Technology

• Experts observe a 19x quarter-over-quarter rise in .es usage for malicious campaigns
• 99% were credential phishing attacks, with 1% relating to remote access trojans
• Microsoft was by far the most commonly impersonated brand

Cybersecurity experts from Cofense have revealed a 19x increase in malicious campaigns using .es domains between Q4 2024 and Q5 2025, making it the third-most abused top-level domain (TLD) after .com and .ru.

Typically reserved for businesses and organizations in Spain, or Spanish-speaking audiences, researchers found nearly 1,400 malicious subdomains across nearly 450 .es base domains between January and May.

An overwhelming majority (99%) of the campaigns involved credential phishing, with most of the remaining 1% delivering remote access trojans (RATs) like ConnectWise RAT, Dark Crystal and XWorm.

.es domains are proving popular for phishing attacks

Although the rise of .es domains in cyberattacks is noteworthy, attack vectors remain unchanged. Malware was seen to be delivered by C2 nodes or spoofed emails, with most (95%) impersonating Microsoft (an attacker's favorite). Adobe, Google, Docusign and the Social Security Administration made up the top-five most commonly impersonated websites. Email lures often mimicked HR and document-related requests.

Interestingly, the malicious .es subdomains were randomly generated, not crafted manually, making them easier to identify as being fake. Examples include ag7sr[.]fjlabpkgcuo[.]es and gymi8[.]fwpzza[.]es.

Despite researchers suggesting that no similarities can be used to link attacks to a single group, 99% of the malicious .es domains were hosted on Cloudflare.

"If one threat actor or threat actor group were taking advantage of .es TLD domains then it is likely that the brands spoofed in .es TLD campaigns would indicate certain preferences by the threat actors," the researchers wrote.

Cofense explained that "significant restrictions" on the usage of .es TLDs were in place until 2005, adding that the recent rise in .es-related attacks could be a cause for concern, marking a new trend exploiting the authority that country-related TLDs unofficially carry.

You might also like

• We've listed the best identity theft protection services
• Keep yourself protected online by installing the best business VPN
• ChatGPT and other AI tools could be putting users at risk by getting company web addresses wrong

Years of insufferable summers and five AC units later, I've finally found the only window unit that keeps my home cool (and looks good while doing it).

Ahead of Galaxy Unpacked, Samsung tells CNET what's in store for its upcoming foldables -- and why it's making some key design changes.

Spending time on the internet in 2025 is proving to be more difficult than ever before. After all, image generation tools are readily accessible in AI products like ChatGPT, and video generation software like Google's Veo 3 is available to the general public.

We're still at a point in the AI-powered internet journey where I'd argue it's still easy enough to tell what's real and what's not. That said, today's AI image and video tools are more powerful than ever, and the scary thing is: this is the worst they'll ever be.

The other day, while browsing Instagram, I stumbled across a new post from France and Real Madrid soccer superstar, Kylian Mbappé. Widely regarded as one of the best players in the world, the image showcased an incredible bicycle kick he had performed in the Club World Cup quarter-final the night before.

The thing is, while the image appeared normal at first glance, a closer inspection revealed some seriously bad editing, or quite possibly AI image generation.

You see, zoom into the crowd watching Mbappé's acrobatics and you'll see a Picasso-esque abstract work of art, people merging into each other, and something more akin to a horror movie than a soccer field. All of the faces blur together like a nightmare.

Initially, I couldn't quite believe that a social media admin for an Instagram account with over 124 million followers would use AI to enhance an image. But then I thought about it some more, and of course, AI has infiltrated the social profiles of public figures. It's everywhere.

A post shared by Kylian Mbappé (@k.mbappe)

A photo posted by on

What's real anymore?

You might be wondering why Kylian Mbappé's social media team used AI to edit this photo in the first place. After all, according to official reports, there were over 76,000 fans in MetLife Stadium for the game.

Since it started a few weeks ago, the FIFA Club World Cup has been plagued with stories of low attendance rates, with one report claiming the tournament had over 1 million empty seats in the group stage alone.

Now, 76,000 is closer to MetLife's 82,500 capacity than some of the other examples of low attendance in the tournament, but considering the terrible photo edits to the crowd, did Mbappé's incredible goal fall into the line of sight of empty seats?

Thinking about the context of the tournament and the badly edited photo, this is the conclusion that I'd feel most comfortable putting forward. After all, sometimes you have to call a spade a spade. Why would anyone play with the crowd capacity other than to fill gaps?

Whether Mbappe's photo was AI-edited or badly edited by a human, the world of photo editing on social media is nothing new. Since the inception of platforms like Facebook and Instagram, we've all had to adapt to the 21st-century way of living, which often includes curating your best angles and happiest moments online.

AI-generated instagram influencer (Image credit: @millasofiafin / Instagram)

In the past, influencers would slim down their waists using apps on their iPhones or by manually editing the photo. Now, AI can do it all with the press of a button. And that's not all, AI-edited photos are only the beginning.

Whether that's fully AI-generated influencers scoring brand deals, AI-generated YouTube channels garnering millions of views and subscribers, or AI beauty pageants, we no longer live in a world where it's easy to tell what's real on social media and what's not. The thing is, have we ever been able to? Or is AI just making it easier than ever before to bend reality to fit a narrative?

As someone who has used multiple AI photo editing tools for testing purposes, I'm fairly confident that Mbappé's terrible editing is the work of artificial intelligence, and most probably Apple's mediocre Clean Up tool.

This clear example of heavily-edited photography shines a light, not only on bad editing, but also on the fact that if bad edits slip through the cracks, there's a high chance most of the posts we see aren't actually rooted in real life.

You might also like

• Can an AI travel agent remove the stress of vacation booking? I tried three to find out
• I tried a ChatGPT prompt that 'unlocks 4o’s full power', and I don’t know why I didn’t try it sooner
• I asked AI to plan the perfect day in London – here’s what it got right and hilariously wrong

Spending time on the internet in 2025 is proving to be more difficult than ever before. After all, image generation tools are readily accessible in AI products like ChatGPT, and video generation software like Google's Veo 3 is available to the general public.

We're still at a point in the AI-powered internet journey where I'd argue it's still easy enough to tell what's real and what's not. That said, today's AI image and video tools are more powerful than ever, and the scary thing is: this is the worst they'll ever be.

The other day, while browsing Instagram, I stumbled across a new post from France and Real Madrid soccer superstar, Kylian Mbappé. Widely regarded as one of the best players in the world, the image showcased an incredible bicycle kick he had performed in the Club World Cup quarter-final the night before.

The thing is, while the image appeared normal at first glance, a closer inspection revealed some seriously bad editing, or quite possibly AI image generation.

You see, zoom into the crowd watching Mbappé's acrobatics and you'll see a Picasso-esque abstract work of art, people merging into each other, and something more akin to a horror movie than a soccer field. All of the faces blur together like a nightmare.

Initially, I couldn't quite believe that a social media admin for an Instagram account with over 124 million followers would use AI to enhance an image. But then I thought about it some more, and of course, AI has infiltrated the social profiles of public figures. It's everywhere.

A post shared by Kylian Mbappé (@k.mbappe)

A photo posted by on

What's real anymore?

You might be wondering why Kylian Mbappé's social media team used AI to edit this photo in the first place. After all, according to official reports, there were over 76,000 fans in MetLife Stadium for the game.

Since it started a few weeks ago, the FIFA Club World Cup has been plagued with stories of low attendance rates, with one report claiming the tournament had over 1 million empty seats in the group stage alone.

Now, 76,000 is closer to MetLife's 82,500 capacity than some of the other examples of low attendance in the tournament, but considering the terrible photo edits to the crowd, did Mbappé's incredible goal fall into the line of sight of empty seats?

Thinking about the context of the tournament and the badly edited photo, this is the conclusion that I'd feel most comfortable putting forward. After all, sometimes you have to call a spade a spade. Why would anyone play with the crowd capacity other than to fill gaps?

Whether Mbappe's photo was AI-edited or badly edited by a human, the world of photo editing on social media is nothing new. Since the inception of platforms like Facebook and Instagram, we've all had to adapt to the 21st-century way of living, which often includes curating your best angles and happiest moments online.

AI-generated instagram influencer (Image credit: @millasofiafin / Instagram)

In the past, influencers would slim down their waists using apps on their iPhones or by manually editing the photo. Now, AI can do it all with the press of a button. And that's not all, AI-edited photos are only the beginning.

Whether that's fully AI-generated influencers scoring brand deals, AI-generated YouTube channels garnering millions of views and subscribers, or AI beauty pageants, we no longer live in a world where it's easy to tell what's real on social media and what's not. The thing is, have we ever been able to? Or is AI just making it easier than ever before to bend reality to fit a narrative?

As someone who has used multiple AI photo editing tools for testing purposes, I'm fairly confident that Mbappé's terrible editing is the work of artificial intelligence, and most probably Apple's mediocre Clean Up tool.

This clear example of heavily-edited photography shines a light, not only on bad editing, but also on the fact that if bad edits slip through the cracks, there's a high chance most of the posts we see aren't actually rooted in real life.

You might also like

• Can an AI travel agent remove the stress of vacation booking? I tried three to find out
• I tried a ChatGPT prompt that 'unlocks 4o’s full power', and I don’t know why I didn’t try it sooner
• I asked AI to plan the perfect day in London – here’s what it got right and hilariously wrong

From massage guns to your favorite sunscreens, you’ll want to stock up now.

The UK Public Accounts Committee’s recent report sends a clear and urgent message: cyber threats are evolving faster than defenses can keep up. The digital infrastructure underpinning our critical infrastructure is increasingly exposed—not only due to external threats but because of internal gaps in strategy, capability, and legacy system management.

Replacing outdated technology may be part of the solution, but it’s far from the full picture. We need a fundamental shift in mindset—toward continuous assurance, smarter system design, and a dynamic approach to skills development that anticipates the challenges of tomorrow, not just today.

Beyond ‘build and forget’: cybersecurity as an ongoing commitment

For too long, cybersecurity has followed a static, compliance-driven model—deploy once, tick the box, and move on. In today’s evolving threat landscape, this ‘build and forget’ mentality is no longer viable, if it ever was.

Security must be woven into every stage of design, development, and operations through a Secure by Design approach. With Cyber Physical Systems and enterprise IT environments in constant flux, reassessing security posture regularly ensures defenses remain adaptive and effective.

The UK government has rightly prioritized Secure by Design in its Defending the UK in a Digital World: Cyber Security Strategy 2022–25. Yet, despite this ambition, adoption across sectors remains uneven, with many organizations still relying on outdated risk frameworks and reactive measures—essentially attempting to counter modern threats with legacy solutions.

Cybersecurity must evolve beyond static processes. It requires continuous evaluation, proactive defense, and resilient security strategies to stay ahead of emerging risks.

Legacy systems: balancing risk and progress

Few areas illustrate the tension between innovation and practicality more clearly than legacy systems. Originally built for a different technological landscape, many were air-gapped, manually operated, and completely isolated from external networks—never designed to withstand the level of connectivity and cyber threats seen today.

In pursuit of efficiency and cost reduction, organizations have increasingly networked and remotely managed these systems, often without implementing adequate security safeguards. While this enhances operational flexibility, it also exposes critical infrastructure to new vulnerabilities, opening doors to sophisticated cyber threats.

The solution isn't as simple as replacing old systems outright. The decision to upgrade or extend the life of legacy platforms requires careful cyber risk evaluation, ensuring the right balance of mitigation strategies, isolation measures, and continuous monitoring to maintain security. Organizations must also weigh financial constraints, applying appropriate risk controls to optimize security investments without excessive costs.

Secure by design: a strategic imperative

Secure by Design isn’t just a cybersecurity buzzword—it’s an essential principle for building resilient digital infrastructure. It ensures that an appropriate level of security is built in from the ground up, integrated at every design, development, and operational phase to create adaptable, auditable, and testable systems.

Yet, despite its inclusion in policy frameworks and industry guidelines, implementation is often incomplete or superficial. Many organizations pay lip service to security but fail to embed it across teams and processes, treating it as an isolated function rather than an organizational priority – much like organizations approach health and safety.

Regulation will play a vital role in closing this gap. The upcoming Cyber Security and Resilience Bill is set to improve oversight, enforce stronger standards, and introduce mandatory incident reporting for high-risk sectors. Coupled with enhanced threat intelligence sharing, this legislation could shift cybersecurity strategies from reactive defense to proactive resilience.

Cyber talent: securing the future workforce

Another major challenge in cybersecurity is building a workforce capable of responding to evolving threats. As technology advances, skills become obsolete faster than ever, requiring ongoing investment in cyber capability development.

While direct government hiring plays a role—especially in sensitive security domains—industry partnerships offer a scalable alternative. Specialist organizations, like Thales, are well-positioned to train and upskill professionals through apprenticeships, immersive simulations, and diverse sector exposure, which traditional public-sector roles often lack.

This model provides dual benefits: it equips trainees with experience across various threat scenarios, while also fostering an agile workforce that can quickly adapt to emerging technologies like artificial intelligence and quantum computing. Cybersecurity isn't just about fixing vulnerabilities of the past—it's about anticipating and securing against the threats of the future.

Transparency and collaboration: strengthening cyber resilience

Cybersecurity can no longer be viewed as a standalone issue—threats don’t respect organizational boundaries, and weaknesses in one system can expose an entire network. Transparency, collaboration, and mandatory incident reporting are essential for national security, ensuring vulnerabilities are addressed before they escalate into widespread risks.

Reporting cyber incidents, much like the practice of reporting near misses and accidents in health and safety, strengthens overall resilience. Just as greater visibility has helped organizations reduce safety incidents, increased cyber reporting makes it harder for hackers to exploit gaps, reinforcing defenses before real damage occurs.

The Thales Data Threat Report underscores the growing risks to Critical National Infrastructure (CNI) and highlights the value of compliance—organisations that passed cybersecurity audits had significantly fewer breaches than those that failed. With the Cyber Security and Resilience Bill raising standards, stronger protections for essential infrastructure, including data centers that support AI innovation and national healthcare, will become the norm.

Each unreported cyber-attack is a missed opportunity to refine security strategies. Enhanced visibility into cyber threats improves collective intelligence, allowing organizations to make faster, smarter decisions in the face of emerging risks.

Conclusion: building security for the future

Modern cybersecurity requires more than patching vulnerabilities or checking compliance boxes—it demands a strategic rethink of how systems are designed, maintained, and safeguarded. Understanding how legacy and modern environments interact is key, alongside developing cyber talent that can anticipate and mitigate future risks.

Secure by Design must be the standard, not the exception. With the right balance of policy, proactive security measures, and skilled professionals, organizations can shift from reactive defense to sustainable resilience, ensuring they are prepared for the threats ahead, not just responding to the ones they see today.

We've featured the best endpoint protection software.

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

The UK Public Accounts Committee’s recent report sends a clear and urgent message: cyber threats are evolving faster than defenses can keep up. The digital infrastructure underpinning our critical infrastructure is increasingly exposed—not only due to external threats but because of internal gaps in strategy, capability, and legacy system management.

Replacing outdated technology may be part of the solution, but it’s far from the full picture. We need a fundamental shift in mindset—toward continuous assurance, smarter system design, and a dynamic approach to skills development that anticipates the challenges of tomorrow, not just today.

Beyond ‘build and forget’: cybersecurity as an ongoing commitment

For too long, cybersecurity has followed a static, compliance-driven model—deploy once, tick the box, and move on. In today’s evolving threat landscape, this ‘build and forget’ mentality is no longer viable, if it ever was.

Security must be woven into every stage of design, development, and operations through a Secure by Design approach. With Cyber Physical Systems and enterprise IT environments in constant flux, reassessing security posture regularly ensures defenses remain adaptive and effective.

The UK government has rightly prioritized Secure by Design in its Defending the UK in a Digital World: Cyber Security Strategy 2022–25. Yet, despite this ambition, adoption across sectors remains uneven, with many organizations still relying on outdated risk frameworks and reactive measures—essentially attempting to counter modern threats with legacy solutions.

Cybersecurity must evolve beyond static processes. It requires continuous evaluation, proactive defense, and resilient security strategies to stay ahead of emerging risks.

Legacy systems: balancing risk and progress

Few areas illustrate the tension between innovation and practicality more clearly than legacy systems. Originally built for a different technological landscape, many were air-gapped, manually operated, and completely isolated from external networks—never designed to withstand the level of connectivity and cyber threats seen today.

In pursuit of efficiency and cost reduction, organizations have increasingly networked and remotely managed these systems, often without implementing adequate security safeguards. While this enhances operational flexibility, it also exposes critical infrastructure to new vulnerabilities, opening doors to sophisticated cyber threats.

The solution isn't as simple as replacing old systems outright. The decision to upgrade or extend the life of legacy platforms requires careful cyber risk evaluation, ensuring the right balance of mitigation strategies, isolation measures, and continuous monitoring to maintain security. Organizations must also weigh financial constraints, applying appropriate risk controls to optimize security investments without excessive costs.

Secure by design: a strategic imperative

Secure by Design isn’t just a cybersecurity buzzword—it’s an essential principle for building resilient digital infrastructure. It ensures that an appropriate level of security is built in from the ground up, integrated at every design, development, and operational phase to create adaptable, auditable, and testable systems.

Yet, despite its inclusion in policy frameworks and industry guidelines, implementation is often incomplete or superficial. Many organizations pay lip service to security but fail to embed it across teams and processes, treating it as an isolated function rather than an organizational priority – much like organizations approach health and safety.

Regulation will play a vital role in closing this gap. The upcoming Cyber Security and Resilience Bill is set to improve oversight, enforce stronger standards, and introduce mandatory incident reporting for high-risk sectors. Coupled with enhanced threat intelligence sharing, this legislation could shift cybersecurity strategies from reactive defense to proactive resilience.

Cyber talent: securing the future workforce

Another major challenge in cybersecurity is building a workforce capable of responding to evolving threats. As technology advances, skills become obsolete faster than ever, requiring ongoing investment in cyber capability development.

While direct government hiring plays a role—especially in sensitive security domains—industry partnerships offer a scalable alternative. Specialist organizations, like Thales, are well-positioned to train and upskill professionals through apprenticeships, immersive simulations, and diverse sector exposure, which traditional public-sector roles often lack.

This model provides dual benefits: it equips trainees with experience across various threat scenarios, while also fostering an agile workforce that can quickly adapt to emerging technologies like artificial intelligence and quantum computing. Cybersecurity isn't just about fixing vulnerabilities of the past—it's about anticipating and securing against the threats of the future.

Transparency and collaboration: strengthening cyber resilience

Cybersecurity can no longer be viewed as a standalone issue—threats don’t respect organizational boundaries, and weaknesses in one system can expose an entire network. Transparency, collaboration, and mandatory incident reporting are essential for national security, ensuring vulnerabilities are addressed before they escalate into widespread risks.

Reporting cyber incidents, much like the practice of reporting near misses and accidents in health and safety, strengthens overall resilience. Just as greater visibility has helped organizations reduce safety incidents, increased cyber reporting makes it harder for hackers to exploit gaps, reinforcing defenses before real damage occurs.

The Thales Data Threat Report underscores the growing risks to Critical National Infrastructure (CNI) and highlights the value of compliance—organisations that passed cybersecurity audits had significantly fewer breaches than those that failed. With the Cyber Security and Resilience Bill raising standards, stronger protections for essential infrastructure, including data centers that support AI innovation and national healthcare, will become the norm.

Each unreported cyber-attack is a missed opportunity to refine security strategies. Enhanced visibility into cyber threats improves collective intelligence, allowing organizations to make faster, smarter decisions in the face of emerging risks.

Conclusion: building security for the future

Modern cybersecurity requires more than patching vulnerabilities or checking compliance boxes—it demands a strategic rethink of how systems are designed, maintained, and safeguarded. Understanding how legacy and modern environments interact is key, alongside developing cyber talent that can anticipate and mitigate future risks.

Secure by Design must be the standard, not the exception. With the right balance of policy, proactive security measures, and skilled professionals, organizations can shift from reactive defense to sustainable resilience, ensuring they are prepared for the threats ahead, not just responding to the ones they see today.

We've featured the best endpoint protection 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

In the world of technology, open source refers to software which is made publicly available, and is free for anyone to use or modify. It is a very popular approach to software development, enabling an open and community engagement to building software projects.

One of the main advantages of open source is transparency, a property that is widely embraced by tech enthusiasts. Open source code means that anyone can inspect the code before usage, meaning that anyone could spot malicious, or broken, code. This is a perfect example of the mantra “Don't trust, verify”, which is popular in various technology spaces.

One of the main drawbacks of open source is considered to be lack of interested involvement from external developers. Many projects typically start with one, or a few, developers – and this base can struggle to grow, leading to a low “bus-factor” (dependency on too few developers) for the majority of open source projects.

This can lead to developer burnout or exhaustion, and can eventually cause the ultimate abandonment of projects. Alongside this, funding can also play a crucial role; many of the most successful projects are backed by some degree of funding from organizations and/or non-profits, effectively supporting developers throughout the time required to maintain the project.

Thankfully, while these issues are very much real there are several examples of successful and pioneering projects in the open source space which have overcome these obstacles. The Linux Foundation is the most prominent example, and there are also some interesting ones in the cryptography space, a field that can particularly benefit from adapting the open source approach. Let’s take a look at some of them individually below.

The Linux Foundation

The most notable example of a successful open source project is the open source development of the Linux Kernel. Originally published in 1991 under a custom license which did not allow commercial usage, this was quickly shifted to an open source-friendly license (GPL), to enable developers to collaborate on the code and ensure wide and free availability of the software.

The open source version which can now be found on Github, is one of the most contributed projects in the history of open source. With over 15,000 contributors, this project has clearly been a huge success. Whilst only ~4% of Desktop computers worldwide are now running Linux, it is still the platform of choice for many computer scientists and cryptographers.

Let’s encrypt

In the cryptography space, a popular example is Let’s Encrypt. Let’s Encrypt is an influential open source certificate authority (CA) which helps enable secure communication via HTTPS over the web. This project was started by the Internet Security Research Group (ISG), which is a small non-profit launched in 2015 and backed by various industry players.

Since their launch in 2015, HTTPS usage on the internet has risen from ~40% to ~90%+ today. This, of course, is not down solely to Let’s encrypt, but their mission to “democratize HTTPS by making SSL/TLS certificates free, automated, and easy to obtain” has clearly had a major impact on this. Today, Let’s Encrypt powers secure communications for millions of sites, APIs, and services and is viewed as a major success in the cryptography space.

Open science

Open science is a similar idea which draws influence from open source. The goal of open science is to make academic publications (as well as any associated source code, data, etc) public and free for anyone to view or build on. Particularly in the cryptography space, this is something that academics tend to view as a must-have for their work: all experiments should be reproducible with available code. The website https://paperswithcode.com/ collects works which open-source their code in the Machine Learning space.

Such an approach can be critical for successful outcomes in science and research. Attempting to build on something “closed” is almost impossible, which may lead researchers to either completely rebuild something from scratch, or giving up entirely.

An additional element that can help promote this collaborative approach is the promotion of platforms like events and conferences that aim to gather experts and researchers and foster exchange. In the world of cryptography - and of Fully Homomorphic Encryption in particular - FHE.org is one such community, hosting regular online meetups and an annual conference to bring together the best and brightest minds.

Looking forward

The next frontier in this space, in my view, is AI. AI is an area showing immense growth over the past few years; there is a tangible improvement in models. OpenAI recently closed the largest funding round in history (40b at a 300b valuation). There is currently an AI-arms race between big-tech firms, and specialized start-ups who are vying to release the best model (Polymarket is currently backing Google, at the time of writing).

Clearly, this is a sector where open source principles could be applied well, with collaboration encouraged to ensure correct management of ethics, as opposed to control by a single organization or board.

We've featured the best encryption 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

In the world of technology, open source refers to software which is made publicly available, and is free for anyone to use or modify. It is a very popular approach to software development, enabling an open and community engagement to building software projects.

One of the main advantages of open source is transparency, a property that is widely embraced by tech enthusiasts. Open source code means that anyone can inspect the code before usage, meaning that anyone could spot malicious, or broken, code. This is a perfect example of the mantra “Don't trust, verify”, which is popular in various technology spaces.

One of the main drawbacks of open source is considered to be lack of interested involvement from external developers. Many projects typically start with one, or a few, developers – and this base can struggle to grow, leading to a low “bus-factor” (dependency on too few developers) for the majority of open source projects.

This can lead to developer burnout or exhaustion, and can eventually cause the ultimate abandonment of projects. Alongside this, funding can also play a crucial role; many of the most successful projects are backed by some degree of funding from organizations and/or non-profits, effectively supporting developers throughout the time required to maintain the project.

Thankfully, while these issues are very much real there are several examples of successful and pioneering projects in the open source space which have overcome these obstacles. The Linux Foundation is the most prominent example, and there are also some interesting ones in the cryptography space, a field that can particularly benefit from adapting the open source approach. Let’s take a look at some of them individually below.

The Linux Foundation

The most notable example of a successful open source project is the open source development of the Linux Kernel. Originally published in 1991 under a custom license which did not allow commercial usage, this was quickly shifted to an open source-friendly license (GPL), to enable developers to collaborate on the code and ensure wide and free availability of the software.

The open source version which can now be found on Github, is one of the most contributed projects in the history of open source. With over 15,000 contributors, this project has clearly been a huge success. Whilst only ~4% of Desktop computers worldwide are now running Linux, it is still the platform of choice for many computer scientists and cryptographers.

Let’s encrypt

In the cryptography space, a popular example is Let’s Encrypt. Let’s Encrypt is an influential open source certificate authority (CA) which helps enable secure communication via HTTPS over the web. This project was started by the Internet Security Research Group (ISG), which is a small non-profit launched in 2015 and backed by various industry players.

Since their launch in 2015, HTTPS usage on the internet has risen from ~40% to ~90%+ today. This, of course, is not down solely to Let’s encrypt, but their mission to “democratize HTTPS by making SSL/TLS certificates free, automated, and easy to obtain” has clearly had a major impact on this. Today, Let’s Encrypt powers secure communications for millions of sites, APIs, and services and is viewed as a major success in the cryptography space.

Open science

Open science is a similar idea which draws influence from open source. The goal of open science is to make academic publications (as well as any associated source code, data, etc) public and free for anyone to view or build on. Particularly in the cryptography space, this is something that academics tend to view as a must-have for their work: all experiments should be reproducible with available code. The website https://paperswithcode.com/ collects works which open-source their code in the Machine Learning space.

Such an approach can be critical for successful outcomes in science and research. Attempting to build on something “closed” is almost impossible, which may lead researchers to either completely rebuild something from scratch, or giving up entirely.

An additional element that can help promote this collaborative approach is the promotion of platforms like events and conferences that aim to gather experts and researchers and foster exchange. In the world of cryptography - and of Fully Homomorphic Encryption in particular - FHE.org is one such community, hosting regular online meetups and an annual conference to bring together the best and brightest minds.

Looking forward

The next frontier in this space, in my view, is AI. AI is an area showing immense growth over the past few years; there is a tangible improvement in models. OpenAI recently closed the largest funding round in history (40b at a 300b valuation). There is currently an AI-arms race between big-tech firms, and specialized start-ups who are vying to release the best model (Polymarket is currently backing Google, at the time of writing).

Clearly, this is a sector where open source principles could be applied well, with collaboration encouraged to ensure correct management of ethics, as opposed to control by a single organization or board.

We've featured the best encryption 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

It’s good news for those of us who can’t wait to head back to Virgin River, because seasons 1-2 of Sullivan’s Crossing – also created by author Robyn Carr – are now available to stream on Netflix. You might have missed it in everything new coming to Netflix in July 2025, but there’s a good reason why you shouldn’t overlook it (and not just that season 3 is still to come).

Virgin River’s success alone is nothing short of astonishing. It’s now the longest-running original series on the platform, still pulling in almost 16 million viewers when season 6 dropped in December 2024. Virgin River season 7 has officially wrapped filming, but no release date is confirmed as of writing.

Bad news, right? Wrong! Sullivan’s Crossing is the ideal new TV show to tide us over in the meantime. Set in Canada rather than California, it’s the epitome of cozy comfort drama with plenty of famous faces thrown in for good measure.

Why I recommend watching Sullivan’s Crossing on Netflix

Neurosurgeon Maggie (Morgan Kohan) is looking to move away from her big city life in Boston after a terrible tragedy affects her personal life. She moves back to her childhood hometown of Sullivan’s Crossing, Nova Scotia, where her dad Sully (Scott Patterson) runs a campground. Maggie meets new resident Cal (Chad Michael Murray), and sparks soon fly. Where have we heard that storyline before?

Sure, Sullivan’s Crossing is exactly the same as Virgin River if you strip it back to its bare bones, but in reality, the hit TV show is so much more than its big sister. For starters, the cast unites When Hope Calls’ Kohan, noughties teen heartthrob Murray and Gilmore Girls legend Patterson in family conflict that always pulls on the heartstrings, allowing each actor to do what they’re best at. Obviously, it doesn’t hurt that any drama takes place against the jaw-droppingly picturesque vistas of rural Canada.

Our main anchor points differ from Virgin River, too. Maggie is falling for Cal, but still has tenuous links to her ex, Andrew (Allan Hawco). In essence, this acts as the reverse of Jack Sheridan (Martin Henderson), who falls for newcomer Mel (Alexandra Breckenridge) but hasn’t fully broken things off with Charmaine (Lauren Hammersley). Sullivan’s Crossing takes its romantic triangle one step further, embroiling the trio in a messy pregnancy plot destined to change all of their lives for good.

What’s the best thing about Sullivan’s Crossing, I hear you ask? The beautiful mundanity of their day-to-day lives. Where rival TV shows like Riverdale and the reboot of Dynasty take the drama in their dynamics to the extreme, Sullivan’s Crossing always returns to the simplicity of what real life really looks like, and that pays off in its popularity. You can blissfully lose your mind to the till breaking at the campground reception, inane chatter happening in the local diner and the many whimsical local events that bring the community together.

If you just can’t wait for Virgin River to return, or want to immerse yourself in more of the same, there’s no question that Sullivan’s Crossing will scratch that itch.

You might also like

• I watch TV for a living and these are my top 3 Netflix shows that you should watch now
• Netflix’s #1 movie is a K-pop demon hunting anime with 97% on Rotten Tomatoes that I can’t get out of my head
• Netflix will start streaming NASA content soon – and the results look out of this world

It’s good news for those of us who can’t wait to head back to Virgin River, because seasons 1-2 of Sullivan’s Crossing – also created by author Robyn Carr – are now available to stream on Netflix. You might have missed it in everything new coming to Netflix in July 2025, but there’s a good reason why you shouldn’t overlook it (and not just that season 3 is still to come).

Virgin River’s success alone is nothing short of astonishing. It’s now the longest-running original series on the platform, still pulling in almost 16 million viewers when season 6 dropped in December 2024. Virgin River season 7 has officially wrapped filming, but no release date is confirmed as of writing.

Bad news, right? Wrong! Sullivan’s Crossing is the ideal new TV show to tide us over in the meantime. Set in Canada rather than California, it’s the epitome of cozy comfort drama with plenty of famous faces thrown in for good measure.

Why I recommend watching Sullivan’s Crossing on Netflix

Neurosurgeon Maggie (Morgan Kohan) is looking to move away from her big city life in Boston after a terrible tragedy affects her personal life. She moves back to her childhood hometown of Sullivan’s Crossing, Nova Scotia, where her dad Sully (Scott Patterson) runs a campground. Maggie meets new resident Cal (Chad Michael Murray), and sparks soon fly. Where have we heard that storyline before?

Sure, Sullivan’s Crossing is exactly the same as Virgin River if you strip it back to its bare bones, but in reality, the hit TV show is so much more than its big sister. For starters, the cast unites When Hope Calls’ Kohan, noughties teen heartthrob Murray and Gilmore Girls legend Patterson in family conflict that always pulls on the heartstrings, allowing each actor to do what they’re best at. Obviously, it doesn’t hurt that any drama takes place against the jaw-droppingly picturesque vistas of rural Canada.

Our main anchor points differ from Virgin River, too. Maggie is falling for Cal, but still has tenuous links to her ex, Andrew (Allan Hawco). In essence, this acts as the reverse of Jack Sheridan (Martin Henderson), who falls for newcomer Mel (Alexandra Breckenridge) but hasn’t fully broken things off with Charmaine (Lauren Hammersley). Sullivan’s Crossing takes its romantic triangle one step further, embroiling the trio in a messy pregnancy plot destined to change all of their lives for good.

What’s the best thing about Sullivan’s Crossing, I hear you ask? The beautiful mundanity of their day-to-day lives. Where rival TV shows like Riverdale and the reboot of Dynasty take the drama in their dynamics to the extreme, Sullivan’s Crossing always returns to the simplicity of what real life really looks like, and that pays off in its popularity. You can blissfully lose your mind to the till breaking at the campground reception, inane chatter happening in the local diner and the many whimsical local events that bring the community together.

If you just can’t wait for Virgin River to return, or want to immerse yourself in more of the same, there’s no question that Sullivan’s Crossing will scratch that itch.

You might also like

• I watch TV for a living and these are my top 3 Netflix shows that you should watch now
• Netflix’s #1 movie is a K-pop demon hunting anime with 97% on Rotten Tomatoes that I can’t get out of my head
• Netflix will start streaming NASA content soon – and the results look out of this world

The shift comes after nearly 20 years of mandatory shoe removal, a rule put in place after the 2001 "shoe bomber" attempt reshaped airport security protocols.

• The KEF XIO soundbar costs £1,999 / $2,499 / AU$3,600
• Dolby Atmos, DTS:X and Sony 360 Reality Audio
• 5.1.2 channels with 12 class D amplifiers

Audio legend KEF has been busy: not only did its team collaborate with Nothing on the Headphone (1), but it's also unveiled its first-ever Dolby Atmos soundbar. The team is clearly pretty proud of it, and promises a "transcendent" audio experience.

At $2,499 / £1,999 this is clearly one for high-end home theaters. It's a 5.1.2-channel all-in-one soundbar with 12 class D amplifiers putting out a whopping 820W of total power.

Despite that, the XIO soundbar is still shallow enough to look good on a wall. And where many soundbars are made specifically for movies, KEF says that the XIO has been made with music in mind too.

(Image credit: KEF)KEF XIO soundbar: key features

The soundbar has six Uni-Q MX drivers, which are smaller versions of the Uni-Q drivers you'll find in larger KEF speakers. These are special because they're two speaker in one – a smaller driver sits at the center of the cone of the large speaker, creating a wider soundstage between them that forms one clear audio signal.

Three of these drivers are placed on top for upward-firing audio, and the other three face forwards. They're teamed up with four P185 bass drivers, which are rectangular drivers that KEF says have the same radiating area as a standard circular 10cm driver. The rectangular shape means the soundbar can accommodate more drivers without adding bulk – similar racetrack drivers are used in lots of the best soundbars.

Despite that serious low-end power, you shouldn't get much in the way of unwanted vibration: KEF says the way it has arranged the drivers, with two pairs facing in opposite directions, meaning that each driver cancels out the vibration of its opposite.

The XIO also features KEF's patent-pending Velocity Control Technology, which it calls VECO. It's a sensor that monitors the P185 driver cones, adjusting the movement to compensate for any errors it detects in order to minimize distortion and compression.

Another key acronym is MIE, short for the Music Integrity Engine. This is a suite of DSP algorithms made specifically for the XIO to handle multi-channel processing, virtualization and object placement. The result, KEF says, is "transcendent soundscapes".

The XIO works with all the major streamers including Tidal, Amazon Music, Qobuz and Deezer, and it has HDMI eARC, optical and subwoofer outputs. You can also add any KEF subwoofer via the optional KW2 wireless receiver.

The KEF XIO is available via early access to myKEF members from today, 8 July, and it goes on sale on the 18th of July. It'll also be available from today in certain retailers, including Peter Tyson AV and Sevenoaks Sound & Vision in the UK, with wider retail availability starting on 8th August 2025.

You might also like

• The best soundbars: TV audio upgrade for all budgets
• The best Dolby Atmos soundbars for all budgets
• Why owning a soundbar has been a game-changer for me

• The KEF XIO soundbar costs £1,999 / $2,499 / AU$3,600
• Dolby Atmos, DTS:X and Sony 360 Reality Audio
• 5.1.2 channels with 12 class D amplifiers

Audio legend KEF has been busy: not only did its team collaborate with Nothing on the Headphone (1), but it's also unveiled its first-ever Dolby Atmos soundbar. The team is clearly pretty proud of it, and promises a "transcendent" audio experience.

At $2,499 / £1,999 this is clearly one for high-end home theaters. It's a 5.1.2-channel all-in-one soundbar with 12 class D amplifiers putting out a whopping 820W of total power.

Despite that, the XIO soundbar is still shallow enough to look good on a wall. And where many soundbars are made specifically for movies, KEF says that the XIO has been made with music in mind too.

(Image credit: KEF)KEF XIO soundbar: key features

The soundbar has six Uni-Q MX drivers, which are smaller versions of the Uni-Q drivers you'll find in larger KEF speakers. These are special because they're two speaker in one – a smaller driver sits at the center of the cone of the large speaker, creating a wider soundstage between them that forms one clear audio signal.

Three of these drivers are placed on top for upward-firing audio, and the other three face forwards. They're teamed up with four P185 bass drivers, which are rectangular drivers that KEF says have the same radiating area as a standard circular 10cm driver. The rectangular shape means the soundbar can accommodate more drivers without adding bulk – similar racetrack drivers are used in lots of the best soundbars.

Despite that serious low-end power, you shouldn't get much in the way of unwanted vibration: KEF says the way it has arranged the drivers, with two pairs facing in opposite directions, meaning that each driver cancels out the vibration of its opposite.

The XIO also features KEF's patent-pending Velocity Control Technology, which it calls VECO. It's a sensor that monitors the P185 driver cones, adjusting the movement to compensate for any errors it detects in order to minimize distortion and compression.

Another key acronym is MIE, short for the Music Integrity Engine. This is a suite of DSP algorithms made specifically for the XIO to handle multi-channel processing, virtualization and object placement. The result, KEF says, is "transcendent soundscapes".

The XIO works with all the major streamers including Tidal, Amazon Music, Qobuz and Deezer, and it has HDMI eARC, optical and subwoofer outputs. You can also add any KEF subwoofer via the optional KW2 wireless receiver.

The KEF XIO is available via early access to myKEF members from today, 8 July, and it goes on sale on the 18th of July. It'll also be available from today in certain retailers, including Peter Tyson AV and Sevenoaks Sound & Vision in the UK, with wider retail availability starting on 8th August 2025.

You might also like

• The best soundbars: TV audio upgrade for all budgets
• The best Dolby Atmos soundbars for all budgets
• Why owning a soundbar has been a game-changer for me

• Atomic Stealer, or AMOS, is no longer just a pure infostealer, experts warn
• The tool now comes with a backdoor and a persistence mechanism
• A new variant was seen circulating in the wild

Atomic Stealer (AMOS), one of the most dangerous infostealer malware threats on the macOS ecosystem, just got a significant upgrade that makes it even more dangerous, experts have warned.

A new version of the malware was spotted sporting a backdoor that not only allows persistent access and survives reboots, but also grants the attackers the ability to deploy any other malware on the compromised device, as well.

The news comes courtesy of MacPaw’s cybersecurity arm, Moonlock, who were tipped off by an independent researcher with the alias g0njxa., who noted the backdoored version of Atomic macOS Stealer now has the potential to gain full access to thousands of Mac devices worldwide.

Get 55% off Incogni's Data Removal service with code TECHRADAR

Wipe your personal data off the internet with the Incogni data removal service. Stop identity thieves
and protect your privacy from unwanted spam and scam calls.View Deal

A popular infostealer

AMOS has been around for years, establishing itself as the go-to stealer malware used in many major hacking campaigns. Until now, it was capable of extracting a wide range of data, including browser-stored passwords and keychains, autofill data, cryptocurrency wallet information, system data, and different files. It was also able to bypass macOS protections, tricking Gatekeeper and other macOS security features.

It was sold as MaaS (malware-as-a-service) on underground forums, and often distributed via fake apps and malicious websites.

We last heard of AMOS in early June 2025, when Russian threat actors used the popular ClickFix method to deploy it against their targets. At the time, security researchers from CloudSek reported multiple websites spoofing Spectrum, a US-based telecommunications provider, to deliver the malware.

In early January, software developer Ryan Chenkie spotted a malicious campaign on Google, promoting a fake version of Homebrew, an open source package manager for macOS and Linux that was, in fact, AMOS.

"AMOS malware campaigns have already reached over 120 countries, with the United States, France, Italy, the United Kingdom, and Canada among the most affected," the researchers warned.

Via BleepingComputer

You might also like

• Dangerous new MacOS malware is targeting Apple users everywhere - here's what you need to know
• Take a look at our guide to the best authenticator app
• We've rounded up the best password managers

• Atomic Stealer, or AMOS, is no longer just a pure infostealer, experts warn
• The tool now comes with a backdoor and a persistence mechanism
• A new variant was seen circulating in the wild

Atomic Stealer (AMOS), one of the most dangerous infostealer malware threats on the macOS ecosystem, just got a significant upgrade that makes it even more dangerous, experts have warned.

A new version of the malware was spotted sporting a backdoor that not only allows persistent access and survives reboots, but also grants the attackers the ability to deploy any other malware on the compromised device, as well.

The news comes courtesy of MacPaw’s cybersecurity arm, Moonlock, who were tipped off by an independent researcher with the alias g0njxa., who noted the backdoored version of Atomic macOS Stealer now has the potential to gain full access to thousands of Mac devices worldwide.

Get 55% off Incogni's Data Removal service with code TECHRADAR

Wipe your personal data off the internet with the Incogni data removal service. Stop identity thieves
and protect your privacy from unwanted spam and scam calls.View Deal

A popular infostealer

AMOS has been around for years, establishing itself as the go-to stealer malware used in many major hacking campaigns. Until now, it was capable of extracting a wide range of data, including browser-stored passwords and keychains, autofill data, cryptocurrency wallet information, system data, and different files. It was also able to bypass macOS protections, tricking Gatekeeper and other macOS security features.

It was sold as MaaS (malware-as-a-service) on underground forums, and often distributed via fake apps and malicious websites.

We last heard of AMOS in early June 2025, when Russian threat actors used the popular ClickFix method to deploy it against their targets. At the time, security researchers from CloudSek reported multiple websites spoofing Spectrum, a US-based telecommunications provider, to deliver the malware.

In early January, software developer Ryan Chenkie spotted a malicious campaign on Google, promoting a fake version of Homebrew, an open source package manager for macOS and Linux that was, in fact, AMOS.

"AMOS malware campaigns have already reached over 120 countries, with the United States, France, Italy, the United Kingdom, and Canada among the most affected," the researchers warned.

Via BleepingComputer

You might also like

• Dangerous new MacOS malware is targeting Apple users everywhere - here's what you need to know
• Take a look at our guide to the best authenticator app
• We've rounded up the best password managers

Exclusive: The Rubik's Pulse Cube from Spin Master transforms from a bland mirror cube to a light show puzzle.

• Foundation season 3 will see Gaal Dornick sacrifice everything to stop The Mule
• Hari Seldon's protégé really steps up this season, actor Lou Llobell says
• It remains to be seen if she'll succeed in thwarting this season's main villain

Lou Llobell has indicated Gaal Dornick will stop at nothing to ensure The Mule doesn't become the new galactic tyrant in Foundation season 3.

Speaking to TechRadar, Llobell, who has portrayed Dornick since the Apple TV+ sci-fi series began, says her character is willing to "sacrifice everything" to defeat this season's primary antagonist.

First teased in last season's finale, The Mule is an incredibly dangerous Mentalic – people who possess superhuman psychic abilities – who wants to bring the entire galaxy under his iron fist. You can read my Foundation season 2 ending explained article and/or exclusive chat with Pilou Asbaek, who says The Mule needs to be seen as "a big threat" to The Foundation and The Imperium, for more details.

However, Dornick, who's also a Mentalic, has had visions of The Mule since Foundation season 2. Alongside mentor and psychohistorian Hari Seldon (Jared Harris), Dornick has spent the past 152 years in and out of cryosleep to prepare a secret Ignis-based community of Mentalics – also known as Second Foundation – for The Mule's arrival.

Gaal has spent the past 152 years preparing for The Mule's arrival (Image credit: Apple TV+)

With the unhinged and ambitious villain finally showing his face in one of the best Apple TV+ shows' third season, it's time for Dornick to fully step up and prevent him from achieving his goal of becoming the galaxy's latest authoritarian.

To combat the demonstrably evil character played by Asbaek – the Game of Thrones alumnus replaced Mikael Persbrandt as The Mule in a season 3 cast shake-up, FYI – though, Dornick will have to put everything else aside, including her own happiness, to thwart his rapid and successful conquering of Foundation's fictional universe.

"Gail has grown so much," Llobell told me, "And you see her really take a massive leap into this new role as a leader and someone who's going to sacrifice everything to make sure the plan stays on track and that The Mule is stopped.

"It's beautiful to see her with all these different people [on Ignis] and the relationships that she has [with them]," Llobell continued. "But, her one goal is still to deal with The Mule. She's ready to take this on, even though things might not go according to plan as the season unfolds."

Find out what I thought about the first half of the show's latest chapter by reading my Foundation season 3 review. Be sure to read the below section, too, for more coverage, exclusive or otherwise, of the Apple TV Original before season 3 arrives on July 11, too.

You might also like

• 'It cracks me up': Foundation actor Lee Pace teases Brother Day's The Big Lebowski-like transformation in season 3 of the Apple TV+ show
• 'We're the last to find out': Foundation stars don't know if it'll return for season 4 – but they haven't given up hope for the Apple TV+ show's renewal
• Foundation season 3 trailer reveals the Apple TV+ show's most dangerous villain yet – and an unexpected alliance

• Foundation season 3 will see Gaal Dornick sacrifice everything to stop The Mule
• Hari Seldon's protégé really steps up this season, actor Lou Llobell says
• It remains to be seen if she'll succeed in thwarting this season's main villain

Lou Llobell has indicated Gaal Dornick will stop at nothing to ensure The Mule doesn't become the new galactic tyrant in Foundation season 3.

Speaking to TechRadar, Llobell, who has portrayed Dornick since the Apple TV+ sci-fi series began, says her character is willing to "sacrifice everything" to defeat this season's primary antagonist.

First teased in last season's finale, The Mule is an incredibly dangerous Mentalic – people who possess superhuman psychic abilities – who wants to bring the entire galaxy under his iron fist. You can read my Foundation season 2 ending explained article and/or exclusive chat with Pilou Asbaek, who says The Mule needs to be seen as "a big threat" to The Foundation and The Imperium, for more details.

However, Dornick, who's also a Mentalic, has had visions of The Mule since Foundation season 2. Alongside mentor and psychohistorian Hari Seldon (Jared Harris), Dornick has spent the past 152 years in and out of cryosleep to prepare a secret Ignis-based community of Mentalics – also known as Second Foundation – for The Mule's arrival.

Gaal has spent the past 152 years preparing for The Mule's arrival (Image credit: Apple TV+)

With the unhinged and ambitious villain finally showing his face in one of the best Apple TV+ shows' third season, it's time for Dornick to fully step up and prevent him from achieving his goal of becoming the galaxy's latest authoritarian.

To combat the demonstrably evil character played by Asbaek – the Game of Thrones alumnus replaced Mikael Persbrandt as The Mule in a season 3 cast shake-up, FYI – though, Dornick will have to put everything else aside, including her own happiness, to thwart his rapid and successful conquering of Foundation's fictional universe.

"Gail has grown so much," Llobell told me, "And you see her really take a massive leap into this new role as a leader and someone who's going to sacrifice everything to make sure the plan stays on track and that The Mule is stopped.

"It's beautiful to see her with all these different people [on Ignis] and the relationships that she has [with them]," Llobell continued. "But, her one goal is still to deal with The Mule. She's ready to take this on, even though things might not go according to plan as the season unfolds."

Find out what I thought about the first half of the show's latest chapter by reading my Foundation season 3 review. Be sure to read the below section, too, for more coverage, exclusive or otherwise, of the Apple TV Original before season 3 arrives on July 11, too.

You might also like

• 'It cracks me up': Foundation actor Lee Pace teases Brother Day's The Big Lebowski-like transformation in season 3 of the Apple TV+ show
• 'We're the last to find out': Foundation stars don't know if it'll return for season 4 – but they haven't given up hope for the Apple TV+ show's renewal
• Foundation season 3 trailer reveals the Apple TV+ show's most dangerous villain yet – and an unexpected alliance