TechRadar News

• ChatGPT might not be ad-free forever
• Head of ChatGPT Nick Turley says ads could be implemented in the future in some OpenAI products
• ChatGPT could remain ad-free, but at what cost?

The head of ChatGPT says OpenAI could implement ads on its products in the future, but only if they are "thoughtful and tasteful."

Speaking to The Verge, Nick Turley, the head of ChatGPT, was quizzed on the possibility of the AI chatbot getting ads in the future. While reluctant to respond, he did give an insight into how he views advertising for ChatGPT, and couldn't rule it out completely.

Turley said, "Look, since you’re really trying to get me to comment on ads, I have become humble enough not to make crazy, extreme, long-term statements on a question like that, because maybe there is a certain market where people aren’t willing to pay us, yet we want to offer the best, latest, and greatest. Maybe that would be a place to consider other indirect forms of monetization."

He added that if the company were to ever implement ads, they would need to do so in a "very, very careful and deliberate because I really think that the thing that makes ChatGPT magical is the fact that you get the best answer for you, and there’s no other stakeholder in the middle."

While that sounds pretty positive for the future of an ad-free ChatGPT experience, Turley couldn't rule it out altogether, instead opting for a more diplomatic approach. He said, "I’m humble enough not to rule it out categorically, but we’d have to be very thoughtful and tasteful about it."

Maybe more positive, however, is that he seems to think that if ads were to appear on OpenAI products, then it's likely to be on other offerings from the company, not the world's most popular AI chatbot with over 700 million weekly users.

He said, "We will build other products, and those other products can have different dimensions to them, and maybe ChatGPT just isn’t an ads-y product because it’s just so deeply accountable to your goals."

ChatGPT is ad free, but subscriptions will likely ramp up

While Nick Turley's comments leave me feeling hopeful for an ad-free ChatGPT experience for the foreseeable future, I feel less positive about an even more heavily tiered platform that paywalls its best features.

Just yesterday, ChatGPT started to roll out Gmail connectivity, but its best capabilities are paywalled behind ChatGPT Pro, a $200 / £200 a month subscription.

As a ChatGPT Plus subscriber who pays $20 / £20 a month, I've noticed the sheer difference in capabilities between the paid version and the free one, and while I understand OpenAI needs to make money as a business, $240 / £240 a year feels like a steep ask.

I'm worried about the future of AI being paywalled behind premium subscriptions, and I think it's the direction we're heading in. Even companies like Google have offered premium AI subscriptions with the best Pixel smartphones for a year; however, following that initial period, the same Gemini AI capabilities become paid.

If subscriptions and basic free tiers are the way AI companies justify ad-free chatbots, then I'm more pro-ads than I ever thought I'd be. In fact, give me an ad-riddled ChatGPT with full access to the same as a Plus account for free; I'd take that over forking out a hefty amount of money every single month.

You might also like

• GPT-5 free tier – 3 cool features you can try now for nothing
• GPT-5 is here – 5 things you need to know about OpenAI’s ‘most useful’ model yet
• Yes, Sam Altman will be played by Andrew Garfield in 'Artificial', but all we want to know is who’s playing Elon Musk

• Code in macOS Tahoe hints that a 5G MacBook could be on the way
• Apple is apparently testing it in a MacBook with M5 Pro chip
• There’s no guarantee a 5G MacBook will ever launch, though

Apple’s best iPhones and iPads have all been imbued with 5G goodness, but MacBooks remain left out in the cold. There are hints now, though, that that could soon change, and I think this could bring with it some serious benefits.

The idea of a 5G MacBook hasn’t been put forward by a dodgy leaker or an industry analyst – no, it comes from Apple itself. According to Macworld, Apple’s macOS Tahoe code contains references to an unreleased Mac codenamed “t6050” that will feature an M5 Pro chip, indicating that it’s going to be part of the next generation of MacBook Pro laptops.

Crucially, the code suggests that this MacBook will come with a “Centauri” chip, Macworld claims. Centauri is the codename for Apple’s 5G modem, which found its way into the iPhone 16e under the C1 name. If Macworld’s report is accurate, that would make the MacBook Pro with M5 Pro chip the first Apple laptop to feature 5G connectivity.

The timing makes sense, as Apple outfitted its iPhone 16e with its first in-house 5G modem last year. With more control over the chip’s features and manufacturing process, Apple might feel that the time is right to bring 5G capabilities to its laptop range.

And with the efficiency of Apple silicon added to the mix, there would be less concern over the chip unduly draining your laptop’s battery.

Connectivity off the grid

(Image credit: Apple)

On the face of it, adding 5G to a MacBook might seem like a strange idea. After all, no one is about to hold a clunky laptop up to their ear to make a phone call. And when I first heard this rumor, I was pretty unconvinced.

But the idea has been growing on me over the past few months. Far from being an anachronism, 5G could actually bring some pretty upsides to your Apple laptop, both in terms of connectivity and security.

For instance, it could help you stay connected to the internet in places where Wi-Fi is unavailable, such as when you’re traveling or off the grid. As well as that, it would mean you wouldn’t need to connect to free Wi-Fi in bars and restaurants, avoiding what can be a serious security risk to your laptop.

Still, there’s no guarantee that Apple will ever actually release a 5G MacBook. While it has been rumored for many months now, the company could simply be testing the device and ultimately decide not to launch it.

But with the M5-series MacBooks expected to arrive in either late 2025 or early 2026, we might find out relatively soon.

You might also like

• Apple could finally give us a MacBook that can get online from almost anywhere – but not until 2026 at the earliest
• Bad news, Apple fans – the M5 MacBook Pro may not launch until 2026
• The iPhone 16e’s 5G performance seemingly has the iPhone 16’s beat

• Windows 11 just got a new test build in the Release Preview channel
• It comes with a fair chunk of AI features, but only one I actually want
• That's the AI agent in Settings, which is rolling out to more people - but sadly, I won't get this, as it's for Copilot+ PCs only

Windows 11 just got a new preview release, and there's quite a bit of work with AI here, including a new home page for the Recall ability and a wider deployment of the AI agent in the Settings app.

This is preview build 26100.506,1, which has been pushed out to testers in the Release Preview channel, and as you might guess, much of the fresh introductions on the AI front are for Copilot+ PCs only.

One of the most important moves is that the AI agent in Settings - which works as a natural language search to find the options you want easily - is now rolling out to Copilot+ laptops with AMD and Intel processors.

Previously, this ability was only available to Copilot+ PCs with Snapdragon X (Arm) CPUs, but it's now functioning with all of these devices (though the feature still only works when your primary language is set to English in Windows 11). Sadly, I don't have a Copilot+ device, so I'm more than a bit jealous, as I feel like this is very much an AI feature worth having - and those are a rarity.

Speaking of which, still in Copilot+ PCs territory, the Recall feature - for those who've turned it on, and that won't be everyone, even though it's a core pillar of Microsoft's drive with AI - now has a new home page which is based on your recent activity.

So, this surfaces your most-used apps and websites, and will also present you with the latest snapshots taken so you can quickly get back into tasks you were previously engaged in (assuming you have snapshots set to be saved in Recall, of course). Essentially, it's a bank of useful and personalized shortcuts to get back into what you were doing in a swifter manner.

Along with this, Click to Do (AI-powered context-sensitive suggestions) now comes with a short tutorial (an interactive one, apparently). This shows how it works to offer shortcuts with images or text, a useful touch for those new to the idea.

All of this is for Copilot+ PCs, but there are some fresh AI additions coming to all Windows 11 systems. That includes AI actions in File Explorer, which are shortcuts in a similar vein to Click to Do, based around editing images or summarizing documents.

With images (JPG and PNG formats) in File Explorer - which is the app that displays the folders on your desktop - there are now four AI-related actions. That includes a Visual Search (image search on the web), and the Erase Objects ability, which is AI-powered object removal in the Photos app - plus there's the self-explanatory Blur Background option in that app, too. On top of that, Remove Background will do just that with an image in the Paint app.

In the case of summarizing documents, however, this functionality is restricted - you need a Microsoft 365 subscription to benefit from AI actions here.

Away from AI, there are quite a few other changes with this new preview build, and a whole host of minor tweaks all over the place - check out Microsoft's extensive blog post for the full details.

Other nifty non-AI changes worth noting include Windows 11 users outside of Europe getting the ability to fully customize which widgets appear on the lock screen (previously this was exclusive to the European Economic Area).

Windows Hello - the secure login system - now has a revamped interface, and a bug where facial recognition failed has been fixed (plus fingerprint login now works better after the PC wakes from sleep).

(Image credit: Foxy burrow / Shutterstock / Microsoft)Analysis: The importance of being artificially intelligent

The agent for changing Windows 11 settings is one of the more useful applications for AI that I've seen introduced by Microsoft, so while it's good to see it coming to Copilot+ PCs that don't have an Arm processor, I'm still jealous that I'm missing out. Yes, I get it - you need that powerful NPU to make it work - but still, it doesn't feel great to not be receiving one of the genuinely laudable AI concepts that Microsoft has on the boil.

At any rate, as this is the Release Preview channel, this means the AI agent should hit the finished version of Windows 11 soon enough for those with the necessary hardware.

The rest of us can feel like we're missing out, or maybe you won't even care about this - or the other steps forward Microsoft has taken with AI here. Okay, so it's not that there wasn't anything else for normal (non-Copilot+) Windows 11 machines in this preview, but a fair chunk of the major moves were to do with AI.

And it's likely that the focus on AI features here might elicit some groans from quite a few folks - but despite that, and the relatively niche appeal of Copilot+ devices still, we can expect a lot more of Microsoft pushing this kind of tech.

Recent visions of the future of Windows from a couple of Microsoft executives have underlined the importance of AI (as well as the cloud and voice commands for input), so it's clear that this is a path Microsoft is going to be forging down with even more vigor in the coming years. In short: you'd best get used to this.

You might also like

• Latest Windows 11 update packs a great new feature - but I hope I never have to use it
• Microsoft promises to crack one of the biggest problems with Windows 11: slow performance
• Microsoft has all but given up on Windows 11 SE – and it looks like the war against Chromebooks has been lost

• The Pixel 10 Pro Fold is again tipped to be IP68 rated
• It would be the first foldable to reach the standard
• The Pixel 10 phones are getting unveiled on August 20

We're patiently counting down the days until the big Google Pixel 10 showcase on August 20, and ahead of the launch event there's a rumor that the Pixel 10 Pro Fold will be the first foldable to launch with an IP68 rating for dust and water protection.

This comes from WinFuture, in an article where most of the handset's technical details are presented, and it's the IP68 rating that really stands out. No other foldable on the market can match it, not even the IP48-rated Samsung Galaxy Z Fold 7.

An IP68 phone is completely protected against dust getting in, and is watertight too: the exact definition depends on the phone and the manufacturer, but being able to survive for 30 minutes in water at a depth of 1.5 meters (nearly five feet) is typical.

The Google Pixel 9 Pro Fold offers IPX8 protection, like foldables such as the Motorola Razr 50 Ultra: this means a high level of waterproofing, but no guarantee that specks of dust and dirt aren't going to find their way into the housing.

A repeating rumor

Of course, a foldable phone with a flexible screen and a hinge presents some very specific challenges when it comes to keeping out dust and water, and it will be quite the feat if Google has managed to get the Pixel 10 Pro Fold up to the required standard.

Given that we heard the same rumor from separate sources in June and in July, it's looking more and more likely that Google has indeed pulled off that feat – and it's going to be one of the major selling points of the handset, on top of whatever other upgrades we get treated to.

Another interesting tidbit from this report is that the storage will be upgraded from UFS 3.1 to UFS 4.0. It won't make a huge difference, but it will mean a speed increase, and should make the foldable phone feel snappier overall.

We will of course be covering the launch of the Pixel 10 Pro Fold and the other handsets in the series live on August 20, and we're expecting Google to provide a livestream of some description, so you'll be able to watch along wherever you are in the world.

You might also like

• Here's what to expect from the Pixel 10 Pro Fold
• These are the best foldables you can buy
• I want to see even more AI on the Pixel 10 phones

• A threat actor is selling a huge database on the dark web
• They claim it was taken from multiple hotels in Italy
• The country's digital transformation agency confirmed the breach

A cybercriminal has managed to break into the booking system used by numerous hotels in Italy and steal highly sensitive information on thousands of guests, experts have warned.

Recently, a threat actor with the alias mydocs took to underground hacking forums to advertise the sale of roughly 100,000 individual identity documents, including passport scans, ID cards, and more, which they claimed to have stolen it from multiple hotels across Italy.

At first, the claims were met with skepticism from the wider cybersecurity industry, however Italy’s technical agency for digital transformation, AGID, has now confirmed the authenticity of the breach.

Abusing stolen data

At least ten hotels were struck, with the number possibly increasing in the coming weeks, AGID said, hinting it had “intercepted” an illegal sale of the documents.

"This data, once stolen, can be used for fraudulent purposes: from the creation of false documents to the opening of bank accounts, up to social engineering attacks and digital identity theft, with consequences for victims that can also be serious, both from an economic and legal point of view,” AGID said in a press release (machine translated) posted on its website.

It is possible that mydocs either inflated the numbers, or managed to steal years’ worth of sensitive data, since some of the hotels compromised have just a few dozen rooms. In any case, an official investigation is now under way.

The hospitality industry, given that it handles highly sensitive data, continues to be among the most targeted ones out there. Hotels and lodging, restaurants, event planning agencies, and tourism companies, are frequent victims of ransomware, impersonation, and data theft.

Victims are urged to remain vigilant with incoming communications, especially emails claiming to be coming from Italian hotels.

Via The Register

You might also like

• Millions of hotel users see personal info checked out in huge data leak
• Take a look at our guide to the best authenticator app
• We've rounded up the best password managers

• HBO Max has released one final trailer for Peacemaker season 2
• The latest teaser is full of gory and raunchy moments
• It includes a clip that appears to spoil a key story moment

Peacemaker season 2 is less than a week from premiering on our screens and, to celebrate its impending release, HBO Max has unveiled one final trailer that's definitely not suitable to watch in public.

Released yesterday (August 14), the red band trailer is full of the gruesome action and raunchy humor we've come to expect from R-rated projects with DC Studios co-CEO James Gunn at the helm. Indeed, from mutilated bodies and numerous swear words, to an orgy that looks like it'll rival The Boys season 3's 'Herogasm' episode for the amount of nudity on show, it's a sizzle reel that's best seen in private. (Warning: viewer discretion is advised).

With Peacemaker season 2 set to arrive on August 21 (North and South America) and August 22 (everywhere else), this is the last piece of footage-based promotional material we're likely to see.

It's full of interesting clips, too. The returning hit show's first two teasers put this season's multiversal elements at the forefront – and, in the process, did little to address DC fans' concerns over its place in Gunn and Peter Safran's rebooted DC Universe (DCU). The adult-oriented TV Original's final trailer doesn't ignore those aspects, either, but it emphasises Rick Flag Sr's pursuit of the titular character – real name Chris Smith – who, spoiler, killed Flag Sr's son in 2021's The Suicide Squad.

References to Gunn's Superman movie, such as Lex Luthor's near-world-ending plan and the threat that metahumans (read: superpowered beings) pose, are also placed front and center. It seems, then, that the plot of Peacemaker's second season will split its time between the dimension known as the DCU and whatever parallel realities its cast discover via the Quantum Unfolding Chamber (QUC) – i.e. the interdimensional storage unit found in the home of Smith's dead father.

Does Peacemaker season 2's final trailer ruin a major plot point ahead of the show's return?

Well, this doesn't look good... (Image credit: HBO Max/DC Studios)

All in all, the third DCU Chapter One project's final round of footage is getting a lot of love from viewers. However, I think a lot of people have skipped over the fact that Peacemaker 2's latest teaser appears to spoil a huge moment that could have a major impact on its story.

Potentially big spoilers follow for Peacemaker season 2.

The trailer's final clip begins with a four-eyed alien, which appears to be carrying a dead, hairless cast, opening its own interdimensional door to the QUC. As it enters this seemingly infinite realm, it sees Smith and his self-installed bestie Adrian Chase/Vigilante burning... something. Long story short: Smith crudely tells the alien to leave, which sparks a somewhat amusing conversation between Smith and Chase.

...this isn't what it looks like, promise! (Image credit: HBO Max/DC Studios)

Now, we don't know what the pair are doing, but I don't think they're having an interdimensional barbeque. In fact, I believe they're destroying evidence that could implicate them in someone's murder.

Think about it. The duo are wearing aprons and hazmat-style gloves covered in blood. As we learned time and again in season 1, Smith only asks for Chase's help when there's some form of dirty work to be done. Furthermore, the official trailer for Peacemaker season 2 showed Smith leading Chase through the QUC as the latter carried equipment that's usually reserved for cleaning, as well as a tense showdown with an alternate reality's Peacemaker.

To me, all of this evidence suggests Smith has accidentally killed someone. My belief is it's not just an ordinary person, either – indeed, I think it's the aforementioned Peacemaker.

Am I reading too much into this single clip? Perhaps but, before you dismiss the idea completely, let me show you part of the official logline for one of the best HBO Max shows' next entry: "In season 2, Peacemaker discovers an alternate world where life is everything he wishes it could be."

As it's implied, this other Smith/Peacemaker's life is markedly better than the DCU's version. The Smith we know could get jealous and, whether accidental or not, kill his multiversal variant, thereby allowing him to take the place of this parallel world's Smith, which is what season 2's various trailers have hinted at.

I guess we'll find out for sure once Peacemaker 2 launches with a two-episode premiere on HBO Max and more of the world's best streaming services in the near future. For now, get the lowdown on the series' return via my dedicated guide to Peacemaker season 2.

You might also like

• How to watch the DC movies in order: chronological and release date, DCEU films, James Gunn's DCU, and more
• When is James Gunn's Superman movie coming to HBO Max and other streaming services?
• Superman's most surprising cameo has got me asking 3 big questions about James Gunn's next DCU project

A single missile can cost millions of dollars and only hit a single critical target. A low-equity, AI-powered cyberattack costs next to nothing and can disrupt entire economies, degrading national power and eroding strategic advantage.

The rules have changed: the future of warfare is a series of asynchronous, covert cyber operations carried out below the threshold of kinetic conflict. Battles will still be fought over land, sea, and sky, but what happens in the cyber domain could have a greater bearing on their outcome than how troops maneuver on the battlefield.

We were always heading in this direction, but AI has proven a dangerous accelerant. The entire military industrial base must become fortified against these risks, and that starts with continuous, autonomous validation of its cyber security defenses.

Today’s adversaries, whether state-sponsored actors or independent cybercrime syndicates, are deploying AI-driven agents to handicap critical systems across the entire military supply chain.

Stop them with our cybersecurity tools now.

The Case for Autonomous Resilience

Today’s adversaries, whether state-sponsored actors or independent cybercrime syndicates, are deploying AI-driven agents to handicap critical systems across the entire military supply chain. These attackers aren’t focused on headline-making digital bombs, but a slow attrition, applying continuous pressure to degrade functionality over time. They’re also working anonymously: AI-enabled cyberattacks are executed by autonomous agents or proxies, making attribution slow or impossible.

Consider a hypothetical attack on the U.S. Navy. The Navy depends on a vast, decentralized web of small and mid-sized suppliers for everything from propulsion components to shipboard software systems. While these systems and suppliers may coalesce into the most technologically advanced Navy in the world, their interdependence is almost akin to human biology, in the way that a hit to one system can thoroughly destabilize another.

An adversary doesn’t need to breach the Navy directly. Instead, they can launch persistent cyberattacks on the long tail of maritime subcontractors, degrading national capability over time instead of in one massive, headline-making blow.

Third-party vendors, which often lack the financial resources to properly patch vulnerabilities, may be riddled with unsewn wounds that attackers can use as an entry point. But major security vulnerabilities aren’t the only way in. AI-driven agents can autonomously compromise outdated email systems, misconfigured cloud services, or exposed remote access portals across hundreds of these suppliers.

The impacts of these attacks can look like “normal” disruptions, the result of human error or some missing piece of code: delayed component deliveries, corrupted design files, and general operational uncertainty. However, the ill effects accumulate over time, delaying shipbuilding schedules and weakening overall fleet readiness.

Emerging threats

That’s not even accounting for sanctions. If equipment is damaged, and replacement parts or skilled maintenance teams are restricted, one attack has just crippled a nation’s chip manufacturing capacity—potentially for months or years.

These attacks also get smarter over time. AI agents are designed for continuous improvement, and as they sink deeper into a system, they become more adept at uncovering and exploiting weaknesses. The cascading damage limits recovery efforts, further delaying defense production timelines and dragging entire economies backwards.

Despite these emerging threats, most defense and industrial organizations still rely on traditional concepts of deterrents, built around visible threats and proportional response: think static defenses, annual audits, and reactive incident response. Meanwhile, adversaries are running autonomous campaigns that learn, adapt, and evolve faster than human defenders can respond. You cannot deter what you cannot detect, and you cannot retaliate against what you cannot attribute.

Facing such dire stakes, defense contractors must exploit their own environments before attackers do. That means deploying AI-powered agents across critical infrastructure—breaking in, chaining weaknesses, and fixing them—to achieve true resilience. If the window for exploitation narrows, and the cost of action rises. “Low equity” means little against a high chance of failure.

Leveraging AI in Proactive Defense

Fighting fire with fire sounds simple enough, but there are serious risks involved. The same AI tools that bolster organizations’ defenses against smarter, more covert attacks can also create new vulnerabilities. Large language models (LLMs) may cache critical weaknesses in their model architecture, and third-party components that contribute to the models’ effectiveness can also introduce new vulnerabilities.

Any AI-powered security tools should undergo a comprehensive vetting process to identify potential risks and weaknesses. Model architecture and history, data pipeline hygiene, and infrastructural requirements–such as digital sovereignty compliance–are all factors to consider when augmenting security with AI-enabled tools.

Even the cleanest, most secure AI program is not a failsafe. Defenders that rely too heavily on AI will find themselves facing many of the same problems that plague their counterparts who use outdated scanners.

A mix of false confidence and alert fatigue from automated risk notifications can lead to missed critical vulnerabilities. In a national security scenario, that can lose a battle. That can lose a war. Real, attack-driven testing makes up for where AI lacks, and when used in tandem with it, creates an ironclad shield against AI-enabled adversaries.

Artificial intelligence is a boon for society and industry—but it is also a weapon, and a dangerous one at that. Fortunately, it’s one that we can wield for ourselves.

• Microsoft Teams adding new Chat Notes feature
• New addition should help users collaborate better
• Only desktop Microsoft Teams users will initially get the new tool

Your workplace Microsoft Teams chats could soon actually be productive, thanks to a new service coming to the video conferencing platform.

The company has announced Chat notes, a "new collaborative space" where users will be able to share ideas and collaborate on work without having to leave Microsoft Teams.

Microsoft says the new feature will allow not just real-time online collaboration, but also will help keep your communications in check, and increase productivity across the board.

Microsoft Teams Chat Notes

In a Microsoft 365 roadmap post announcing the news, the company says Chat Notes will be all about, "enabling real-time collaboration, organization of key information, and streamlined communications."

The tool will be available in group chats along with 1:1 conversations, with the ability to add in extra team members when required.

The update post does not go into a whole lot of detail about what Chat Notes will look like, but does mention a few features, including the ability to add images, format text, share Loop components, and mention chat members when you require their attention.

Chat Notes is still listed as being in development for now, but has an expected release date of September 2025, so users should not have too long to wait to super-charge their productivity tools.

Upon release, Microsoft says Chat Notes will be available for Teams users on Desktop to begin with, with users across the world able to access it.

The launch will be the latest in a series of recent releases aimed at making Microsoft Teams users more productive when using the service.

This includes adding the ability for users to save messages in their conversations and chats, meaning users no longer have to trawl back manually through long conversations to find the message they need.

The company also revealed configurable keyboard shortcuts are finally coming to Microsoft Teams soon, giving users quick and easy access the symbols, icons, and others they may use regularly on a daily basis, but sometime are not close at hand when typing in a work environment.

You might also like

• Microsoft Teams is finally adding a tiny but crucial feature I honestly can't believe it never had
• Microsoft Teams will now let companies block people for good — so you'd better behave
• We've rounded up the best project management software around

Outlander: Blood of my Blood has burst onto Starz and Amazon MGM+ with all the vim and vigor of its predecessor Outlander, and we’ve only seen three out of 10 episodes. If you haven’t caught up yet, Blood of my Blood acts as a prequel to the main series, following Claire (Caitriona Balfe) and Jamie’s (Sam Heughan) parents, from World War I to the rugged Highlands of 18th century Scotland. They’re separate yet connected, in even more ways than we’re currently aware of.

As it stands, we’ve seen both pairs fall in love and l into their respective time travel journeys, but it’s Claire’s parents that I’m the most interested in. Henry (Jeremy Irvine) and Julia (Hermione Corfield) have fallen in love through a beautiful written correspondence before being brutally thrown back in time. They’ve been split up from one another along the way, finding their way back to each other while talking themselves into rewriting history in the process.

While I’m enjoying the hunt for Outlander Easter eggs (and there are a lot of those), I just can’t predict what we’re about to see in the next episode, let alone how the season will wrap up. But keeping one eye on the season 1 finale is exactly how my mind works, so I had to ask Irvine and Corfield about it when I had the chance. So, imagine my surprise when they told me they also have no idea what’s coming, and frankly, the reason why is an incredibly smart move from the producers.

Out of four potential endings, we have no idea which will air in the Outlander: Blood of My Blood season 1 finale

“We get given our scripts while we're filming the block before,” Corfield tells me, fresh from the set of Outlander: Blood of My Blood season 2. “We're currently doing block one and two [of season 2], and we won’t start to see scripts for block three until we're nearly done with block two, so you're constantly learning where your character is going as well. We actually have no clue what’s really coming.”

Irvine adds the punch: “We filmed about four different endings for season one, so we still haven't seen which one they've gone for in the final edit.” It’s the sort of thing RuPaul’s Drag Race would do to keep a season’s eventual winner a secret, filming every queen sobbing on the runway with a crown a sceptre.

This means that outside of the editing room, absolutely nobody knows how season 1 will end, and that’s refreshing. Alongside my fellow press chums, I’ve been lucky enough to have advanced Outlander: Blood of My Blood access, but that left us about four episodes short. In a time when we’re desperate to work out what’s coming, have all information at our fingertips and uncover leaks and spoilers before they hit the small screen, a move like this is now the only foolproof way to keep us in suspense, providing the watercooler moment that’s working so well for weekly releases. And for that, I tip my hat in respect to Starz.

Obviously, we can try to work out what’s coming thanks to the seven existing seasons of Outlander, clearly having to directly intersect with Outlander: Blood of My Blood at some point (especially if Claire is as close in time to her parents as we think). But Irvine didn’t realize the level of pressure that would come with that until season 1 started airing.

“Not until we started doing press,” he tells me about the comparison pressure. “We film it in this little bubble in Scotland, so it's very easy to forget that anyone's ever going to watch it. Now that we're promoting it, and we're filming season two, I'm looking at my scripts for season two going ‘Geez, gotta make sure I do a good job of that,’ because of the fan base.”

So, will there be any real clues to hang onto following the ending of Outlander: Blood of My Blood until we’re into the second season? Only time (travel) will tell.

You might also like

• Sullivan’s Crossing star says she’s game for a ‘fun’ Virgin River crossover episode and I think she’s a genius
• Netflix star rules out wild fan theory about her character in Sweet Magnolias season 5: 'I'm not wearing a pregnancy suit this season'
• I asked a major Netflix star if you should ditch the streaming service for Hallmark+’s romance dramas, and the answer might surprise you

What's the next Nintendo Switch 2 exclusive planned for 2025?

(Image credit: Nintendo)

The next Nintendo Switch 2 exclusive planned for a 2025 release is Kirby and the Forgotten Land Nintendo Switch 2 Edition + Star-Crossed World. Now that the console is out, you can check out our Switch 2 restock guide for info on where to pick one up. Elsewhere, there's TechRadar Gaming's Switch 2 review, where you'll find our final verdict on the new screen, controllers, mouse mode, and battery life.

Upcoming Nintendo Switch 2 exclusives will start to arrive thick over the next few months, with Kirby and the Forgotten Land Nintendo Switch 2 Edition + Star-Crossed World kicking things off in style now that the Switch 2 is out. Nintendo fans will have plenty to keep them busy for the rest of the year as well, with everything from a new Metroid Prime game, to a new Pokémon set to come out.

In terms of upcoming Switch 2 titles and Switch 2 launch games the next couple of months are absolutely stacked with Nintendo Switch 2 games. Some are exclusives like Kirby Air Riders, while others are new versions of old games like Elden Ring: Tarnished Edition. One thing's for sure, there will likely be some new contenders for our best Nintendo Switch 2 games list by the end of the year.

Here are the upcoming Nintendo Switch 2 exclusives to look forward to in 2025 and beyond. We'll only be covering games that have been officially confirmed for release, many of which already have solid release dates locked in. In terms of what games we've chosen to include here, these are either only coming to Nintendo Switch 2, new versions of Nintendo Switch games, or Nintendo Switch 2 versions of upcoming games. We'll indicate where each game sits with regard to these release platforms as we go.

Upcoming Nintendo Switch 2 exclusives

(Image credit: Nintendo)

There are currently quite a few Nintendo Switch 2 exclusives planned and confirmed for future release on the upcoming console. We've listed them below, alongside their release dates where available.

The games below include those only coming to Nintendo Switch 2, new versions of Nintendo Switch games, or Nintendo Switch 2 versions of upcoming games. We'll indicate where each game sits with regard to these release platforms as we go.

• Kirby and the Forgotten Land - August 28, 2025 (Nintendo Switch 2 edition)
• Story of Seasons: Grand Bazaar - August 28, 2025 (Nintendo Switch 2 edition)
• Kirby Air Riders - TBC 2025 (Nintendo Switch 2 exclusive)
• Hyrule Warriors: Age of Imprisonment - Winter 2025 (Nintendo Switch 2 exclusive)
• The Duskbloods - TBC 2026 (Nintendo Switch 2 exclusive)
• Hades 2 - TBC (Nintendo Switch 2, Nintendo Switch exclusive)
• Splatoon Raiders - TBC (Nintendo Switch 2 exclusive)

(Image credit: FromSoftware)The Duskbloods - TBC 2026

New blood

Let's kick off this list of Nintendo Switch 2 exclusives with perhaps the most surprising reveal of the latest Nintendo Direct. The Duskbloods is a brand new game from Bloodborne and Dark Souls developer FromSoftware. It's a Nintendo Switch 2 exclusive, and it'll launch sometime in 2026. Beyond that, we don't know too much about it, other than it. being described as a PvPvE multiplayer game. Whether this will be similar to how things work in Elden Ring, with players invading your game, or more akin to what Elden Ring: Nightreign is going for remains to be seen.

(Image credit: Nintendo)Hyrule Warriors: Age of Imprisonment - Winter 2025

Back to where it all began

Hyrule Warriors: Age of Imprisonment is a third-person action game that serves as a prequel to The Legend of Zelda: Tears of the Kingdom. It all plays out in that classic Dynasty Warriors style, with the story set around the Imprisoning War that kicks off Link's journey to save Hyrule with the help of the Zonai. From the looks of things, we'll be able to fight as Princess Zelda against large groups of Ganon's army. There's other playable characters too, like the masked champions from Tears of the Kingdom.

This one doesn't have a release date yet, only that it'll launch sometime in Winter 2025, exclusively for Nintendo Switch 2.

(Image credit: Nintendo)Kirby Air Riders - TBC 2025

The sequel to a GameCube hidden gem

Kirby Air Riders, a brand new title originated from the Nintendo GameCube classic racing-action game Kirby Air Ride, is coming to Nintendo Switch 2 this year – directed by Masahiro Sakurai. We don't know much about it yet, though it'll be a flying racing game starring Kirby and his pals for sure. The original featured items that transformed Kirby, giving him abilities like Link's sword. Hopefully this mechanic returns when the game launches some time in 2025.

(Image credit: Nintendo)Splatoon Raiders

Spin-off

Nintendo announced Splatoon Raiders via the Nintendo Today App, revealing a short trailer, as well as the fact that it'll be exclusive to Switch 2. Beyond that, we don't know much about it, other than it looks set to be a Splatoon spin-off, possibly with a cooperative survival game focus. We'll have to wait for more details on this one, as currently, it doesn't even have a release window.

(Image credit: Supergiant Games)Hades 2

Titan of Time

Finally, we have Hades 2, which will launch as a console exclusive on Nintendo Switch 2 and Nintendo Switch. Of course, it's already in early access on PC, but its full version will land first on Nintendo platforms. Aside from that we don't have a launch date, but we've certainly spent a lot of time with the early access version of Hades 2 already, and definitely vouch for it.

Nintendo Switch 2 exclusives - new versions of Switch games

(Image credit: Nintendo)

In addition to the Nintendo Switch 2 exclusives listed above, there's also Switch 2 versions of existing and upcoming Switch games to contend with. Here are the games being upgraded and receiving separate Switch 2 editions:

• Kirby and the Forgotten Land + Star-Crossed World - August 28, 2025 (Nintendo Switch 2 edition)
• Story of Seasons: Grand Bazaar - August 28, 2025 (Nintendo Switch 2 edition)
• Pokémon Legends Z-A - October 16, 2025 (Nintendo Switch 2 edition)
• Metroid Prime 4: Beyond - TBC 2025 (Nintendo Switch 2 edition)

With regards to Switch 2 versions of Switch games, TechRadar had access to Breath of the Wild and Tears of the Kingdom a month or so ago, as part of a wider Nintendo Switch 2 preview. Our Hardware Editor Rhys Wood had this to say:

"Playing it (Breath of the Wild) and Tears of the Kingdom at a rock-solid 4K 60fps on console feels like it’s not real. But it is on Switch 2. Both games also have drastically sharper image quality, meaning they’re absolutely the definitive version of each release."

Nintendo Switch 2 exclusives FAQ

(Image credit: Nintendo)Is The Duskbloods only on Nintendo Switch 2?

The Duskbloods is only planned for Nintendo Switch 2 as things stand. This means that it's not likely to arrive on any other platforms, even Nintendo Switch. By the looks of things, this is a collaboration between FromSoftware and Nintendo, so you'll need to pick up the new console to play it in 2026.

Will all new Nintendo games be Switch 2 exclusive?

While some new Nintendo games like Mario Kart World are indeed Nintendo Switch 2 only, there are still games coming to Nintendo Switch. Namely, Metroid Prime 4, and Pokémon Legends Z-A will release on both.

This will likely continue for a while, until eventually, things will move to be Switch 2 only. Even with a new console out now, you'll likely be able to play new games on your old Switch for a while yet, though you will miss out on the bigger titles like Donkey Kong: Bananza, and The Duskbloods.

Will Pokemon Z-A be Switch 2 exclusive?

Pokémon Legends Z-A is not a Nintendo Switch 2 exclusive. At launch, there will be separate versions for both Nintendo Switch, and Nintendo Switch 2. On the new console, it'll run at a better frame rate, and have improved visuals, but the game will be functionally the same on both devices. We now know it's coming out on October 16, 2025. Not long to wait!

You Might Also Like...

• The best Nintendo Switch games to play in 2025
• Without a fresh UI and customization features, the Switch 2 is Nintendo's most boring looking console ever
• Get ready, Tarnished! Elden Ring is coming to the Nintendo Switch 2 this year

• New Moss Green and Sand Shell colors for both wired and wireless ranges
• Also limited-run special edition colors
• From $1,599 / £1,299

When we reviewed (and awarded five sweets stars to) the KEF LS50 Wireless II speakers, we said they were "truly phenomenal" – in fact both the Wireless II and their passive, wired LS50 Meta siblings have garnered a plethora of five-star reviews for their spectacular sound. And now KEF has introduced two new colors for both models, thereby expanding the standard color options to five.

The new hues look quite autumnal to me: Moss Green makes me think of green glass desk lamps, leather chairs and a book-lined study while Sand Shell is more of an open-plan space on a crisp and sunny morning. Both are quite beautiful.

These new colors join the existing range of Carbon Black, Titanium Gray and Mineral White. And KEF also does limited-run special editions in other colors; at the moment that means the bold Crimson Red for the Wireless II and a sober Royal Blue for the LS50 Meta.

(Image credit: KEF)KEF LS50 Wireless II and LS50 Meta: key features

The two models have the same curved baffle and clever mid/bass driver with integrated tweeter, but the Wireless II adds built-in amplification, connectivity for TVs and consoles, plus wireless connectivity between the two speakers. That delivers up to 24-bit/96kHz resolution, and if you need even higher performance you can always connect them with the supplied cable to get 24-bit/192kHz.

Each tweeter in the LS50 Wireless II is powered by a 100W class A/B amplifier, which is designed to deliver a “smooth, incredibly detailed treble performance”, while the mid and bass drivers are driven by dedicated 280W class D amplifiers.

As we said in our KEF LS50 Wireless II review, "the LS50 Wireless II stereo speakers continue to provide the phenomenal audio quality offered by their predecessors – and then some."

Prices for the new colors are the same as for the rest of the ranges. For the LS50 Meta that means an RRP of $1,599 / £1,299 (about AU$2,700) per pair, and the LS50 Wireless II have an RRP of $2,999 / £2,199 (about AU$4,575).

You might also like

• Best stereo speakers 2025: stereo systems for all budgets
• The best wireless speaker, all tested by us
• Audio-Technica's cheap active bookshelf speakers look like the perfect buy for new vinyl fans

• Oracle is adding Gemini to OCI, starting with Gemini 2.5
• Further models, including specialized ones like MedLM, will also be added
• OCI recently signed a similar deal to add xAI's Grok, too

Oracle and Google Cloud have announced a deal to make Google's Gemini models available via Oracle Cloud Infrastructure's (OCI) Generative AI service.

Oracle confirmed the service would kick off offering Gemini 2.5, but the company would continue to add further models for text, image, video, speech and music generation as well as specialized models like MedLM for healthcare.

The company explained customers would now be able to use Gemini models – among others – to help them build their own AI agents for coding, automation, research and more.

Oracle Cloud now offers Gemini models

"Oracle plans to make Google’s entire range of Gemini models available via OCI Generative AI service through new integrations with Vertex AI," Oracle added.

Although the models will continue to run on Google's servers, they can be accessed through Oracle's cloud platform with customers paying with Oracle Universal Credits, but the benefits of this partnership will be twofold.

For Oracle, it means the company can expand its offering of models, including third-party options, rather than pushing its own in-house models, making it more attractive for broader audiences. With the deal, Google can also expand its enterprise reach, getting Gemini models into more hands.

"Oracle has been intentional in offering model choice curated for the enterprise, spanning open and proprietary models,” OCI President Clay Magouyrk explained.

The news comes around two months after Oracle struck a similar deal to make xAI's Grok models available to customers, and while many hyperscale cloud providers are battling it out to secure exclusive or preferential AI model partnerships (like Microsoft with OpenAI and AWS with Anthropic), Oracle's multi-vendor strategy could give it a more stable position in the future.

You might also like

• We've listed the best cloud computing platforms
• Check out the best AI tools and the best AI writers from a range of different models
• Oracle signs a mystery huge cloud contract - one which could net it $30bn in revenue

• Amazon has released the first official poster for Fallout season 2
• It confirms the Prime Video show's next entry will feature a key location from the games
• The series' next trailer and official launch date could be revealed very soon, too

We're less than four months away from Fallout season 2, but Prime Video has already started promoting the return of its hit TV series.

Fallout season 2 won't make its debut until December, but Amazon MGM Studios is kicking off its marketing campaign early with the unveiling of its first official poster.

A post shared by FALLOUT ⚡️ (@falloutonprime)

A photo posted by on

The artwork doesn't provide any new information about the hugely successful Prime Video show's return.

Sure, it confirms the show's lead trio in Ella Purnell's Lucy, Aaron Moten's Maximus, and Walton Goggins' The Ghoul – the last of that trio being accompanied by his trusty canine companion Dogmeat – will be back for more Wasteland-based adventures. The poster also reconfirms New Vegas, which was teased in the mid-credits scene of last season's finale and is the setting of its Fallout video game namesake's most beloved entry, aka Fallout: New Vegas, will be a key location in season 2. Those details aside, though, we know relatively little else about one of the best Prime Video shows' next chapter.

Well, for the time being, because I believe we're about to get some huge news about the Fallout TV show's forthcoming installment in the days ahead.

When could Fallout season 2's release date and first trailer be revealed?

Please don't make us wait much longer for more news on Fallout season 2, Amazon... (Image credit: JoJo Whilden/Prime Video)

I'll preface this section by saying I have no insider knowledge on these matters, so read whatever you want into what I outline below.

With that little bit of housekeeping out of the way, I'm convinced Fallout season 2's first trailer and official launch date will be unveiled on Tuesday, August 19.

There's evidence to back up my claim – the main piece being that the above Instagram post confirms season 2 will be part of Opening Night Live at gamescom 2025.

Europe's biggest video game convention runs from August 20 to 24 this year. However, with Opening Night Live set to take place on August 19, I'm expecting any members of the series' cast and/or crew invited onto the stage to announce an actual release date and, at the very least, drop some form of video teaser to whet our appetite for Fallout's return.

It makes perfect sense for a big announcement or two to be made at Gamescom. Microsoft, which owns Bethesda, aka the studio behind the Fallout game franchise, always has a major presence at Gamescom. Considering how well Fallout's first season was received, it's in Microsoft and Bethesda's best interests, as much as it is Prime Video's, to start ramping up excitement for its sophomore season. What better way to do that than at the last big gaming convention of the year?

But, what do you think? Will we get a launch date confirmation and trailer reveal next Tuesday? Or am I jumping the gun? Let me know in the comments. Then, read my dedicated guide for Fallout season 2 and check out the below section for more coverage of the incredibly popular Amazon TV Original.

You might also like

• Fallout season 1 is Prime Video’s explosively entertaining addition to the golden age lineup of video game adaptations
• One of Fallout season 1's most heart-breaking finale scenes was reshot because 'it didn't feel right', star says
• 'Forever changed': Amazon's Fallout cast discuss their characters' 'violent' season 1 voyages across the Wasteland

2025 has been an awesome year for movies. In the last month alone, we’ve been blessed with huge hits like Superman and The Naked Gun, alongside high-quality smaller fare in a range of genres like Nobody 2 (action/comedy), Eddington (comedy/western) and A Big Bold Beautiful Journey (fantasy/romance).

Prime Video's War of the Worlds aside, where 2025 has really shone, though, is in horror.

I'm often more lenient when judging horror films compared to other genres. As long as I get goosebumps and it isn't a complete mess, I'll generally have a good time.

Plenty of this year's horrors fit that description, from genuinely good flicks like Companion, Dangerous Animals and Final Destination: Bloodlines, to good-bad movies that I still had fun with like Clown in a Cornfield and Wolf Man.

Since the beginning of the year, however, there’d been one horror movie I'd been counting down the days to release – Weapons.

The new film is the first after director Zach Cregger's 2022 debut, Barbarian, which was one of my favorite movies of that year – and one of the best horror films in recent memory. So, when Creggers said Weapons "is more – and in a good way" it immediately shot to the top of my must-see list.

Finally hitting theaters last week, it didn't disappoint. Weapons is thoughtful, sad, and downright creepy. As opposed to many other good horrors, it's also a great movie in its own right.

While I didn't quite think it hit the heights that my colleague Lucy Buglass did – who awarded it five stars in our Weapons review and called it her favorite movie of the year – it’s undeniably one of the year’s most impactful films.

Another 2025 horror that's winning deserved praise is Together – and I absolutely agree that it’s a body horror flick that shouldn't be missed.

Only catching the trailer a few weeks earlier when watching I Know What You Did Last Summer (which I actually fell asleep during), I went into Together mostly blind – and had an absolute blast. It's one of those movies that makes you want to look away while not being able to, and the weirdly heartfelt ending was a pleasant surprise.

While I really enjoyed Weapons and Together and believe they both deserve the accolades they're receiving, they're far from my favorite scary movies of the year.

In fact, there are three horrors from 2025 that I liked more: my favorite movie of the year, my favorite horror of the year (yes, they're different!) and another that was just a darn good time.

The best part is you can avoid a bad audience that laughs at the wrong time or brings in noisy food and stream all three from the comfort of your home.

"You keep dancing with the devil... one day he's gonna follow you home"

(Image credit: Warner Bros. Pictures)

Without a shadow of a doubt, Sinners is my favorite movie of 2025 (so far). And, thankfully, its classification as a horror flick lets me include it here and talk about it.

I just watched it for the third time and loved it just as much as I did the first, if not more. Let's make this clear, though: this vampire flick isn't 30 Days of Night.

Directed by Ryan Coogler, who has sat at the top of my 'watch everything this person directs' list since Fruitvale Station, Sinners stars Michael B. Jordan as the Smokestack Twins – playing both Smoke and Stack.

At its face, the premise is simple. Set in the American South in the early 1930s, the two brothers arrive back in their hometown after years of war, adventure and making both money and trouble – and looking to hide from that trouble by opening up a 'juke joint', headlined by their younger cousin and gifted musician Sammie.

Only to find evil waiting for them.

Beyond its simple premise, Sinners is also a deliberate exploration of folklore, America's racial history, ancestry, and the liberating power of music.

While I question its status as a horror because I personally didn't find it that scary, a lack of spooks isn't a bad thing. For those after a gory vampiric slasher, you will get a hefty dose of gnarly horror action – but personally, my favorite moments in the film are found in the quiet, world-building first hour.

It's the banter between Smoke and Stack as they recruit their staff, how they treat their Sammie, any moment Delta Slim (Delroy Lindo) is on-screen and the music that I fell in love with.

Arguably worth watching just to hear Sammie sing – and to bask in Mike B's coolness – Sinners is available to stream on HBO Max, but you can also rent or purchase it on Apple TV or Prime Video, depending on your location.

"Some people believe the spirit stays in the body for months after death"

(Image credit: A24 / TCD/Prod.DB / Alamy Stock Photo)

Beating out Weapons, Sinners and Together for my number one horror of 2025 is Bring Her Back and, frankly, it isn't even close.

While the Philippou brothers' first film, Talk To Me, was a genuinely freaky, fun, and original take on possession, Bring Her Back invokes dread right from its opening scene, with no let-up for its 99-minute runtime.

The Australian film follows Andy and his blind younger step-sister, Piper, after the recent death of their father, whose body Andy discovered. The two siblings are incredibly close, but Andy can't become her legal guardian because he's only 17. So, after much convincing because of Andy's troubled past, they are placed into foster care together.

What follows is the foster mother's deep manipulation of Andy against Piper, as her sinister plot slowly unfolds.

It features some of the most cringe-inducing body horror I've ever seen, along with jumpscares, a supernatural element, and maddeningly real manipulation by a person in power.

Yes, I’ll admit it – it gets a few brownie points from me being a fellow Aussie, but if it's actual horror you're after, nobody is currently doing it better than the Philippou brothers.

Bring Her Back should be available on HBO Max in the coming weeks, but it's available to rent or buy on Apple TV and Prime Video right now.

"The more you kill, the easier it gets"

(Image credit: Sony Pictures Releasing)

On the other end of the spectrum from Bring Her Back, 28 Years Later is just a good, zombie-killing time. (And, yes, I know they aren't zombies.)

I should make it clear that I'm not a fan of the series. In fact, I made it through two-thirds of 28 Days Later before I decided I didn't want to waste a Friday night with a movie I wasn't enjoying, and I haven't seen 28 Weeks Later.

So, when my sister dragged me to see 28 Years Later starring Aaron Taylor-Johnson off the back of his Oscar-worthy performance in Kraven the Hunter, I didn't exactly have high hopes.

I don't love the over-the-top blood, the slow-mo kills, or the chaotic editing during action scenes. But I still thoroughly enjoyed the near-two-hour journey.

It follows a small village of survivors who live comfortably on a small island away from the mainland, focusing on Spike (Alfie Williams), his father Jamie (Taylor-Johnson) and his mum Isla (Jodie Comer), who has fallen ill.

While I loved the film for its fun action and any scene involving Ralph Fiennes (which I won't dare spoil), the story is also about a young boy braving unknown horrors to save his mother.

Still, in terms of zombie flicks that will have you watching the screen between your fingers, crying and cheering for the heroes at different times, 28 Years Later is an easy recommendation, and I'm looking forward to whatever this new trilogy may hold.

28 Years Later is available to rent or purchase on Apple TV and Prime Video. Sony Pictures films usually come to Netflix, so you can expect 28 Years Later on the red streamer at some point soon in the US, UK and Australia. I suspect it might come to Binge Down Under, too.

You might also like...

• I’ve picked the 5 best Netflix TV shows to supercharge your mood, mindset and overall wellbeing – including one with 100% on Rotten Tomatoes
• I watched Nobody 2 but this dull as dishwater comedy movie is no summer blockbuster and had me longing for Bob Odenkirk’s Better Call Saul days
• I initially ignored Your Friends & Neighbors, but this underrated Jon Hamm-led dark comedy is more proof Apple TV+ is in a league of its own

A single missile can cost millions of dollars and only hit a single critical target. A low-equity, AI-powered cyberattack costs next to nothing and can disrupt entire economies, degrading national power and eroding strategic advantage.

The rules have changed: the future of warfare is a series of asynchronous, covert cyber operations carried out below the threshold of kinetic conflict. Battles will still be fought over land, sea, and sky, but what happens in the cyber domain could have a greater bearing on their outcome than how troops maneuver on the battlefield.

We were always heading in this direction, but AI has proven a dangerous accelerant. The entire military industrial base must become fortified against these risks, and that starts with continuous, autonomous validation of its cyber security defenses.

The Case for Autonomous Resilience

Today’s adversaries, whether state-sponsored actors or independent cybercrime syndicates, are deploying AI-driven agents to handicap critical systems across the entire military supply chain. These attackers aren’t focused on headline-making digital bombs, but a slow attrition, applying continuous pressure to degrade functionality over time. They’re also working anonymously: AI-enabled cyberattacks are executed by autonomous agents or proxies, making attribution slow or impossible.

Consider a hypothetical attack on the U.S. Navy. The Navy depends on a vast, decentralized web of small and mid-sized suppliers for everything from propulsion components to shipboard software systems. While these systems and suppliers may coalesce into the most technologically advanced Navy in the world, their interdependence is almost akin to human biology, in the way that a hit to one system can thoroughly destabilize another.

An adversary doesn’t need to breach the Navy directly. Instead, they can launch persistent cyberattacks on the long tail of maritime subcontractors, degrading national capability over time instead of in one massive, headline-making blow.

Third-party vendors, which often lack the financial resources to properly patch vulnerabilities, may be riddled with unsewn wounds that attackers can use as an entry point. But major security vulnerabilities aren’t the only way in. AI-driven agents can autonomously compromise outdated email systems, misconfigured cloud services, or exposed remote desktops across hundreds of these suppliers.

The impacts of these attacks can look like “normal” disruptions, the result of human error or some missing piece of code: delayed component deliveries, corrupted design files, and general operational uncertainty. However, the ill effects accumulate over time, delaying shipbuilding schedules and weakening overall fleet readiness.

Emerging threats

That’s not even accounting for sanctions. If equipment is damaged, and replacement parts or skilled maintenance teams are restricted, one attack has just crippled a nation’s chip manufacturing capacity—potentially for months or years.

These attacks also get smarter over time. AI agents are designed for continuous improvement, and as they sink deeper into a system, they become more adept at uncovering and exploiting weaknesses. The cascading damage limits recovery efforts, further delaying defense production timelines and dragging entire economies backwards.

Despite these emerging threats, most defense and industrial organizations still rely on traditional concepts of deterrents, built around visible threats and proportional response: think static defenses, annual audits, and reactive incident response. Meanwhile, adversaries are running autonomous campaigns that learn, adapt, and evolve faster than human defenders can respond. You cannot deter what you cannot detect, and you cannot retaliate against what you cannot attribute.

Facing such dire stakes, defense contractors must exploit their own environments before attackers do. That means deploying AI-powered agents across critical infrastructure—breaking in, chaining weaknesses, and fixing them—to achieve true resilience. If the window for exploitation narrows, and the cost of action rises. “Low equity” means little against a high chance of failure.

Leveraging AI in Proactive Defense

Fighting fire with fire sounds simple enough, but there are serious risks involved. The same AI tools that bolster organizations’ defenses against smarter, more covert attacks can also create new vulnerabilities. Large language models (LLMs) may cache critical weaknesses in their model architecture, and third-party components that contribute to the models’ effectiveness can also introduce new vulnerabilities.

Any AI-powered security tools should undergo a comprehensive vetting process to identify potential risks and weaknesses. Model architecture and history, data pipeline hygiene, and infrastructural requirements–such as digital sovereignty compliance–are all factors to consider when augmenting security with AI-enabled tools.

Even the cleanest, most secure AI program is not a failsafe. Defenders that rely too heavily on AI will find themselves facing many of the same problems that plague their counterparts who use outdated scanners.

A mix of false confidence and alert fatigue from automated risk notifications can lead to missed critical vulnerabilities. In a national security scenario, that can lose a battle. That can lose a war. Real, attack-driven testing makes up for where AI lacks, and when used in tandem with it, creates an ironclad shield against AI-enabled adversaries.

Artificial intelligence is a boon for society and industry—but it is also a weapon, and a dangerous one at that. Fortunately, it’s one that we can wield for ourselves.

We list the best Large Language Models (LLMs) for coding.

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

• Framer introduced a new tool called On-Page Editing
• It allows anyone, with no prior knowledge, to edit websites in real-time
• Typos, content, new pages, can all be done quickly and seamlessly by any team member

One of the best website builders, Framer, just introduced a new feature that lets users update websites directly on the live page. Called On-Page Editing, the new tool is designed to have anyone, not just designers, make changes to websites quickly and safely.

In a press release shared with TechRadar Pro, Framer said that with On-Page Editing, users can fix typos, update text, swap images, and even create new pages without opening the design canvas, or navigating the Content Management System (CMS). Perhaps more importantly - they can do it without relying on someone else to implement changes.

“This isn’t just about making edits easier,” said Koen Bok, CEO and co-founder of Framer. “It’s about unlocking a whole new way to collaborate. With On-Page Editing, we’re laying the foundation for websites where designers build the system, but anyone can contribute with confidence.”

Removing bottlenecks

On-Page Editing integrates directly with Framer’s platform, offering single-click editing that syncs changes to the project, instantly. Rich text formatting, links, lists, and CMS page creation can all be done visually, without switching interfaces, the company explained. Framer also said that since teams can submit edits for review, bottlenecks between marketing, content, and design departments could be removed altogether.

Framer says that its end-to-end control of the stack, from canvas, across CMS, to hosting, allows the kind of workflow that keeps design integrity intact, while still allowing non-technical staff to contribute, and in real-time, at that. It expects template creators to benefit from the new offering as well, since they’ll now be able to create more static designs, while the flexibility of the system will enable anyone to edit and publish without coding knowledge.

On-Page Editing is available immediately for all paid Framer plans. Prices range from $75/month/site, to $200/month/site. There is a also the option of custom pricing for enterprise clients.

You might also like

• Framer aims to increase productivity but maintain "magic" with new website building features
• Check out our list of the best AI website builders
• We've rounded up the best password managers

With IT infrastructure growing more complex and teams under pressure to do more with less, it’s time for organizations to rethink their observability strategy before costs, burnout, and blind spots spiral out of control.

Across industries, legacy observability tools are buckling under the weight of today’s dynamic infrastructure. These traditional monitoring systems were designed for a world where environments barely moved, data trickled in manageable amounts, and collecting more metrics felt like progress.

But that era is long gone, and teams stuck in ‘collect everything’ mode are paying the price with runaway costs, spiraling complexity, and blind spots that turn small hiccups into full-blown outages.

In today’s fast-moving, containerized world, this strategy is backfiring. What once felt like a safety net has morphed into a data landfill, drowning teams in noise, burning them out, and surprising them with cost overruns that deliver the only visibility nobody wants: a meeting with the CFO to justify the bill.

The observability promise that fell apart

For years, engineering teams were sold a simple idea: more data meant more control. That advice made sense when infrastructure was static and applications evolved slowly – capturing everything often delivered the insights teams needed. But the rise of cloud technology changed everything, turning environments ephemeral and accelerating the pace of change and telemetry growth. Yet many teams still cling to the old ‘collect everything’ strategy, even as it drags them down.

Modern systems don’t wait. They scale instantly, shift constantly, and produce overwhelming volumes of telemetry. The tools that once brought stability are falling behind; they weren’t built for today’s level of scale or complexity. They’ve become rigid, noisy, and expensive, and the cracks are starting to show.

In sectors like aviation, even brief outages can result in millions of dollars in losses within minutes. Elsewhere, the fallout is just as real: frustrated customers, eroded trust, and reputational damage.

What once felt like a smart investment has quietly become a liability. Many organizations are waking up to the uncomfortable truth: their observability stack is no longer fit for purpose. Instead of becoming the true utility teams can rely on, it adds to the technical debt they’re actively trying to mitigate.

When teams can’t separate signal from noise, dashboards become cluttered with irrelevant metrics, alerts never cease, and real issues slip through the cracks. This constant stream of distractions imposes a steep distraction tax: every context switch, every false alarm, every hunt for meaning chips away at an engineer’s productive time and mental energy.

Over time, this chaos breeds reliance on tribal knowledge from a few seasoned ‘heroes’ who know where the bodies are buried. These heroes become the crutch that props up the system, celebrated for their late-night saves. However, a hero culture comes at a high price, with burnout, a lack of knowledge sharing, and stalled innovation, as teams spend more time firefighting than building differentiating features.

Observability should enable innovation, not kill it. When engineers are drowning in data without clarity, the best they can do is react. And in a world moving this fast, organizations that can’t move past constant triage will find themselves leapfrogged by the competition.

What does good observability look like?

Solving this problem isn’t just about new tools – it demands a strategic approach to your business pain. A strong observability strategy helps you deliver a better customer experience, enhance employee productivity, and increase conversion rates and revenue.

It delivers clear insights into the performance of your digital investments by revealing feature adoption trends, capacity and scaling gaps, and release quality and velocity issues. Done right, observability fuels a culture shift where teams embrace it as an enabler, not another distraction tax.

A clear telemetry collection methodology is essential to make observability a strategic asset rather than an operational burden. This methodology should be guided by well-defined Service Level Objectives (SLOs) and error budgets, which set the standard for what matters most to your business and customers.

By aligning telemetry collection with these objectives, you ensure your observability strategy surfaces only the data that helps measure and improve outcomes. This disciplined approach connects engineering efforts directly to business value, enabling teams to confidently invest in features, optimize performance, and scale systems without getting lost in the data deluge.

Even the best telemetry strategy will fail if observability is treated as an afterthought or siloed concern. Successful organizations make observability a shared responsibility by embedding it into team norms, workflows, and incentives. That starts with clear executive sponsorship to set expectations, coupled with training that gives every engineer confidence in reading, interpreting, and acting on telemetry data.

Organizational Change Management (OCM) practices help teams adopt observability incrementally, shifting the culture from reactive heroes to proactive, data-driven improvement. When observability becomes part of how everyone builds and operates software, it transforms from a distraction into a force multiplier for innovation and resilience.

Observability done right isn’t optional – it’s a competitive advantage. Teams that treat it as a strategic utility, guided by clear objectives and supported by a culture of shared responsibility, will outpace those stuck in reactive firefighting.

Now is the time to rethink your observability strategy: invest in disciplined telemetry, align it with what matters most to your business, and empower your teams to build with confidence. Strong observability turns bold strategies into market leadership and keeps your teams focused on the future.

We've listed the best business intelligence platform.

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

While unmanned aerial vehicles have been in operation for nearly 200 years, drones, as we know them today, were first used by the military in 2006 before technological innovations turned them into in-demand children’s toys.

Now, industries, including mining, oil and gas, ports, utilities, and public safety teams, are starting to recognize their capabilities. When used for surveillance or inspection, they can protect workers from entering hazardous areas and deliver greater situational awareness about operations or incidents when every second counts.

Today, the industrial drone sector is rapidly evolving, with 2025 marking significant advancements in technology and applications across various industries. Several countries are making drones even more attractive for industrial use, allowing them to be flown beyond visual line of sight (BVLOS) at low heights close to buildings and infrastructure.

Drones for safer, more efficient operations

The global drone market is projected to grow from $36.7 billion in 2024 to $44.32 billion in 2025, with industrial applications being a significant contributor to this growth. This has the potential to unlock new Industry 4.0 use cases for operational efficiency and worker safety improvements.

For example, at vast mining sites, drones can handle site surveys more safely, effectively and sustainably than on foot or by plane. They can be dispatched multiple times, equipped with payloads including LiDAR sensors, HD cameras, magnetometers and thermal imaging cameras to gain greater intelligence or terrain mapping

Once a mine is operational, drones can inspect waste and stockpiles, relieving workers, often in the harshest of temperatures. They can also more accurately and rapidly calculate stock volumes for better management. Operational disruption and worker hazards are reduced when drones are flown to inspect the impact of blasting.

Using drones to handle perimeter surveillance at large industrial sites reduces reliance on people, vehicle wear and tear and fuel consumption which contributes to industrial enterprises sustainability efforts. Oil and gas companies can use them to inspect storage tanks, pipelines, cooling towers and substations.

Dispatching drones from one or multiple operation centers ahead of first responders enables early assessment of a situation's severity, helping to support efficient resource allocation while also receiving AI-powered intelligence that can inform and accelerate decisions to help keep workers, property, and premises safer.

For instance, in a wildfire, drones can be flown lower than a manned aircraft, using thermal cameras to identify heat spots through the smoke. Using data accessed through drone flights, teams are better equipped to handle these situations and protect themselves, the community, and the environment. Today hundreds of US police departments have drone programs. Skilled police pilots operate drones for search and rescue missions, crowd and event monitoring, and various other tasks.

Integrating edge and AI technology for reliable drone flights

Connected to an edge and AI platform, drones can benefit industrial enterprises with automated activities that enhance efficiency and safety.

For example, leveraging AI, analytics and machine learning, changes in sensor data could trigger a drone flight – whether to monitor a perimeter breach, an equipment malfunction, or to notify the correct first responder teams. Using data and video from those drones, any team will be better informed, ensuring that the right people get to where they need to be faster with the right equipment.

To do this, industries must deploy an integrated digitalization platform that uses the right mix of technologies. That means reliable connectivity, real-time edge data processing, ruggedized drone hardware and an extensive range of software to enable new use cases.

To fly beyond BVLOS, connectivity must be robust, meaning seamless handovers at speed. Using Wi-Fi alone, as networks become overloaded, handovers between access points can be delayed, meaning that drones fail. However, by implementing a platform that integrates multiple technologies, including public and private 4G and 5G wireless and Wi-Fi, enterprises will benefit from redundancy and reliable connectivity even as the drone flies BVLOS.

An on-premises industrial edge processing solution allows data to be processed in real time and consumed by applications for new efficiencies. Real-time data will better inform teams of the latest situation, while enhancing productivity and ensuring the right people and equipment are in the right place at the right time.

Drone hardware must be ruggedized, built to withstand harsh industrial environments and weather conditions at sites including mines, agricultural areas and oil and gas facilities. A drone solution that can be used with a mix of payloads and software will unlock flexibility for any enterprise. Efficiency is enhanced further using drones that are charged remotely and flown further BVLOS.

The drone use cases will vary depending on the industrial enterprise's needs, but in most cases, they will allow them to achieve multiple goals simultaneously. For example, industries can access more accurate data for operational improvements while also protecting their workers.

We list the best Enterprise Resource Planning (ERP) 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

Super Apps are gaining traction globally, combining shopping, banking, messaging, and more into seamless mobile experiences. These platforms are reshaping how consumers live, engage, and transact, particularly in mobile-first environments.

At the same time, global economic shifts, including rising tariffs, geopolitical uncertainty, and growth plateaus in traditional markets, are prompting brands to explore opportunities in new markets. A pattern is emerging, with mobile-first businesses pushing to expand beyond their home turf, not just for growth but to reduce exposure to risk.

I refer to this emerging trend as internationalization. It is not a theoretical concept, but a real and active shift that we are already seeing play out. Brands are entering and testing new regions through mobile channels, especially in places where Super Apps have already become the central gateway to mobile life. Mobile apps offer a fast, scalable way to reach new audiences, particularly where mobile commerce dominates.

Understanding the intersection of these trends is crucial for anyone helping to shape how global brands will continue to grow.

Why internationalization is accelerating

Internationalization is no longer just a growth strategy. It is becoming a necessity. As trade restrictions increase and supply chains become more fragile, relying on a single market creates vulnerability. Regulatory shifts, currency swings, and platform volatility can all impact performance overnight.

Recent advertising patterns make this clear. In response to new U.S. tariffs, platforms like Temu and Shein have adjusted their strategies by pulling back ad spend in the region and ramping up investment elsewhere. Shein has reportedly increased its ad spend by more than a third in the UK and France, while Temu boosted its budgets by 40% in France and 20% in the UK.

Expanding into new regions spreads risk and opens access to mobile-first populations with advanced digital behaviors. And for good reason. Localizing an app experience is one of the fastest and most effective ways to test, learn, and scale.

The benefit is not just about protection. It is also about growth through diversification, reaching new audiences, unlocking cultural insights, and building relevance beyond your base market.

Super Apps as the platform advantage

In many high-growth regions such as Southeast Asia, Latin America, and South Asia, Super Apps are not just part of the consumer journey. They are the consumer journey. These platforms function as commerce, payment, and engagement ecosystems, often all within a single app.

Even in markets that are yet to fully embrace Super Apps, platforms are increasingly bundling services and moving toward more integrated, mobile-first experiences. For brands, this represents an opportunity to meet users where their digital lives are already converging and to shape that convergence in new regions.

Uber is a prime example of this evolution. It has moved far beyond ride-hailing to offer food delivery, groceries, freight, and even financial services − all from a single interface. This expansion has helped power Uber’s advertising business, which surpassed $1 billion in annual revenue in 2023, driven by rich first-party data and high-frequency user interactions.

For brands looking to expand internationally, these platforms offer direct access to engaged, mobile-native audiences. But success takes more than presence. It requires thoughtful localization aligned with local preferences, payment behavior, and platform expectations.

That might mean adapting checkout flows to local payment providers, surfacing regionally relevant promotions, or aligning creative with cultural shopping habits; all of which depend on having access to the right regional insights and data. When brands get this right, they can scale faster and more sustainably across markets that might otherwise be disrupted by economic or policy shifts.

Super Apps offer reach, but they are not plug-and-play. Brands need to treat them as local infrastructure, shaped by how people shop, pay, and engage in each market. Success depends on designing for that context, not just adapting language or layout.

Measurement with trust and depth

As Super Apps bring together commerce, content, communication, and payments, they create both a challenge and an opportunity for measurement. The customer journey becomes more consolidated but also more complex. Traditional attribution models often fall short when interactions span multiple functions within one app environment.

At the same time, these platforms offer something rare. Because users stay logged in and conduct a wide range of activities within a single ecosystem, Super Apps create a unified stream of high-quality, first-party data. When managed responsibly, this data can offer brands a richer, more accurate understanding of what drives performance across the entire customer lifecycle.

But this opportunity requires a shift in how measurement is approached. Brands need to move beyond fragmented analytics and build a framework that accounts for the full breadth of user behavior. This is not just about reporting. It is about harmonizing data across channels and touchpoints to create a more complete view of impact.

Equally important is how that data is handled. Privacy must be built into the infrastructure, not layered on after the fact. When a user's entire digital experience takes place within a single app, privacy is not just a compliance issue. It becomes an essential commitment.

For marketers operating inside Super Apps, measurement is no longer a technical task. It is a strategic imperative. The brands that win in this environment will be those that can gain from insights without compromising integrity.

Navigating complexity with strategic intent

Super Apps deliver scale but also come with complexity. Closed ecosystems can restrict access to data and reduce flexibility. Platform dependency raises questions about control and long-term resilience.

This is not a reason to hold back. It is a reason to engage strategically, with visibility, adaptability, and a clear approach to measurement and integration. Sustainable growth means knowing how to work within these environments while maintaining your own strategic independence.

Global expansion today is not about replicating what worked at home. It is about adapting, aligning, and operating with a deep understanding of where you want to grow next.

Compete with context

Super Apps are not just a trend. They are becoming the infrastructure for mobile-first economies and the entry point to digital life in many parts of the world. At the same time, internationalization is accelerating as brands seek new revenue, new users, and greater resilience.

Those who succeed will not just be available in a region. They will localize with purpose, measure with precision, and deliver value within the platform experience itself.

We've listed the best mobile payment app.

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 potential benefits and opportunities from AI agents in both our business and personal lives center around one word: productivity. Being able to do more with the help of AI seems like a no-brainer and a road to a less busy and hectic life.

But we’ve heard this before, haven’t we?

Like the introduction of the internet, smartphones, and app stores – all had the promise to simplify our lives and boost productivity. While they have had a profound impact on the way we work, it hasn’t changed how busy we are. In fact, Pew Research found that 47 percent of office-based employees said technology has actually increased their work hours. We always seem to fall back into the “busyness trap,” using technology to get things done faster, but filling our newly-found spare time by adding more menial tasks to our lists.

Why? Because we live and work in a world where being busy is equated with success. However, this new AI era presents a chance to approach things in a completely different way.

The fascination around AI being able to unlock new levels of productivity is absolutely warranted – it is an opportunity to help employees save time on administrative tasks and focus on delivering more strategic value through bigger, more impactful projects.

But in order to achieve that, we need to spark a mindset shift around what productivity truly means and to become comfortable with a “fewer but better” philosophy. Otherwise, we’re setting ourselves up to repeat the same busyness cycles we’ve seen before.

Business leaders now have the chance to rethink their approach, create an employee experience built on impact vs. to-do lists, redefine success standards, and optimize an AI-driven workforce. Here are a few ways to get started:

It’s time to redefine productivity

By automating routine and tedious tasks, AI adoption is opening the door to new levels of productivity, but not in the sense of checking more boxes. Real productivity means doing better, bigger, and more valuable things with less.

This requires a cultural shift. Leaders must be willing to rethink what productivity actually is. Allowing people to focus more time on more meaningful initiatives can result in higher value and better execution across their organizations. By fostering a culture that emphasizes meaningful results, quality outcomes, and impactful contributions, leaders can inspire employees to align their efforts with the broader goals of the organization.

Break the mold: normalize not being busy

“There are not enough hours in the day.” “I am working as hard as I can.” “I wish I could, but I’m just too busy.” We all know these familiar phrases because more often than not, we are all busy being busy.

Busy people are too often seen as important people, but with the help of AI, we don’t have to be so busy to be accomplished. AI offers an opportunity to focus on essential, strategic, and high-value work, but it’s all for naught unless we can alter the perception around not being busy. We have to move away from the idea that busyness is a badge of honor and instead embrace intentionality, balance, and strategic prioritization.

Look for ways to reduce the noise and sit in silence

One of the key opportunities of integrating AI into workflows is letting AI agents handle the “noise.” All of those small yet time-consuming tasks that are keeping us busy can be automated. Some examples include data entry, scheduling and calendar management, time zone coordination, invoice generation, and more. Letting AI eliminate the noise of our everyday seems easy enough, as long as we don’t replace them with more tedious tasks.

Being OK with allowing AI to take over the mundane and not replacing that time with other filler tasks is a skill business leaders will need to adopt, foster, and encourage their teams to exhibit as well. It requires continuously re-evaluating the essential vs. non-essential tasks, and assigning the non-essential to AI agents. A day with three strategic priorities completed can be just as – if not more – successful as a day with ten or more to-dos checked off.

Move away from multi-tasking

We’re now so used to constant context switching between devices, screens, conversations, and projects. The idea of being productive has centered around accomplishing multiple things, and that is traditionally done by multi-tasking. Although constantly switching between tasks may seem productive, it often results in mental exhaustion, diminished concentration, and a lower quality of work.

But with the growing adoption of AI agents, we have the opportunity to lean into hyper-focused productivity and hone in on specific initiatives that require and deserve our undivided attention. Over time, this focused approach is what drives more meaningful outcomes, greater personal fulfillment, and material improvement in productivity and effectiveness.

Embrace a better way forward

Despite the potential of past technological advancements, the reality is they haven’t alleviated our busy lives as we hoped. However, AI agents offer an incredible opportunity, not just to do more, but to do better. But business leaders need to be willing to redefine cultural and workplace norms.

By focusing on quality over quantity, reducing the constant noise, and moving away from multi-tasking, we can unlock the real promise of AI and create a future where productivity energizes rather than exhausts the workforce, all resulting in better business outcomes.

We've listed the best productivity apps for iPad.

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