apple TV+)

Next thing we know, Mark S is standing in the elevator leading to Lumon's severed floor. Exiting the elevator, Mark S is greeted by a giant, creepy mural of him completing Cold Harbor as numerous Lumon employees past and present, plus the Eagan family, watch on.

But Mark S doesn't have time to take it all in as he's quickly reunited with Helly. The duo head to MDR and, after reading a card from their manager Seth Milchick – held by a creepy-looking statue of Kier Eagan – Mark S restarts his work on Cold Harbor. He's incredibly reluctant to do so because he knows it'll mean the death of the innies but, after a tear-jerking heart-to-heart with Helly, he begins to process the final few number clusters.

Meanwhile, Dylan G is revealed to be alive. Remember, he tried to resign from Lumon after Gretchen, his outie's wife, rejected his marriage proposal. Episode 9 ended on something of a cliffhanger for Dylan G, with the implication that, by resigning, he was effectively committing suicide.

As I predicted, though, Dylan G can't simply walk away. Milchick meets him at the severed floor's elevator with his outie's response to his resignation. The pair clearly don't get on, but the heartfelt reply from Dylan's outie reignites the fire within Dylan G, who heads back to MDR to find his friends.

I've got to admit, I laughed out loud at this entire sequence (Image credit: Apple TV+)

But, back to Mark S and Helly R. As Mark completes the final file, Helly hands him Irving's note – i.e., the one that contains directions to the black hallway that leads to the testing floor, where Gemma is being held captive.

Before their rescue mission can begin, though, an amusingly surreal sequence distracts them. Milchick converses with the previously mentioned Kier statue (it's actually an animatronic) and then leads Lumon's Choreography and Merriment (C&M) division in a dance number that puts season 1 episode 7's Music Dance Sequence to shame. Hey, the Apple TV+ show was originally billed as a dark comedy, so absurd, laugh-out-loud moments like this are bound to happen intermittently!

To help Mark S slip away, Helly creates her own distraction by stealing Milchick's walkie-talkie and heading to the bathroom. He follows her but she manages to evade him and temporarily lock him in it. Milchick almost escapes, but Helly, with help from Dylan G and a vending machine, blocks the entrance, trapping him.

With no one to stop him, all Mark S has to do is follow Irving's directions, take the testing floor elevator, find Gemma, and help her escape, right? Well...

Who dies in Severance season 2 episode 10?

So long, Mister Drummond (Image credit: Apple TV+)

Don't worry, Mark S, Gemma, and the rest of MDR survive. In fact, only one person dies in 'Cold Harbor' and that's Mister Drummond.

Taking the testing floor elevator up to the severed floor after Mark S completes Project Cold Harbor, Drummond opens a secret room located behind the wall opposite the testing floor elevator hallway. There, he meets Mammalian Nurturables division chief Lorne and one of her goats (I'll explain what the goats are used for later).

"Where do you think you're going, Mark S?" (Image credit: Apple TV+)

Drummond and Lorne are interrupted by Mark S' attempts to break down the door to the testing floor elevator. His key card doesn't grant him access to it, you see.

Drummond investigates and, realizing what Mark S is trying to do, tries to stop him. A scuffle ensues and, after overpowering Mark S, Drummond starts strangling him to death.

Don't do it, Lorne! (Image credit: Apple TV+)

He would've succeeded, too, if not for Lorne's intervention. Remember, Mark S and Helly left a lasting impression on Lorne and her crew in Severance season 2 episode 3, so it feels right that she'd save Mark S' life.

A fight breaks out between Drummond and Lorne, which Lorne wins. Before she can kill Drummond, though, Mark S stops her because he can use Drummond as leverage to get to Gemma.

Hold on, Mark, aren't you going to cross a severance barr- oh, never mind... (Image credit: Apple TV+)

Or so Mark S thinks. After using Drummond's all-access key card on the hallway door, he holds Drummond at gunpoint as a hostage. As they travel down in the elevator, Mark S starts to explain what's going to happen when they reach the testing floor.

However, when the elevator passes through its severance barrier, Mark's outie re-emerges, causing him to pull the trigger on the gun, and the bullet *ahem* severs the artery in Drummond's neck. Long story short: Mark accidentally kills him.

Does Mark choose Gemma or Helly in the Severance season 2 finale?

Gemma and Mark, reunited at last – but not for long (Image credit: Apple TV+)

Above all others, this is the moment that's going to divide Severance fans. Why? Because Mark chooses Helly over Gemma.

Stepping out of the testing floor elevator, Mark's bloodied outie makes his way to the Cold Harbor testing room where he's confronted by Cecily, the nurse who keeps tabs on Gemma. Cecily refuses to open the door and runs away when Mark brandishes his gun.

How's he going to get in now? With Drummond's blood, of course. Like Cecily, Drummond's blood signature grants him access to all the test rooms and it just so happens that Mark is covered in it. So, he uses some blood from his tie to gain entry to the Cold Harbor test room.

Oh, Mark... (Image credit: Apple TV+)

There, he finds a terrified Gemma, who doesn't recognize Mark because her 25th (!) innie persona is active in this room. Despite Doctor Mauer's attempts to convince her to stay away from him, Mark manages to get her to leave with him. Passing through the test room's severance barrier, Gemma's outie re-emerges. Cue a soul-stirring reunion between the couple that won't leave a dry eye in the house.

Severance doesn't let us have happy things, though, so you know a gut-punch event is moments away.

Traveling back to the severed floor, Mark and Gemma pass through the testing floor's severance barrier, which causes Mark S and Ms Casey to take over. Mark S is still aware of the plan to get Gemma out, though, so he leads her to the severed floor's exit stairwell. Its doorway contains another severance barrier that, once crossed, will allow their outies to emerge again and finally be rid of Lumon.

Fly, you fools! (Image credit: Apple TV+)

And here comes the kicker. Mark S convinces Ms Casey to leave, which she does and then turns back into Gemma. However, when Gemma tries to get Mark S to follow her, he hesitates. That lets Helly R – after leaving Dylan and the C&M team to deal with Milchick – find Mark S and Gemma at the exit stairwell.

Helly's arrival reminds Mark S that he wants to be with her – after all, they love each other. As a confused and devastated Gemma watches, Mark S and Helly flee back through Lumon's labyrinthine halls to locations unknown.

What is Project Cold Harbor, exactly?

What does Gemma find in the Cold Harbor test room? (Image credit: Apple TV+)

One of the sci-fi mystery-thriller's most talked-about mysteries, Project Cold Harbor is finally solved in this episode.

After Gemma passes through the test room's severance barrier, Doctor Mauer tells her to walk into the room. There, she's greeted by a crib that's either the exact same one that Mark and Gemma were going to use for their baby or a near-identical replica of it. Either way, considering Gemma lost her and Mark's baby years earlier, it's disturbing and sinister that Lumon would make Gemma go through that again, albeit via her latest innie persona.

Well, that's not disturbing at all, Lumon (Image credit: Apple TV+)

Mauer instructs Gemma's newest innie to take the crib apart with a screwdriver. She duly obliges and, as she does so, Mauer watches her progress from the secret monitoring room we first saw in season 2 episode 7. Jame Eagen, Lumon's current CEO, also watches from his own monitoring room.

Later, Mauer asks if Gemma knows who she is, to which she replies "I don't know". That confirms Lumon's latest test is a success, but Mark's arrival puts paid to the final stage of whatever evil scheme they'd concocted.

What are the goats used for in Severance's season 2 finale?

I swear if anything happens to this little fella, Lumon... (Image credit: Apple TV+)

Ever since they were first teased in season 1, fans have longed to learn more about what Lumon Industries is doing with its goats. It's high time we found out, too. The last time we saw Lumon's goats was in season 2 episode 3 but, even then, we learned next to nothing about what they'll be used for.

Well, we finally have an answer – and, after a new Apple teaser confirmed they'd return in Severance's season 2 finale, it turns out one of the five best Severance goat theories was actually right.

During episode 10's final act, we learn that Lumon's Mammalian Nurturables division has hand-reared the goats for sacrificial offerings. Indeed, Lorne says as much after she meets Drummond in the secret room I mentioned earlier.

Justice for Emile! (Image credit: Apple TV+)

Drummond asks Lorne if the goat is "full of verve and wiles", to which Lorne says it has the "most of its entire flock". Interestingly, verve and wiles are two of the word-based cards we saw in the Cobel household in Severance season 2 episode 8, which signals their importance to the Kier Eagan doctrine.

Anyway, animal sacrifices have played an integral role in many real-world religions, so it's no great surprise that they do so in Lumon's religion-like cult. What is a surprise, though, is what purpose the goats appear to play as part of their sacrifice.

Drummond tells Lorne that the goat she's brought – it's called Emile, by the way – will be "entombed with a cherished woman whose spirit must be guided to Kier's door". The suggestion here is that, once the goat is sacrificed (read: shot in the head), it'll somehow be imbued with Gemma's spirit when she's also bumped off following Project Cold Harbor's completion.

Thankfully, Emile is spared by Mark S' inadvertent intervention, and Lorne teams up with Mark S to overpower Drummond. Will we ever see one of Lumon's goats get the chop? Hopefully not, but never say never.

Does Severance season 2's final episode have a mid-credits or post credits scene?

"What? Are you waiting for an end credits scene to play?" (Image credit: Apple TV+)

End credits stingers have become par for the course for many franchises, including the Marvel Cinematic Universe (MCU), as well as certain shows on the world's best streaming services, but not here.

Apple TV Originals steer clear of teasing what's to come next season, and that's certainly true of the Severance season 2 finale. There's no mid- or post-credits sequence to stick around for.

You should sit through the end credits to appreciate how many people worked on such a compelling series like this. But, if you're only hanging around for an end credits stinger, you'll be sorely disappointed.

Has Apple announced Severance season 3 yet?

Waiting for that official season 3 announcement like... (Image credit: Apple TV+)

Not yet, but it's a case of when, not if, Apple will renew one of the best Apple TV+ shows for a third season.

For one, Severance has become more successful with each passing week. The three-year gap between seasons 1 and 2 was excruciating, but that allowed more people to check out its debut season and become obsessed with its numerous mysteries.

With more viewers jumping on the bandwagon after the release of its sophomore installment, Severance is now more popular than Ted Lasso. The number of breakout Apple TV+ shows pales in comparison to some of the tech giant's streaming rivals, including Netflix and Prime Video. So, if Apple really wants to prove it's a serial entertainment industry contender, greenlighting new seasons for hit shows like Severance should be a foregone conclusion.

Tell us when you'll be back, Milchick! (Image credit: Apple TV+)

For what it's worth, the series' next entry is in the works. In February, director/executive producer Ben Stiller confirmed season 3's scripts are currently being written.

Unfortunately, there's not as much good news on the filming start date front. Dichen Lachman, who plays Gemma/Ms. Casey, exclusively told me that she doesn't know when principal photography will begin on Severance season 3.

More details could be revealed at a post-season 2 finale event in LA this Saturday (March 22), though. Speaking on the New Heights podcast, Stiller said: "Hopefully we’ll be announcing what the plan is very soon". I'm keeping my fingers crossed that it's sooner rather than later!

How does 'Cold Harbor' set up Severance season 3's story?

What will become of Jame and Helly after this season's finale? (Image credit: Apple TV+)

We don't know, but there are plenty of unanswered questions that'll need, well, answering.

For starters: where have Mark S and Helly gone? They can't leave Lumon as their outies will take over, but Lumon will be using all of its resources to track down the pair after events that transpired in 'Cold Harbor'.

Then there's the wider fallout and implications of what happened in this episode. Lumon should come down hard on Dylan G and C&M following their rebellion. Milchick will have something to say and/or do with them, too, but he'll also be in the firing line as this all happened under his watch. That's before we get onto how Lumon proceeds with the severance program and its malicious plans for the technology and those who signed up for the procedure.

Mark's problems are only just beginning (Image credit: Apple TV+)

That's just the tip of the iceberg. Will Gemma seek out Devon and Ricken, and try to enlist their help to rescue Mark? What's Harmony Cobel's endgame? What do the season 2 finale events mean for Jame in a leadership capacity? Is Reghabi going to show up again and attempt to help Gemma reintegrate? When will we see Irving and Burt again? And what is the Grand Agendum that Kier Eagan's animatronic mentions after Mark finishes Project Cold Harbor?

Those questions and many other mysteries need to be solved before Severance's end credits roll for the final time. For what it's worth, creator Dan Erickson "has a sense" of what the show's final scene will be and how many seasons would be "ideal" to reach it. Whether Severance has one or more seasons of story left to tell, then, is perhaps the biggest mystery of all.

• YouTube might be adding another feature behind its Premium paywall
• A new report says an audio quality control with three levels is arriving
• It could bolster the Premium feature set but continue the trend of putting more features behind a membership

YouTube is seemingly pulling out all the stops to remain at the top of the streaming game for both video content and YouTube Music and while it answered our requests for adjustable video quality a few years ago, the platform has yet to offer the same for audio.

However, this could be on the horizon for YouTube, as new hints point to a forthcoming feature that would allow you to control audio quality when watching videos.

Thanks to Android Authority, which has spotted new strings in the YouTube beta app, there’s fresh evidence that hints at YouTube’s next big upgrade. It would essentially give you the liberty to adjust the audio quality of whatever video you're watching.

These could come in three different options; Normal, YouTube’s standard audio, High, an improved bitrate option, and Auto, which could simply be an automatic setting depending on your internet speed. It seems too good to be true, doesn’t it? Well ,with YouTube, there’s always a catch.

Video quality settings are free to adjust for all users, but audio control may only be available to YouTube Premium subscribers. (Image credit: YouTube)

According to Android Authority’s findings, YouTube’s audio quality feature will only be available for those who are subscribed to YouTube Premium, and even then, there’s a possibility that this feature may only be applicable to certain videos in its endless library of content.

It’s hard to pinpoint when YouTube will launch this feature since it only exists as a few lines of coding at the moment, but if YouTube decides to proceed with it, it could be one of the platform’s most notable upgrades of the past few years.

It seems as though YouTube will do almost anything to get more people signed up for its YouTube Premium service, and these attempts to lure you in have been cropping up quite frequently. A few weeks ago, YouTube launched its cheaper YouTube Premium Lite tier in the US, packing ad-free content on ‘most videos’ but excluding offline or background video playback.

For as long as I can remember, adjustable video quality settings have been part of YouTube’s array of video enhancements, but they have had no effect on audio playback. The audio quality of YouTube videos has always depended on the uploader, so if the audio control rumors are true, it could do wonders to get more audiophiles to jump on the YouTube Premium bandwagon.

You might also like

• Gemini AI can now watch YouTube videos for you, and this changes everything
• I signed up for YouTube Premium, and it was the best tech decision I've made all year
• I was watching YouTube on my TV before it became more popular than phones – here are 3 reasons why it’s better on the big screen

• Hugging Face has launched HuggingSnap, an iOS app that can analyze and describe whatever your iPhone's camera sees.
• The app works offline, never sending data to the cloud.
• HuggingSnap is imperfect but demonstrates what can be done entirely on-device.

Giving eyesight to AI is becoming increasingly common as tools like ChatGPT, Microsoft Copilot, and Google Gemini roll out glasses for their AI tools. Hugging Face has just dropped its own spin on the idea with a new iOS app called HuggingSnap that offers to look at the world through your iPhone’s camera and describe what it sees without ever connecting to the cloud.

Think of it like having a personal tour guide who knows how to keep their mouth shut. HuggingSnap runs entirely offline using Hugging Face’s in-house vision model, smolVLM2, to enable instant object recognition, scene descriptions, text reading, and general observations about your surroundings without any of your data being sent off into the internet void.

That offline capability makes HuggingSnap particularly useful in situations where connectivity is spotty. If you’re hiking in the wilderness, traveling abroad without reliable internet, or simply in one of those grocery store aisles where cell service mysteriously disappears, then having the capacity on your phone is a real boon. Plus, the app claims to be super efficient, meaning it won’t drain your battery the way cloud-based AI models do.

HuggingSnap looks at my world

I decided to give the app a whirl. First, I pointed it at my laptop screen while my browser was on my TechRadar biography. At first, the app did a solid job transcribing the text and explaining what it saw. It drifted from reality when it saw the headlines and other details around my bio, however. HuggingSnap thought the references to new computer chips in a headline were an indicator of what's powering my laptop, and seemed to think some of the names in headlines indicated other people who use my laptop.

(Image credit: Hugging Snap Screenshot)

I then pointed my camera at my son's playpen full of toys I hadn't cleaned up yet. Again, the AI did a great job with the broad strokes in describing the play area and the toys inside. It got the colors and even the textures right when identifying stuffed toys versus blocks. It also fell down in some of the details. For instance, it called a bear a dog and seemed to think a stacking ring was a ball. Overall, I'd call HuggingSnap's AI great for describing a scene to a friend but not quite good enough for a police report.

(Image credit: Hugging Snap Screenshot) See the future

HuggingSnap’s on-device approach stands out from your iPhone's built-in abilities. While the device can identify plants, copy text from images, and tell you whether that spider on your wall is the kind that should make you relocate, it almost always has to send some information to the cloud.

HuggingSnap is notable in a world where most apps want to track everything short of your blood type. That said, Apple is heavily investing in on-device AI for its future iPhones. But for now, if you want privacy with your AI vision, HuggingSnap might be perfect for you.

You might also like

• Microsoft Copilot Vision is the perfect holiday shopping buddy, and it’s finally here
• ChatGPT adds eyes to its voice with new screen and video sharing feature
• Grok gets glasses to see what you're talking about

• Citroen boss says the future plan is to be “daring and shocking”
• As a result, the 2CV is poised to return as a modern EV
• Looking backwards is the new looking forwards

Following hot on the heels of the revived Renault 5 and its madcap R5 Turbo 3E big brother, Citroen is the latest brand to state that it is to remake more of its history.

Speaking to Autocar, Citroen's chief executive, Thierry Koskas, said the company would draw on “one of the richest histories in the world” among automakers and that the 2CV was one of the most widely recognized cars it had produced.

Stopping short of confirming a release date, Koskas claims that in the future, Citroen needs more iconic models that will “surprise.”

Earlier this year, Autocar also revealed that design work had already begun on the retro-futuristic 2CV. At the same time, Citroen’s brand chief said that we can expect a new concept car to arrive later this year - but it won’t necessarily be a preview of what we can expect from an electric Tin Snail, as the 2CV was affectionately known.

Old school is the new school of car design

(Image credit: Renault)

It is no coincidence that several European carmakers are delving into the history books for EV inspiration.

Currently, the threat from cheaper and more technologically accomplished Chinese competition is causing concern that buyers will be tempted to opt for the best value options as the cost of living continues to rise.

“But buyers still want good design,” Renault’s design chief, Laurens van de Acker, told me at the recent Dacia Bigster launch. “Design and heritage or having a story to tell,” he added.

It’s also no coincidence that the same European manufacturers are rebooting vehicles that were once regarded as practical, affordable people movers with the power to carry entire nations through hard times.

The Renault 5, for example, was born during the oil crisis of the 1970s, when people were crying out for a cheap and efficient set of wheels to use daily.

Similarly, the 2CV was designed to coax farmers away from horses, haul eggs over rough terrain, and generally act as the ultimate do-it-all vehicle. Arguably the world’s first SUV.

With combustion engine cars being phased out in many countries, customers are crying out for similarly affordable and practical options in the EV space.

You might also like

• My new favorite thing in EVs? This Citroen Ami that's been made to look like a 1950s H-Van
• Citroen reveals Europe's first 'affordable' electric car – and it has a decent range too
• Renault unveils its wildest EV to date and it comes with in-wheel motors and a rally-style vertical handbrake for drifting

• Bloober Team has unveiled its next game I Hate This Place, launching Q4 2025 for PS5, Xbox Series X, Xbox Series S, Nintendo Switch, and PC
• I Hate This Place is a survival horror game based on Skybound's award-nominated comic book series of the same name
• The studio also shared a sneak peek at Cronos: The New Dawn's upcoming dev diary

Bloober Team, the studio behind the Silent Hill 2 remake and Layers of Fear, has unveiled its brand new horror game I Hate This Place, as well as offered a new look at its upcoming title Cronos: The New Dawn.

During the Future Games Show Spring Showcase today, Bloober Team announced that the game is inspired by Skybound Entertainment's Eisner award-nominated comic book series of the same name, created by writer Kyle Starks and artist Artyom Topilin.

I Hate This Place is launching in Q4 2025 for PS5, Xbox Series X, Xbox Series S, Nintendo Switch, and PC and will be developed by Rock Square Thunder.

The first trailer revealed the game to be an all-new isometric survival horror game with a striking visual art style inspired by the 80s retro comic book aesthetics.

In the game, you play as Elena, whose fight for survival begins when she unwittingly awakens a malevolent force, the Horned Man.

"Outmatched and hunted in a hostile world, players must think fast, manipulating their surroundings to even the odds against formidable foes," Bloober Team describes.

Players will also engage in tense gunfights and melee combat "that rewards ingenuity over brute force", as well as craft essential tools to survive the wilderness long enough to reach safehouses before nightfall. There's a day and night cycle too that impacts gameplay dynamics, but players need to beware and avoid the Horned Man at all costs.

Bloober Team is also expected to launch its next game, Cronos: The New Dawn, this year for PS5, Xbox Series X, Xbox Series S, and PC.

To prepare for the upcoming release, the studio has also shared a sneak peek at the first episode of the game's upcoming dev diary; a behind-scenes video offering a more in-depth look at the survival horror experience with new gameplay footage.

This dev diary will explore the world of Cronos, game mechanics, as well as reveal more about the game's protagonist and the monsters players will encounter.

Cronos: The New Dawn was revealed in October 2024 and is described as a unique horror experience that blends mystery, time travel, and intense survival elements.

You might also like...

• Best horror games: the scariest titles to play in 2025
• The PS5 is currently selling faster than the PS4 did in the US, but I'm surprised to discover that the Xbox Series X and S are trailing behind Xbox One
• I was already sold on Assassin's Creed Shadows on PS5 Pro, but now the devs are teasing that the game will soon get a boost from PSSR

• Equal1’s Bell-1 quantum computer runs on 1600W and plugs into a standard power socket
• Deploying Bell-1 is as easy as setting up workstations
• Equal1’s system offers future-proof quantum power without infrastructure changes

Equal1 has introduced the Bell-1, the first quantum computer designed for real-world deployment, integrating seamlessly into existing high-performance computing (HPC) environments without requiring specialized infrastructure or extensive cooling.

Operating on just 1600W and plugging into a standard socket, it installs easily and stands as one of the best workstations, bringing quantum computing from research labs to practical applications in AI, financial modeling, and scientific research.

It's the first fully rack-mounted system that fits into standard data centers, deploying alongside classical hardware to become one of the most powerful computers for tackling complex problems.

A quantum server ready for any data center

Built on Equal1’s silicon-based quantum processor technology, Bell-1 functions as a plug-and-play solution that delivers quantum acceleration for advanced simulations, cryptography, and optimization without requiring infrastructure modifications.

Bell-1 overcomes one of quantum computing’s biggest engineering challenges by using a self-contained cryogenic cooling system to maintain its silicon quantum processor at 0.3 Kelvin (-272.85°C) without an external dilution refrigerator, making it an engineering marvel that operates efficiently within the noisy and thermally demanding environment of an enterprise data center.

Designed to work alongside classical HPC systems, Bell-1 integrates seamlessly into existing infrastructure with its 6-qubit UnityQ processor, handling specialized workloads that classical computers struggle with, making it a compelling option for industries requiring high-performance computing and a strong contender for the best small business servers as a future-proof quantum solution.

The system fits within a standard 600mm x 1000mm x 1600mm server rack and weighs approximately 200kg, making it comparable in size to existing HPC servers.

“Bell-1 represents a paradigm shift in how quantum computing is deployed and utilized. We’ve taken quantum technology out of the lab and into real-world environments where it can drive innovation. This is the dawn of Quantum Computing 2.0—where accessibility, scalability, and practicality take center stage,” said Jason Lynch, CEO of Equal1

“By eliminating the barriers of cost, infrastructure, and complexity, Equal1 is empowering businesses to harness the exponential power of quantum computing today—not in some distant future. Bell-1 is not just an advancement—it’s a revolution in computing.”

You may also like

• These are the best cloud computing services
• Infamous ransomware hackers reveal new tool to brute-force VPNs
• We've rounded up the best cloud hosting providers

• GMKtec EVO-X2 PC delivers flagship performance with the Ryzen AI Max+ 395
• Lisa Su’s autograph marks a milestone as AMD backs GMKtec
• GMKtec secures its place among top vendors

While major brands dominate the mini PC market, smaller vendors like GMKtec are gaining recognition, with the EVO-X2 establishing itself as a high-performance PC distinguished by its AI capabilities.

Per GMKtec, the EVO-X2 is the first mini PC to feature AMD’s Ryzen AI Max+ 395 processor, a 16-core, 32-thread CPU designed for advanced computing.

With a maximum clock speed of 5.1GHz and an 80MB cache, it delivers the power needed for demanding tasks.

AMD CEO Lisa Su signs the first EVO-X2 unit

Despite its small form factor, the EVO-X2 rivals some of the most powerful computers, with the integrated Radeon 8060S iGPU and 40 RDNA 3.5 compute units enhancing rendering efficiency.

Combined with the XDNA 2 architecture NPU, the system achieves 50 TOPS of AI computing power, making it well-suited for AI inference workloads and complex computational tasks.

In benchmark tests involving large-scale AI models, the Ryzen AI Max+ 395 outperformed NVIDIA’s desktop GPUs, including the 5080, making the EVO-X2 a compelling choice for users seeking the best video editing computer.

With its AI-driven architecture and high-performance specifications, the EVO-X2 is more than just a mini PC, offering a strong solution for professional workflows, creative applications, and AI development, making it one of the best workstations in its category.

You may also like

• We've listed the best business laptops right now
• We've also listed the the best mobile workstations around
• Microsoft is MIA as Amazon, Meta, Google and others join consortium to triple nuclear energy output by 2050

• New iPhone 17 dummy units have appeared online
• They show the glass and metal outlines for the Pro and Pro Max
• A substantial redesign is expected for the top two models

The season of iPhone 17 leaks and rumors is well and truly underway, and just days after dummy units of Apple's 2025 flagship phones leaked online, we've got a different batch to pore over – which shows the iPhone 17 glass and metal redesign in a bit more detail.

These new dummy unit images come courtesy of well-known tipster @SonnyDickson, and it looks as though we've got all four models: the iPhone 17, the iPhone 17 Air, the iPhone 17 Pro, and the iPhone 17 Pro Max (or iPhone 17 Ultra, perhaps).

Dummy units like the ones shown here are based on schematics sourced from the supply chain and give accessory makers a chance to get their products ready for launch day. They also allow the rest of us to check out phone designs in advance.

When it comes to the redesign of the iPhone 17 Pro and the iPhone 17 Pro Max – with chunky rear camera bars – it's a look that's been extensively leaked already. We can also see the smaller camera bar expected to appear on the back of the iPhone 17 Air.

New metal and glass design

Here's another look at some iPhone 17 dummies, Notice on the Pro models where the glass will change. pic.twitter.com/lJDc5KXsV9March 20, 2025

What these dummy units add, that previous leaks haven't really shown us, is the way that glass and metal will be combined on these iPhone designs. You can see a faint outline on the Pro and Pro Max models, just below the camera bar.

That outlined section will apparently be glass, enabling wireless MagSafe charging. The other section is rumored to be made of aluminum. That's quite a change from the iPhone 16 Pro and the iPhone 16 Pro Max.

Meanwhile, it looks like the design of the iPhone 17 will be along very similar lines to the iPhone 16, while the iPhone 17 Air is a brand new phone, replacing the iPhone 16 Plus – rumored to be super-slim – so we don't have any predecessor to compare it to.

We've also heard rumors of memory and camera upgrades for these phones, improvements in video recording, and perhaps a useful reverse wireless charging feature. The phones should make their debuts sometime in September, and we'll need to wait until then to see what actually comes to fruition when Apple fully unveils them.

You might also like

• The iPhone 17 Air could have an affordable price
• Why I want the iPhone 17 to get rid of the Dynamic Island
• Leaked iPhone 17 schematics show a speaker redesign

• SanDisk Slim Dual Drive offers 2TB storage, dual USB ports, and fast performance
• Fast read and write speeds enable smooth file transfers and backups
• Universal compatibility makes this SSD a versatile storage solution

SanDisk, known for its high-speed storage and portable SSDs, is set to launch the new SanDisk Slim Dual Drive, offering up to 2TB of storage with a sleek, stick-like form factor that eliminates the need for cables.

Per PC Watch (originally in Japanese), this device features both USB Type-C and USB Type-A ports, making it a versatile option for easy file transfers between older and newer devices without requiring adapters or dongles.

As one of the smallest external SSDs for Mac users, it delivers read speeds of up to 1,000MB/s and write speeds of up to 900MB/s, ensuring smooth performance for large file transfers, video editing, and data backups despite its compact size.

High-speed dual-port connectivity

Measuring just 80 x 18.4 x 10.7mm and weighing only 23g, this device features supports USB 3.2 Gen 2 for fast and reliable data transfer.

Beyond performance, SanDisk has focused on durability and security by equipping the drive with a metal body for added protection and a premium feel, while dedicated encryption software ensures sensitive data remains secure.

Set to launch on the 28th, the SanDisk Slim Dual Drive 1TB model is priced at 18,480 yen (approximately $134) while the 2TB variant is at 38,000 yen (around $256).

You may also like

• These are the largest SSDs and hard drives in the market
• Need a microSD card? Here are the best microSD cards for your microstorage needs
• How Google's new anti-scraping measures are forcing an industry evolution

• Bimawen B15.6 TV Pro integrates both smart TV mode and Google TV modes
• Works as both a monitor and media hub
• A budget-friendly monitor with a 60 Hz refresh rate and HDR

Several portable monitors from popular brands offer standard display features, but among the lesser-known options, the Bimawen B15.6 TV Pro integrates Google TV, a remote control, and an Android-based system.

Per Androidpc (originally in Spanish), it functions as a smart entertainment hub with built-in Chromecast, Wi-Fi 5, Bluetooth 5.0, and access to Google’s app store.

This business monitor includes 2 GB of RAM and 8 GB of expandable storage, making it a capable media device for streaming and apps.

A little-known brand with a unique feature set

This device features a 15.6-inch IPS panel with a 1920 x 1080 resolution, 16:9 aspect ratio, a 60 Hz refresh rate, a 178-degree viewing angle, and a 1000:1 contrast ratio.

While it may not be a high-end graphics solution, the inclusion of HDR400 support and an IPS panel could make it a budget-conscious monitor for digital design.

Designers and content creators can benefit from its accurate color reproduction and wide viewing angles, though the 6-bit + FRC color depth is not ideal for professional-grade work. Connectivity options include a mini HDMI port, two USB Type-C connectors, and a 3.5 mm audio jack, ensuring compatibility with a range of devices. With VESA support, it can also be mounted for a desktop setup.

Users can switch between monitor mode and smart TV mode via a remote control, enabling direct access to Netflix, YouTube, and other streaming services. Unlike a business tablet, it lacks a touchscreen and internal battery, reinforcing its role as a dedicated monitor.

The Bimawen B15.6 TV Pro is available on AliExpress for approximately 118 euros, with additional discounts using the promo code (ESAS20).

You may also like

• We've rounded up the best monitors for photo editing on the market today
• And here are the best monitors for programming we've found
• How AI-powered remediation can help tackle security debt

• Star Wars Andor creator Tony Gilroy has canceled plans to publish the scripts of the show
• The scripts from season one were set to release, but AI is the reason they're not being published
• It's part of a growing trend of the content that AI models are being trained on

AI companies will do anything to get their hands on material to train AI models - just look at what’s happening with music streaming services. It’s only a matter of time before these companies start targeting other areas in the entertainment industry, but the creator of one of the best Disney Plus shows is already taking a stance against the looming reign of AI.

Tony Gilroy, the creator of the Star Wars spin-off series Andor, has pulled the plug on his initial plans to publish the scripts from the show's first season out of fear that his work could be used as material for training AI models. In an interview with Collider, Gilroy shared, “I wanted to do it. We put it together. It's really cool. I've seen it, I loved it. AI is the reason we're not."

The decision to back out doesn’t come easy for Gilroy, who announced his plans to launch a website showcasing not only Andor’s written elements but also a selection of concept art back in 2023.

Still, despite this difficult decision ahead of the imminent release of Andor season two on April 22, it’s clear that Gilroy has grave concerns about the threat that AI has on creative individuals - and I’m glad to see that more artists are taking action to protest AI’s place in the entertainment industry.

(Image credit: Lucasfilm/Disney) I’ll say it again, we’re witnessing a cultural genocide

For AI, ‘permission’ is a word that ceases to exist. Gilroy is just one of the many concerned figures who have actively expressed a need to protect their work in a culture where AI is infiltrating streaming services daily – mainly for user experience reasons. Still, films and TV shows themselves remain open targets for AI to steal for training purposes.

While notable figures such as Christopher Nolan have called for tighter restrictions on AI, this is the first time I’ve seen the threat of AI cause a screenwriter to take action like this since the SAG strikes. But AI has only scratched the surface with film and TV - you get things like this all the time with music.

The Make It Fair campaign has pressured the UK government to ensure that creative rights are protected amidst its threatening AI proposal. (Image credit: Future)

A prime example is the silent protest album, Is This What We Want?, which was released in light of a UK government proposal allowing AI companies to use copy-righted music to train models and algorithms.

Artists from Kate Bush to Annie Lennox banded together to compile a silent album of ambient tracks of empty creative spaces. Though this won’t actively get the government to change its mind on the proposal, it’s succeeding in showing what it could mean for the future of the music industry while using the streaming profits to raise money for music charities.

Screenwriters and musicians alike should be able to share and publish their work without fear of it being used to enhance AI models. Some of us actually like to read the scripts behind our favorite shows.

I have no doubt that AI developers will eventually get bored of taking existing scripts and music to enhance algorithms and will then find a way to infiltrate the creative process altogether.

Though plenty of AI music software can already whip up a quick beat (thanks, Suno), how long will it be before AI-generated music albums go mainstream or when feature AI-generated movies become a thing?

Disney+ and Hulu ad-supported bundle: was $10.99 per month now $2.99 at Hulu and Disney+

This bundle is normally $10.99 per month, so you're saving 72% with this great offer. It's for the ad-supported versions of both services, and it lasts for four months, after which the price goes up to $10.99 per month if you don't cancel. This offer runs until 31 March, 2025.View Deal

You might also like

• Moana 2 achieves demigod status as the movie sets new streaming record on Disney+
• What time is Daredevil: Born Again episode 4 going to be released on Disney+?
• Disney+ renews Percy Jackson and the Olympians for season 3, unleashing an epic new odyssey based on the The Titan's Curse

• Kioxia launches 122.88TB SSD with PCIe Gen5 and dual-port support
• The LC9 Series NVMe SSD is designed for AI workloads and hyperscale storage
• The new drive comes in a compact 2.5-inch form factor

After nearly seven years at the top, Nimbus Data’s massive Exadrive 100TB 2.5-inch SSD has been dethroned by Kioxia, which has unveiled a new 122.88TB model that not only offers a higher storage capacity but also supports PCIe Gen5, a first for this category.

Several companies have previously announced 120TB-class SSDs, including Solidigm, but Kioxia's LC9 Series 122.88TB NVMe SSD stands out by pairing its ultra-high capacity with a compact 2.5-inch form factor and a next-gen interface with dual-port capability for fault tolerance or connectivity to multiple compute systems.

"AI workloads are stretching the capabilities of data storage, asking for larger capacities and swifter access to the extensive datasets found in today's data lakes, and Kioxia is ready to offer the necessary advanced technologies including 2 Tb QLC BiCS FLASH generation 8 of 3D flash memory, CBA and the complimenting AiSAQ," said Axel Störmann, VP & Chief Technology Officer for SSD and Embedded Memory products at Kioxia Europe GmbH.

Supporting AI system developers' needs

The 122.88TB SSD is aimed at hyperscale storage systems, AI workloads, and other data-intensive applications that rely on capacity and speed. There’s no word on availability or pricing yet, but the company does plan to showcase the new drive at "various upcoming conferences".

"This new LC9 Series NVMe SSD is an instrumental Kioxia product expansion that will support AI system developers' needs for high-capacity storage, high performance, and energy efficiency for applications such as AI model training, inference, and Retrieval-Augmented Generation on a vaster scale," Störmann said.

Reporting on the new SSD, ServeTheHome notes, “This is a hot segment of the market, and it is great to see Kioxia joining. As AI clusters get larger, the shared storage tier is usually measured in Exabytes. Teams have found that replacing hard drives with SSDs often reduces power, footprint, and TCO compared to running hybrid arrays. Moving from lower-capacity drives to the 122.88TB capacity in a 2.5-inch drive form factor really highlights the advantage of flash in these systems.”

You might also like

• These are the fastest SSDs you can buy right now
• And these are the largest HDDs and SSDs on the market at the moment
• Solidigm exits consumer SSD market as it prepares to go big on enterprise SSDs

• Nintendo is expanding its partnership with the baseball team the Seattle Mariners ahead of the Opening Season 2025
• The Nintendo logo will be featured on the team's home jerseys and a Nintendo Switch 2 logo on away jerseys
• Seattle Mariner centerfielder Julio Rodríguez will act as a brand ambassador to further the partnership

Nintendo is expanding its partnership with the Seattle Mariners ahead of MLB Opening Season 2025, by becoming the baseball team's first-ever official jersey sleeve partner.

Nintendo announced the new collaboration with a new promotional video released on X / Twitter today, showing Seattle Mariner centerfielder Julio Rodríguez sporting his new jersey with Nintendo's red logo on the sleeve.

The upcoming season is set to begin on March 27, aka Opening Day, and each member of the Seattle Mariners baseball team will be showcasing the Nintendo racetrack logo on their home jerseys, and a Nintendo Switch 2 logo on away jerseys.

"Nintendo and the Mariners have been inextricably linked since 1992," said Kevin Martinez, Seattle Mariners president of business operations, in a press release.

"Now, each time the Mariners take the field, our jersey sleeves will help serve as a reminder of all that Nintendo of America has done for the Northwest community and the team. There isn’t a better partnership in Major League Baseball. We are grateful for our incredible relationship with Nintendo."

Play ball ⚾We’re thrilled to share we’re expanding our long-standing relationship with the @Mariners and will be featured on their jersey sleeve this season!See it for the first time on Opening Day, March 27th! pic.twitter.com/Lig4Aw4bmfMarch 20, 2025

Julio Rodríguez will also further the partnership between the Mariners and Nintendo by serving as a brand ambassador.

"It’s a genuine thrill to be able to partner with such a talented team, including bright stars like Julio," said Doug Bowser, president and chief operating officer of Nintendo of America.

"Our history with the Mariners goes back decades so this really feels like coming home. Nintendo has a lot of exciting news this year and through collaborations like this one, we can’t wait to create even more smiles across all generations. Go Mariners!"

This expanded partnership comes just ahead of the Nintendo Direct Showcase on April 2, which will likely reveal a launch date for the Nintendo Switch 2.

You might also like...

• The best Nintendo Switch games to play in 2025
• Nintendo Switch 2 Direct - five things I want to see
• The PS5 is currently selling faster than the PS4 did in the US, but I'm surprised to discover that the Xbox Series X and S are trailing behind Xbox One

• New research claims downtime from ransomware attacks is costing government agencies thousands
• The average downtime after an attack is over 27 days
• RansomHub is the top offender

New research from Comparitech has outlined that a key underestimated aspect of ransomware is costing Government agencies thousands of dollars per day - downtime.

Of the over 1,000 confirmed ransomware attacks tracked, nearly a month’s worth of downtime was lost per attack on average, costing nearly $84,000 per day.

For government agencies, the average downtime is higher than other sectors, with healthcare averaging 16 days, and 12 days for manufacturers. Although fewer than half of all ransomware incidents end in payment, the total amount paid in previous years has hit over $1 billion, so attacks can be seriously expensive all-round.

Key targets

The research points to government entities being less equipped to overcome attacks, most likely from budget constraints leading to longer recovery times.

On average, attackers demanded $2.2 million in ransom, but it's very rarely confirmed whether a ransom is paid or not - and some governments like the UK have launched proposals which would effectively ban state run organizations from paying ransoms in order to dissuade groups from targeting them.

But this doesn’t stop attackers, with government agencies frequently topping the list of most popular targets.

Critical infrastructure organizations are increasingly at risk of geopolitically motivated attacks, looking to disrupt service providers and undermine trust in public institutions.

The report outlines that notorious group RansomHub is responsible for stealing the most data, with over 730,000 records taken since 2018. Close behind them is ALPHV with 700,000, and Brain Cipher with 650,000 records stolen.

Although it's a fairly new group, RansomHub continues to find success in attacking both public and private organizations, with the group claiming to have made a victim of MetLife earlier in 2025.

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

• Qobuz has published its average streaming rates
• Many rivals talk about the total, not what individual rights holders get paid
• At just over 1.8¢ per stream, Qobuz pays more than most

Streaming service Qobuz has done something none of its rivals have been so far willing to do: it's said how much it pays for streaming – and it's had its claims independently examined too.

Publishing independently verified figures is a first for an industry that prefers to talk about the big picture rather than the small details, and it's a very welcome development not just for artists, but for anyone who's trying to be careful about the impact of their money.

According to Qobuz, in fiscal year 2024 it paid an average of US$0.01873 per stream.

That money doesn't go directly to artists – streamers pay the rights owners, typically record companies and music publishers; what artists actually get from that depends on how good or bad their deals with those rights owners are – but it's a significantly higher average rate than other streamers are believed to be paying.

1.8 cents per stream might not sound like much, but Apple – one of the better payers – says it averages 1.0 cents. Some other streamers are believed to pay considerably less, and to offer reduced royalties in exchange for exposure.

(Image credit: Qobuz) Opinion: It's great to see rates, yeah!

It's not exactly easy in today's world, but many people do try to shop fairly ethically where they can – and knowing how much money's being paid to artists is useful if like me you want to support the people that make the music you love so much.

That's particularly true if the artists you love aren't megastars: 1,000 plays on Qobuz pays an average of $18.73, which isn't much. But it's $18.73 more than the rights owner gets on some other services. And that definitely makes me feel even more positive about Qobuz.

Transparency is something I like to see, and it's definitely a factor in the sites and services I choose to use. For example, I increasingly use bookshop.org and libro.fm for my book deliveries and audiobooks because I like how they work with independent bookshops and how transparent they are about who gets what.

The sums aren't much, I know, but it's nice to know that even when I can't make it to my favorite book nook I can still throw a little money their way instead of making a corporation even more cash.

Of course, people choose – or choose to avoid – the best music streaming services for all kinds of reasons. But with everyone essentially offering the same songs, this is another way in which streaming firms can differentiate themselves from their rivals by telling you what kind of company they are. More, please.

You might also like

• Pocket Casts makes its web player free to all
• Spotify seems set to let you speak to its AI DJ
• Qobuz proves once again it's the hi-res music lover's choice

• Apple is shaking up its Siri development team
• Mike Rockwell, the man behind Vision Pro, is taking over from John Giannandrea
• Could this be the pivotal change Apple needed to bring an upgraded Siri to market?

There's a major shake-up at the top of the Apple food chain as Tim Cook opts for a new leader to help the company bring Apple Intelligence-powered Siri to life.

Following reports from earlier this week that Apple Intelligence would be a focal point of Apple's off-site Top 100 exec meeting, Bloomberg's Mark Gurman is now reporting that AI head John Giannandrea is no longer going to be overseeing Siri's overhaul.

According to Gurman, "[Apple CEO] Tim Cook has lost confidence in the ability of AI head John Giannandrea to execute on product development, so he’s moving over another top executive to help: Vision Pro creator Mike Rockwell."

Rockwell will now be in charge of Siri, according to Gurman's sources, who have asked not to be named, and he'll report directly to software chief Craig Federighi.

With Giannandrea no longer in charge of Siri, he'll now be working on other AI projects. The decision was made during the Apple Top 100 meeting and is said to be confirmed to employees later this week.

Assuming Gruman is right, this major exec shift comes at a pivotal time for Apple – once a pioneer in the voice assistant game but now just a mere passenger – as it tries to devise a solution for Siri.

Apple had advertised its flagship iPhone 16 Pro as the best device for Apple Intelligence, yet nearly seven months after its launch, customers have yet to see the real benefits of AI.

With Apple's Siri delay a public disaster, this shift in leadership could be the catalyst for success that's needed to make the personal context-capable Siri a reality.

Siri's redemption arc?

At WWDC 2024 in June, we got to hear directly from Giannandrea about his vision for the then-newly-announced Siri. He said Siri with Apple Intelligence has a "rich understanding of what’s on your device," and that the voice assistant's knowledge base would "only get richer over time."

Nearly a year later, Siri still can't tell you what month of the year it is and definitely doesn't understand what's on your device.

Rockwell, who can take credit for Apple's Vision Pro development, has the potential to get Siri's development back on track. While the mixed reality headset hasn't necessarily been a commercial success it does achieve some incredible feats.

Tim Cook is likely counting on Rockwell's ability to innovate, as evidenced in the Vision Pro, to make Siri as good as the company advertised back at WWDC. And, if he does, then all the naysayers will have to accept that Apple truly is back on track.

You might also like

• I believed in Apple Intelligence, but Apple let me down
• Even without Apple Intelligence, the new iPad is still one of the best tablets you can buy
• The big Siri Apple Intelligence delay proves that maybe we really don't know Apple at all