TechRadar News

• EA executive vice president Vince Zampella has confirmed that Battlefield 6 won't work on the Steam Deck
• He says there are other "non-Steam Deck handheld platforms that it will work on"
• There are no current plans to bring the game to Nintendo Switch 2

EA has confirmed that Battlefield 6 won't work on the Steam Deck, but that there are other PC handhelds on which the game will run on.

That's according to EA executive vice president Vince Zampella, who verified the lack of Steam Deck support in an interview with PCGamesN at the Battlefield 6 multiplayer reveal event, when asked about the game potentially running on PC gaming handhelds like Valve's console or the ASUS ROG Ally.

"It does not work on Steam Deck," Zampella said. "I think there are some non-Steam Deck handheld platforms that it will work on, probably, right? But I don't think, you know, it's probably not the best way to play the game."

Although the executive didn't offer a reason why the upcoming first-person shooter won't run on the Steam Deck, it's likely due to EA's anti-cheat software, which is incompatible with Linux.

In terms of the Nintendo Switch 2, Zampella also confirmed that EA doesn't have any current plans to bring Battlefield 6 to the latest console.

"We're not doing Switch yet," he said. "Well, I'm not saying we are. I'm just saying we're not considering it right now."

During the multiplayer reveal event, it was announced that Battlefield 6 will launch on October 10, 2025, for PS5, Xbox Series X, Xbox Series S, and PC.

There will also be an early access period beginning August 7-8, followed by open betas on August 9-10, then again on August 14-17.

You might also like...

• The Battlefield 6 battle royale map has seemingly leaked, and Warzone's days could be numbered
• The Nintendo Switch 2 is the company’s least ambitious console to date, but its improvements are astronomical
• Microsoft says Xbox Game Pass reached 'nearly $5 billion' in annual revenue over the last year, setting a new record for the service

• YouTube is introducing AI age checks in the US to filter inappropriate content
• Machine learning will base decisions on searched videos and categories, among other criteria
• This announcement follows the recent blocking of YouTube in Australia for minors and restrictions on social media in the UK

YouTube has announced the introduction of AI-based checks to determine the age of viewers.

This news comes days after a complete ban on YouTube access for under-16s in Australia, and the introduction of age verification checks in the UK under the Online Safety Act.

Trialing the technology with a sample audience in the US, YouTube plans to expand it across the country, tailoring experiences based on the viewer’s age.

This doesn’t just mean blocking unsuitable content. Personalized advertising will also be disabled for children, while older viewers will be required to provide government-issued ID to confirm their age if the AI gets it wrong.

YouTube’s “built-in protections”

(Image credit: Shutterstock)

In a blog post, Director of Product Management for YouTube Youth, James Beser, wrote that "machine learning age estimation" is being employed to protect teens on YouTube.

Distinguishing the new measures from previous incentives such as YouTube Kids and supervised accounts for teens, Beser explained that the platform would work towards presenting age-appropriate videos.

"Over the next few weeks, we’ll begin to roll out machine learning to a small set of users in the US to estimate their age, so that teens are treated as teens and adults as adults," said Beser.

Machine learning will record the searches and video categories, and the AI will make decisions based on these and other factors, including the age of the account.

"This technology will allow us to infer a user’s age and then use that signal, regardless of the birthday in the account, to deliver our age-appropriate product experiences and protections," Beser adds.

These protections include disabling personalized advertising when children are watching, and switching on digital wellbeing tools. (These currently include break reminders, daily timers, and even a bedtime reminder.) In addition, recommendations will be safeguarded, with blocks on some repeated content.

These measures would seem to be VPN-proof, too. There would be no point any government banning VPNs if the AI-based measures are universal and cannot be circumvented.

What's next?

Over the years, criticism of YouTube’s granular content controls has resulted in a subscription service (YouTube Kids), supervised accounts, and Google Family Link integration.

The timing of this announcement, following the decision to include YouTube in Australia’s under-16s social media ban, cannot be a coincidence.

After all, it looks like this is not just a US-only endeavor. “We are now bringing it to the US, and as we make progress, we’ll roll it out in other markets,” Beser writes.

Considering recent events, the UK would seem to be high on that list of markets, followed by key EU nations imposing age verification measures.

You might also like

• The rise of AI on YouTube is changing the platform I grew up with, and not for the better
• Age verification requirements have landed in the UK – how the internet will change, and what about your privacy?
• US age verification laws put your privacy at risk – and "VPNs are not a solution”

• Big Sleep AI-powered vulnerability hunter built by DeepMind and Project Zero
• The first batch of 20 vulnerabilities it has spotted have been announced
• Details are under wraps to give devs time to patch them

Google's AI-powered tool designed to find bugs, Big Sleep, has reported its first batch of 20 security vulnerabilities in open source software.

Developed by AI and security teams from Google's DeepMind and Project Zero, the first vulnerabilities were found in the likes of FFmpeg and ImageMagick, however details of those vulnerabilities remain undisclosed until they have been patched.

Google says Big Sleep marks a significant step forward in app security, with AI capable of autonomously uncovering and reporting vulnerabilities more effectively than human security workers.

Big Sleep digs up the dirt on open source software bugs

Each of the 20 bugs was found and reproduced autonomously by Big Sleep, though Google notes that a human expert does review the findings before making reports public - with human review important to temper worries over false positives or hallucinated bugs by ensuring the issues are worthy of being reported to their respective developers.

Finer details like CVE IDs, technical explanations and proofs of concept are withheld for now under Google's 90-day policy to give developers time to patch the vulnerabilities without attackers getting in first.

"By November 2024, Big Sleep was able to find its first real-world security vulnerability, showing the immense potential of AI to plug security holes before they impact users," President of Global Affairs Kent Walker boasted in a blog post.

VP for Security Engineering, Heather Adkins, announced the news in an X post: "Today as part of our commitment to transparency in this space, we are proud to announce that we have reported the first 20 vulnerabilities discovered using our AI-based "Big Sleep" system powered by Gemini."

Google keeps a full list of vulnerabilities, which currently includes the first 20, separated into high, medium and low impact issues.

Google plans a full technical briefing at the upcoming Black Hat USA and DEF CON 33 events, and will donate anonymized training data to the Secure AI Framework so other researchers can benefit from the tech.

You might also like

• Check out the best AI tools and best AI writers
• Downloaded something dodgy? Here's the best malware removal
• One of Google's "big AI" projects uncovered some serious security threats seemingly all on its own

• A limited edition 8BitDo controller celebrating Rare's 40th anniversary has been spotted at Amazon
• The blue/yellow gamepad features emblems from the developer's most iconic titles
• It's revealed shortly after the cancellation of Everwild, which would've been the studio's next big game

It's definitely been a trying year for Rare, but there is at least something to shout about for fans of the legendary Twycross-based developer in the form of a limited edition 8BitDo controller.

This 8BitDo Ultimate controller for Xbox consoles and PC has been spotted over at Amazon (via Eurogamer), and is intended to celebrate the company's 40th anniversary. No release date has been revealed yet, and it seems like it'll only be available to US audiences at present.

It's definitely one of the nicest limited edition 8BitDo controllers to date, featuring a blue and yellow color scheme reminiscent of the developer's logo. It's also emblazoned with iconography from some of its most renowned titles, including Sea of Thieves and Banjo-Kazooie. There's even a banana there as a cheeky nod to the Donkey Kong Country series as well as Donkey Kong 64.

If you're unfamiliar with the 8BitDo controller, it's a bit of a revelation in the third-party gamepad space. It features Hall effect thumbsticks made to combat drift, two remappable buttons for secondary inputs, and a 2.4GHz dongle for wireless connectivity.

The 8BitDo Ultimate also includes a charging dock in the box, for a clever and easy way to keep the pad topped up. I consider it to be one of the best Xbox controllers and best Nintendo Switch controllers on the market today.

I can't help but feel bittersweet about the reveal of this controller, though. While Sea of Thieves is a tremendous game and continues to enjoy a dedicated player base, Rare's 40th year hasn't exactly been smooth-sailing.

The developer's next game, Everwild, was recently canceled. This is in spite of comments from Xbox boss Phil Spencer back in February of this year, where he claimed the game's development was progressing. Rare also created the Perfect Dark series, an immersive sim-esque reboot of which was also shuttered along with its developer The Initiative amidst another round of Xbox Game Studios layoffs.

You might also like...

• Hysilens is here to save DoT in the Honkai: Star Rail 3.5 update – and I couldn’t be happier
• From edgerunners to evil wrestlers, here's every big fighting game announcement at EVO 2025
• The Turtle Beach Victrix Pro BFG Reloaded could make one of the best PS5 controllers even better

Marvel Rivals Season 3.5 will launch next week, adding a brand new Hero into the mix, and refreshing Competitive Mode ranks and rewards. We've now had the full reveal on what's set to change, including details on Blade, the new Team Ups, and all incoming buffs and nerfs.

What's new in Marvel Rivals?

(Image credit: NetEase)

Marvel Rivals will be launching into Season 3.5 later this week, and NetEase has shared the full roadmap for the new update. Looks like we'll be getting a new event alongside the launch of Season 3.5, and can look forward to skins for Peni Parker, Scarlet Witch, and Iron Man in the future. For full info on what's set to be added on August 8, visit our Marvel Rivals patch notes guide.

In my Marvel Rivals review, I praised the game's huge suite of Heroes, and the surprising amount of depth found when trying out each one. With Season 3.5, there's set to be another Hero added, as well as an overhaul to the roadmap. Since its launch, Marvel Rivals has changed quite a bit, adding smart changes to Competitive Mode while giving players something to work toward. It's consistent updates like this that make Marvel Rivals a worthy inclusion in our best PC Games of 2025 list.

Here's everything you need to know about the next season of Marvel Rivals, including when the current one ends, and what to expect from the new Heroes that are being added. As release day draws nearer, I'll update this page with all of the latest news and info.

When is the next season of Marvel Rivals?

(Image credit: NetEase)

Marvel Rivals Season 3.5 will launch following downtime that's planned for August 8, 2025. This is according to recent info given by NetEase alongside the full reveal of Season 3.5.

When does the current season of Marvel Rivals end?

(Image credit: NetEase)

Marvel Rivals Season 3 will end on August 8 at 5AM EDT / 2AM PDT / 10AM BST, for a few hours of downtime before the launch of the new season. Keep an eye on the official Marvel Rivals website for any changes.

Marvel Rivals Season 3.5 new Team Up changes

(Image credit: NetEase)

Marvel Rivals Season 3.5 brings in a bunch of new Team Ups, while removing existing ones. The first new one adds Luna Snow and Adam Warlock, unlocking powerful synergistic healing. Here's what changed:

• Duality Dance: Luna Snow, Adam Warlock (NEW)
• Vibrant Vitality: Mantis, Groot, Loki (NEW)
• Ever-Burning Bond: Human Torch, Spider-Man (Balance Changes)
• Atlas Bond: Luna Snow, Iron Fist (Removed)
• Guardian Revival: Adam Warlock, Star Lord, Mantis (Removed)
• Chilling Assault: Luna Snow, Hawkeye, Iron Fist (Iron Fist added)
• Lunar Force: Cloak & Dagger, Moon Knight, Blade (Blade added)
• Ragnarok Rebirth: Hela, Loki, Thor (Loki adjusted)
• Rocket Network: Peni Parker, Rocket Raccoon, Star Lord (Star Lord added)

Marvel Rivals latest Hero balancing changes

In the Developer Vision video you can view above, some general balance updates applied to Marvel Rivals alongside the latest seasonal update are explained in full. We've collected the key points in the list below:

• Hero balancing (buffs): Star Lord, Iron Fist, Spider-Man, Dr Strange, Groot, Thor, Adam Warlock,
• Hero Balancing (nerfs): Black Panther, Wolverine, Magneto, Mantis, Cloak & Dagger
• New Hero schedule: One per month from Season 3 onwards
• New mode: Resource Rumble
• Maps: Throne of Knull
• Competitive Mode: New ranked rewards, increased penalties for quitting, point compensation
• Misc: Voice-Chat monitoring, text chat filtering, shader compilation changes

Marvel Rivals Blade

Blade is the main Hero being added for Marvel Rivals Season 3.5. He's a Duelist that wields a shotgun and a sword. He can use the sword to block incoming damage, and has an arcing leap ability to close the distance. His Ultimate ability deals massive damage over a large area with the sword.

Marvel Rivals next season FAQWho was the last Hero to be added to Marvel Rivals?

Phoenix was the last Hero to be added to Marvel Rivals. She's a ranged Duelist that deals great burst damage with flaming attacks. She has a dash to escape diving enemies, and her Ultimate is great for clearing the battlefield of support items like spider-mines, Loki clones, and Rocket revive hubs.

Who are the next characters coming to Marvel Rivals?

The next Hero that'll be added to Marvel Rivals is Blade. Beyond that we don't know who's set to arrive in Season 4, though we only have a month to wait now.

What season is Marvel Rivals in?

Marvel Rivals is currently in Season 3. This started on July 11, and should run through to August 8. We'll then be in Season 3.5.

You Might Also Like...

• Will Marvel Rivals be on Nintendo Switch 2?
• Best free games to play in 2025
• When to expect the next season of Fortnite

When I first started as an engineer in the semiconductor industry, we worked on pretty boring electronics. Laptops, portable phones, gaming computers – not the most exciting stuff by today’s standards. Then some smart engineer on the U.S. West Coast took a portable mobile phone and a portable computer, stuffed them into each other and called it a smartphone. It’s basically a data display device. And we were super proud of our smartphones.

And when these data display devices were combined with big storage and big compute in the cloud, well that’s what enabled an on-demand world. A world that allows us to order whatever we want with only a couple of clicks.

Now, after decades of technology advancements, we are moving from an on-demand world, to one that anticipates our needs and automates to address them. In this world an ever-growing number of connected systems – cars, smart homes, factory floor sensors, healthcare devices – process data right where it’s captured.

At the intelligent edge, manual devices transform into autonomous and responsible robots. These robots will be powered by engineering innovations, new design processes, and advances in sensors and AI.

What can we expect in this new world? Think about this for a minute. Over the next few decades, our homes will be able to predict maintenance needs, keep our families safe, and even order food for the fridge. Sound more like dream than reality? I am telling you - it is not. And it’s more than just our homes. Driving will be entirely automated too, and cars themselves will be convenience spaces where you can relax or work while you ride.

This intelligent world is closer than you think. But how will we get there?

Creating our digital twins

One of the most important steps toward autonomous and responsible robots has been underway for decades: the creation of digital twins. These are virtual models of physical objects, located within the cloud. For an individual, that might be the state of your health, wealth, and your physical presence. It also applies to homes, businesses, hospitals, and even cars.

But simply digitizing the physical world will only take us so far. We need to enable this digital world to reach out to its physical counterpart. In doing so, these digital twins will be able to connect with each other, optimize and learn from each other. Then, and this is the critical part, they apply that knowledge in the real world. Only once we achieve this, can autonomous robots truly become a reality.

From manual machines to autonomous robots

We need to enable machines to sense, think, connect, and act in our physical world. And what is most important, we need to make sure they always do so responsibly – with safety and security at the core of everything they do. Because you will never hand over control to a robot that you do not trust.

Largely, building acting machines has been achieved over the past 140 years. These machines have just always required a human to provide input or oversight. More recently, connectivity challenges have been solved in the 1990s and early 2000s. Now our real challenges lie in enabling machines to sense and think.

The automotive industry is a clear example of where we’ve fallen short here so far. Around 2016, everyone thought that self-driving cars were around the corner. Theoretically, we had the technology to make them work. Yet fully autonomous driving still remains out of reach. So, what went wrong?

The gap between the autonomous driving future that was predicted, and today’s reality comes from a fundamental misunderstanding of AI systems. We thought that simply having an AI system that’s trained on how we drive would be enough.

It’s like expecting to be able to hand the keys to your teenager and letting them drive simply because they’d been in the car with you for years. In the real world, people need to train and pass deterministic tests before they are given a driver’s license. That layer of trust, safety, and security is what was missing.

Enabling the brain shift

To get the safety and security part right, today’s AI (the brain of the robot) requires a new approach. And where better to look for inspiration than the human brain itself.

Our brains are largely broken into three areas: the cerebrum facilitates perception, the cerebellum coordinates action and vital functions, and the brain stem regulates real time functions and powers reflexes. For humans, all of these are crucial. But for robots, it depends on their use.

Back to the autonomous car example, the highest priority is function and safety. For this, we need reflexes and coordination in conjunction with sensors. For a self-driving car, that translates at a base level to functional and safe power management and a real-time neural information transportation system.

Or in other words, reliable Power Management Integrated Circuits (PMIC) and processors that can handle all the information that comes in from a vast array of sensors.

Beyond that, you need modular software building blocks. This is because it’s software that defines how autonomous vehicles function. Having pre-built blocks of software also means that production can be adapted even at scale.

If you’ve got building blocks for compute, networking, power management, and more, you can minimize the time spent on basic functionality. With that, you can invest more effort into bringing products to market or solving difficult challenges.

While self-driving cars are the most familiar example today, this brain shift taking place is laying the groundwork for other intelligent machines in the future.

Building on intelligent foundations

Re-engineering the robot brain is important, but it’s not only thing needed for a world that anticipates and automates. We also need ongoing improvements to sensors, along with a common language to enable interoperability across these robots. These are all areas where we are making rapid progress: with high-resolution radar, ultra-wideband signals, and the Matter standard to name just a few innovations.

A better world built around autonomous and proactive robots might feel like a concept from science fiction. But if you look at the advances making our vehicles, our homes and our factories smarter and safer today, it’s clear we’re laying the right foundations.

We have kicked the door open towards a world that anticipates and automates – now industry players, academia, researchers, engineers, and policy makers have the chance to bring this “robot awakening” to reality – and create a truly intelligent and trusted future.

We list the best Robotic Process Automation (RPA) 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 European Accessibility Act (EAA) took effect on June 28, 2025, introducing new legal requirements for digital accessibility across the European Union (EU).

A 2024 WebAIM study found that 97% of the world’s leading websites do not currently meet compliance standards. The EAA applies to companies providing digital products and services in the EU, regardless of where they are based. This means businesses in the US, Asia and elsewhere will also need to ensure compliance.

Many companies are unprepared because accessibility is not seen as a priority. According to a recent Storyblok report, 18.5% of surveyed companies were completely unaware of the EAA, and 16% had not yet started making necessary changes. Even among those that have knowledge of the EAA, only a quarter feel fully prepared.

A lack of resources, technical complexity and low awareness at the leadership level contribute to the slow pace of progress. Accessibility is also often treated as a compliance issue rather than an integral part of digital strategy, which leads to delays in implementation.

Failure to comply

Failing to comply with the EAA comes with financial and operational risks. Fines range from €5,000 to €20,000 per violation, with additional penalties for ongoing non-compliance. Businesses that do not meet the requirements may also face reputational damage, loss of customers and restrictions on selling products or services in the EU.

Putting aside the legal implications, it’s worth remembering that accessibility affects a significant portion of the population, with more than 135 million people in the EU living with some form of disability. Companies that do not address accessibility risk excluding a substantial number of potential customers.

Businesses do not need to overhaul their entire IT infrastructure overnight, but they do need to start making changes. Conducting an accessibility audit is a good first step. Websites and digital products should be tested against Web Content Accessibility Guidelines (WCAG) 2.2 standards.

While automated tools like Lighthouse and Axe can help flag common issues, manual testing is also necessary. Involving users who rely on assistive technologies provides valuable insight into usability challenges.

Improving accessibility

A headless content management system (CMS) can also make accessibility improvements easier. Separating content from design allows businesses to implement accessibility features consistently across different platforms without needing to redo everything from scratch. Companies should also prioritize smaller, manageable updates such as adding alternative text to images, improving color contrast and ensuring keyboard navigation works correctly while planning more extensive updates in the long term.

Employee training is critical. Accessibility needs to be embedded into company culture and everyday business processes. Regular training ensures that developers, designers and content creators understand best practices and apply them in their work. Compliance teams should also keep up with regulatory updates, as accessibility requirements will continue to develop.

Beyond compliance

Beyond compliance, accessibility offers a business advantage. Research from Storyblok shows that 69% of companies recognize accessibility as a way to improve customer experience and engagement. Digital products designed with accessibility in mind tend to be more user-friendly for everyone, not just those with disabilities. Investing in accessibility now can also prepare businesses for future technological developments, such as voice search and AI-driven user interfaces, which will benefit from inclusive design.

A common misconception is that improving accessibility means starting from scratch. In reality, small changes can have a big impact. Adding captions to videos, ensuring text descriptions for all non-text content and enabling text resizing without loss of functionality are straightforward but effective adjustments. Keyboard accessibility is another critical aspect, as many users rely on keyboards or alternative input devices to navigate digital interfaces.

Businesses that prioritize accessibility often see direct commercial gains. For example, retailers that optimize their websites for accessibility have reported increased conversions, as customers with disabilities, often an underserved market, find it easier to complete transactions. Financial institutions that improve accessibility features in online banking have seen an increase in customer retention and trust. The demand for inclusive digital experiences is growing and organizations that act now can establish themselves as leaders in this space.

Increasing scrutiny

Regulatory bodies are also increasing scrutiny on compliance. In some countries, lawsuits related to web accessibility have surged, with high-profile cases leading to costly settlements. Large companies have faced legal action for failing to provide accessible digital experiences and these cases are expected to become more common as awareness grows. For multinational businesses, aligning with accessibility standards is not just about meeting the minimum EU requirements but also establishing best practices that will serve them globally.

There is no time to waste. The transition period before enforcement is running out and businesses that wait until the last-minute risk rushed implementations that may still fall short of compliance. Accessibility should be viewed as an ongoing process rather than a one-time fix. Regular audits, user testing and engagement with accessibility experts can help companies stay on track and adapt to future regulations.

Technology also continues to develop and businesses that embed web accessibility into their digital strategies will be better positioned for emerging innovations. AI-driven accessibility tools, voice interfaces and adaptive design techniques are transforming how users interact with digital content. Organizations that invest in accessibility now will be better prepared for these shifts making digital services easier to use across a wider range of users and devices.

Investing in accessibility

Organizations that take a structured approach - starting with audits, implementing incremental improvements and investing in accessibility training will be in a stronger position to comply with regulations and serve a wider audience.

Accessibility should be treated like any other core compliance risk: it needs clear ownership, regular reporting, and cross-functional collaboration. Working closely with developers, designers, and legal teams will help surface issues early and prevent gaps in future projects. Waiting until next year is likely to mean rushed fixes and greater exposure. Building the right internal processes now will reduce that risk and make ongoing compliance more manageable.

Compliance professionals must act decisively. The EAA is a fundamental shift in how digital services must be designed and delivered. Failing to act in time could expose organizations to legal scrutiny, increased litigation risks and potential exclusion from key markets. Regulators will not view accessibility failures as minor oversights but as breaches of consumer rights.

Key Takeaways

Compliance teams should take a proactive approach, conducting thorough risk assessments to identify where their digital services fall short. Establishing internal policies and governance frameworks will be critical, ensuring accessibility is embedded into all future development work. Training employees, integrating accessibility checks into procurement processes and working closely with IT and legal teams will help organizations maintain compliance beyond the 2025 deadline.

Meeting accessibility standards requires sustained effort. Ongoing updates, oversight and accountability are essential to remain compliant over time. Businesses that prioritize it now will meet regulatory demands but, more importantly, position themselves as leaders in ethical and responsible digital practices.

Key Takeaways:

1.Businesses operating in the EU must comply, regardless of where they are based.

2.A lack of awareness, technical challenges, and limited resources are common barriers.

3.Non-compliance carries financial penalties and reputational risks. It can also limit market access within the EU.

4.Practical steps, such as accessibility audits, CMS updates and staff training, can help businesses meet requirements.

5.Improving accessibility benefits all users and makes digital platforms more user-friendly. It is a long-term investment, not just a compliance issue.

We list the best web design 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

Security operations are under pressure from all sides. Threats are faster, attack surfaces are expanding, and demands on people and tools continue to grow. At the center of it all, one constant holds: Security Information and Event Management (SIEM) remains a cornerstone of the modern SOC. According to a recent 2025 Security Operations Insights Report, nine in ten security and IT leaders still consider SIEM essential to safeguarding their organizations.

This underscores a core truth: SIEM isn’t outdated. It’s indispensable. But like any critical system, it must adapt to meet today’s realities and tomorrow’s risks.

The future of SIEM isn’t about ripping and replacing the industry itself. It’s about reimagining how it can better serve security teams, moving from static log aggregation and reactive alerts to intelligent automation, real-time insight, and proactive defense. The shift is already underway with AI as the catalyst. It’s changing not only what SIEM platforms can do, but how teams interact with them day to day.

The limitations of traditional SIEM

SIEM emerged in response to the industry demand for centralized visibility and log correlation across digital environments, enabling teams to sift through overwhelming event data and generate alerts in early-generation SOCs. And while first-generation SIEM delivered many improvements to SecOps, it has long struggled with more sophisticated capabilities like real-time analysis and alert accuracy.

Over time, these shortcomings have intensified. Security teams receive thousands of alerts per day across countless services, and nearly half of those alerts remain uninvestigated due to their volume and talent scarcity. Workflows are fragmented, triage is time-consuming, and teams are forced to manually gather context across disparate tools.

These daily pressures are contributing to widespread burnout and fatigue across the cybersecurity workforce, costing U.S. enterprises over $600 million in lost productivity each year. The result is slower detection, delayed response, and greater risk exposure.

This points to a growing disconnect between what SIEM delivers and what organizations need. While the core concept behind SIEM remains essential, most tools today fall short of delivering the speed, scalability and intelligence required to defend today’s digital environments. The Security Operations Insights Report also found that, of the security and IT leaders that view SIEM as relevant, three-fourths are actively considering alternatives.

The case for Intelligent SecOps

The growing strain on security teams has made one thing clear: SIEM platforms have an opportunity to evolve into a service that realistically supports the needs and environments that teams work in today. Intelligent SecOps represents this shift: a model where the core principles of SIEM are preserved, but transformed through AI, automation and cloud-native scale.

According to the same survey, 90% of security leaders see AI as an extremely or very important factor in their decision to adopt a new security solution. These leaders are looking for tools that not only collect data, but help them act on it – faster, smarter, and with greater context.

1. Smarter Triage: Less Noise, More Signal

AI models help reduce false positives by continuously learning from threat intelligence, analyst feedback and environmental patterns. By enriching and prioritizing alerts, these systems elevate the most actionable signals, helping teams focus on the threats that truly matter.

2. Automated Investigations and Contextual Enrichment

Modern SIEM platforms powered by AI offer more than detection. Rather, they automate early-stage investigations by enriching alerts with context, mapping related events and visualizing likely attack paths. Assistive tools like AI copilots can surface key insights instantly, reducing manual work and accelerating decision-making.

3. Proactive Threat Detection with Behavioral Analytics

AI tools enable behavior-based detection that goes beyond static rules or known indicators. By identifying deviations from normal patterns across users, endpoints and applications, these systems surface stealthy or evolving threats. Integrated frameworks like MITRE ATT&CK help contextualize behaviors and link them to known adversary tactics.

4. Accelerated Response Through Automation

With enriched alerts and intelligent correlation, teams can move faster from detection to containment. AI-powered workflows and playbooks enable automated responses, such as isolating hosts or disabling credentials, reducing the window of exposure and freeing analysts to focus on strategic analysis.

5. Cross-Environment Correlation and Real-Time Normalization

As digital environments stretch across cloud, on-prem and SaaS, AI helps normalize and correlate telemetry in real-time, surfacing threats that span infrastructure boundaries. This eliminates blind spots and supports unified investigation across an increasingly complex attack surface.

SIEM as a strategic partner

As threats grow more dynamic and resources remain constrained, the tools security teams rely on must become more than just dashboards. They must become intelligent partners. The evolution toward Intelligent SecOps is not just a technology upgrade. It’s a shift in how teams work, how they scale and how they think about risk. It reflects a broader change in mindset, away from reactive firefighting and toward resilient, intelligence-led operations.

The SOCs of tomorrow will not be defined by how many alerts they generate, but by how intelligently and efficiently they respond. AI-powered SIEM is at the heart of that move towards Intelligent SecOps, bringing clarity to chaos and action to insight.

We've listed the best Robotic Process Automation (RPA) 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

• New images claim to show the iPhone 17 Air’s battery
• It looks to be about half the width of the iPhone 17 Pro’s battery
• Apple is reportedly working on advanced battery tech for the device

The iPhone 17 Air could be one of the best iPhones in many years, and with just a month until Apple is expected to unveil it, details and specifications have started to leak out. The latest rumor concerns the phone’s battery, and it raises some important questions about Apple’s slimline device.

As the name suggests, the iPhone 17 Air is tipped to be the thinnest iPhone in recent memory. That means it’s going to need a super-svelte battery on the inside, and that’s exactly what new images from Korean blog site Naver appear to detail.

There, leaker yeux1122 posted two images showing what they purport to be the iPhone 17 Air’s battery. In the first picture, the leaked battery is compared to one from the iPhone 17 Pro, with yeux1122 claiming that the former is just 2.49mm thick. Judging it by eye, it looks to be about half the width of the iPhone 17 Pro’s battery, which itself was leaked just a few days ago.

Battery capacity versus battery life

(Image credit: yeux1122)

Clearly, cutting down the battery dimensions is necessary in a phone as slim as the iPhone 17 Air is expected to be. But that’s not the only consideration for Apple: it will need to ensure that the longevity of the battery is not negatively impacted by its size.

A previous leak from yeux1122 claimed that the iPhone 17 Air’s battery would have a 2,800mAh capacity. That’s notably less than the 3,582mAh battery in the iPhone 16 Pro and well below the 4,685mAh battery you’ll find in the iPhone 16 Pro Max.

Still, that doesn’t necessarily mean that the iPhone 17 Air will have poor battery life – after all, much of it comes down to how demanding the hardware and software are, and Apple is traditionally very good at optimizing these to eke out more battery life – something it’s rumored to be doing in iOS 26.

Also, Apple analyst Ming-Chi Kuo has previously claimed that Apple will use a “high-density” battery, while other rumors have pointed towards Apple using a new kind of advanced battery tech in the iPhone 17 Air.

Right now, this is all hearsay, and we won’t know for sure how well the iPhone 17 Air’s battery will perform until the device goes on sale in September. Despite its supposedly ultra-thin dimensions, it might still post respectable battery numbers – but all we can do for now is wait for further news and leaks to emerge.

You might also like

• The iPhone 17 Air may have a small battery, but l don't think you'll need to worry
• iPhone 17 Air leaks suggest it'll get next-gen battery – and offset the 17 Pro Max's weight gains
• New iPhone 17 Air leak hints at how the super-slim phone might solve its potential battery life issues

• Google just showed off its forthcoming Pixel 10 in a fresh teaser
• The 30-second advertisement not-so-subtly takes aim at Apple delaying the new Siri
• It tells the viewer to "Ask more of your phone."

We’re just 16 days away from Google showing off its Pixel 10 family alongside some other devices, and a new teaser gave us a fresh look at the smartphone, but also a reminder from the tech giant.

Alongside some sneaky shots of the left and right side as well as a nearly full reveal of the rear of a Pixel 10, including the fact that the ‘camera bar’ is sticking around whether you like it or not, Google writes, “Ask more of your phone.”

It’s a common theme of the teaser advertisement dubbed “Google Pixel 10 | Soon” and not so casually calls out Apple over its delay in rolling out the AI-powered Siri. While the advertisement doesn’t mention Apple, the narration kicks things off by saying, “If you buy a new phone because of a feature that’s coming soon…” It goes on to remind the viewer that if you’re still waiting over a year for it, how you define ‘soon’ could very well change.

And on Instagram, where Google shared it from several handles, it invites us to "Get outside your comfort phone". It's clever, I'll give them that.

Google is likely hoping that you’ll get the hint swiftly and switch to a Pixel 10 or one of its other new devices. Not so subtly, the backing track to this 30-second teaser is an instrumental version of “The Next Episode” by Dr. Dre.

Of course, it highlights that the most recent Pixel 9 family – 9, 9 Pro, 9 Pro XL, and 9 Pro Fold – all came with a bounty of AI software and features available out of the box. These devices, which all feature Google’s Gemini AI models, let you take group photos with fewer hands, and all come bundled with a very capable assistant. These phones can also search what’s on your screen – Circle to Search – and have a new tool called Pixel Screenshots.

(Image credit: Google)

Apple originally announced all of its AI features under the Apple Intelligence umbrella in June of 2024, shipped the iPhone 16 family, which was built for Apple Intelligence, in September of 2024, and then launched the first set of those features back in October of 2025.

In the Spring of 2025, Apple announced a delay in the AI-powered Siri. In an interview with TechRadar’s Editor-at-Large, Lance Ulanoff, Apple’s Craig Federighi explained the delay, noting that it was taking a bit longer and was expected to be resolved next year.

Apple has since been a little more transparent on this, with Tim Cook noting on the Q4 earnings call that the teams are making good progress and that it should be set to ship next year, in 2026. Even so, that’s a long wait for the AI-powered Siri, one that many were hoping to get their hands on – myself and many of my colleagues included.

(Image credit: Google)

With Google setting the stage for its device unveil on August 20, 2025, which is likely several weeks before Apple sets a ‘special event’ to show off the iPhone 17 family, it seems that team Pixel just wants to get ahead and poke some fun at the same time.

Google is set to unveil these new devices – we’re expecting the Pixel 10, Pixel 10 Pro, and Pixel 10 Pro XL – just weeks after Samsung dropped the Galaxy Z Fold 7, Galaxy Z Flip 7, and Galaxy Z Flip 7 FE. According to leaked pricing, though, Google might be holding steadfast without increasing the cost for its main phones – that’ll be a delight to our wallets.

We’re also not expecting a drastic redesign for the Pixel 10 family, maybe a few new colors, but the modern look that Google ushered in with the Pixel 9 is likely here to stay, with this teaser basically confirming it. Even so, though, it’s likely Google has a new silicon Tensor processor powering these and some impressive AI features in the pipeline.

Considering there’s still over two weeks before Google’s event, I’d be surprised if we don’t see some more teasers. Let’s just see how direct they get, and if Google opts for a change in artist for the music accompanying them.

You might also like

• The Google Pixel 10 just leaked again, and it looks like it's getting a big charging upgrade
• 6 of the best Android phone accessories for photography, charging, tracking, and more
• Watch the Samsung Galaxy Z Flip 7 go through a torturous durability test that I can't believe it survived

• Micron 9650 SSD reaches 28GB/s, but it’s strictly for data centers, not consumers
• Liquid cooling in SSDs is no longer hypothetical, and Micron just made it standard
• The 7600 SSD delivers ultra-low latency, yet it still depends on how workloads behave in reality

Micron has introduced what it describes as, “the world’s first PCIe Gen6 data center SSD,” with claims of as-yet-unmatched performance tailored for modern AI workloads.

The 9650 SSD reportedly hits sequential read speeds of up to 28 GB/s and sequential write speeds of up to 14 GB/s.

However, it comes in industrial-grade E3.S and E1.S form factors, making it incompatible with standard desktop PCs, limiting its direct accessibility to broader consumer markets.

Designed for performance but constrained by form factor

This new model targets high-intensity AI environments, offering 5.5 million IOPS in random read performance and up to 900K IOPS for random writes.

Micron’s 9650 improves on Gen5 SSDs with up to 25% and 67% greater energy efficiency for random writes and reads.

It also incorporates liquid cooling options for dense server configurations, and its reduced power draw and emissions support both performance gains and sustainability efforts in data centers.

“With the industry’s first PCIe Gen6 SSD, industry-leading capacities and the lowest latency mainstream SSD - all powered by our first-to-market G9 NAND—Micron is not just setting the pace; we are redefining the frontier of data center innovation,” said Jeremy Werner, senior vice president and general manager of Micron’s Core Data Center Business Unit.

While these specifications may sound impressive on paper, the real test will be in sustained, real-world workloads under diverse operating conditions.

Several vendors have highlighted its potential in supporting inference pipelines and retrieval-augmented generation, suggesting the 9650 could serve as a key infrastructure component for GPU-based servers.

Still, wider adoption will likely hinge on pricing, reliability, and actual ecosystem integration.

Alongside the 9650, Micron also unveiled its 7600 SSD based on PCIe Gen5 and the Micron 6600 ION SSD, which focuses on capacity.

The 7600 claims to deliver sub-1 millisecond latency on demanding database applications like RocksDB.

With read speeds reaching 12 GB/s, the 7600 outperforms existing Gen5 SSDs in metrics such as random writes and energy efficiency.

But the claims of having the fastest SSD must be balanced against actual deployment flexibility and sustained workload performance.

“Micron’s cutting-edge storage technologies showcase the importance of fast, efficient storage as AI workloads continue to redefine infrastructure requirements,” said Raghu Nambiar, corporate vice president, Data Center Ecosystems and Solutions, AMD.

You might also like

• Here are the best mobile workstations around today
• Take a look at some of the best external hard drives
• Google Workspace is hitting back against the most prolific methods of account takeover with these super simple changes

• A leaked render of the DJI Mini 5 Pro seems to reveal the drone's design
• A separate image of the box may have also revealed its specs
• The leaks hint at a 1-inch sensor and LiDAR-powered autofocus

The DJI Mini 5 Pro seems unlikely to hit its previously-rumored August 7 launch date, but fans of tiny drones have at least been given some solace in the form of two big leaks – and they hint at a little flying camera that could be worth waiting a little longer for.

Firstly, a render of the rumored successor to one of the world's best drones was shared on the Discord channel for Drone-Hacks, giving us what could be our first look at the drone.

If the image is correct, the Mini 5 Pro will look very similar to its Mini 4 Pro predecessor, aside from a larger camera module and the inclusion of two forward-facing LiDAR sensors for improved obstacle avoidance.

That LiDAR-powered feature has previously appeared on the DJI Mavic 4 Pro and DJI Air 3S, but this would be the first time we've seen it on a Mini series drone – and it could be particularly useful if you fly in low light.

Mini 5 Pro leaked render! The leak didn’t come from our team, some beta tester shared a picture on the public Discord. Thanks for the credit tho! #Mini5Pro https://t.co/LV1eQJCYKDAugust 3, 2025

Even more revealing is the potential list of specs, shared by @JasperEllens and DroneXL. According to some leaked packaging, the DJI Mini 5 Pro will have a 1-inch image sensor – up from the 1/1.3-inch chip on the Mini 4 Pro – and be capable of shooting 4K/120fps footage, another slight boost from its predecessor's 4K/100fps mode.

Elsewhere, there's the promise of a "high quality 48mm med-tele mode", which will presumably be similar to the one that recently arrived on the DJI Pocket 3, thanks to a firmware update. This is effectively a 'lossless' crop mode that can be handy if you need a bit more reach.

The leak also promises a 36-minute flight time, which would be a slight boost on the claimed 34 minutes managed by the Mini 4 Pro (or 45 minutes, if you shell out for the Intelligent Flight Battery Plus, which isn't available in the EU).

The big question

(Image credit: DJI)

While those Mini 5 Pro specs sound very promising indeed – and could make me consider upgrading my Mini 3 Pro drone – there is one slightly worrying omission from the box and specs sheet.

Usually, there's a "less than 249g" tag accompanying the "ultra-light and foldable" description on a Mini series box. That isn't there this time, at least not on the leaked packaging.

Does this mean the Mini 5 Pro could be the first in the series to exceed that crucial weight? In many regions, including the US and UK, a sub-250g weight is a selling point because flying regulations are more relaxed. For example, in the Open A1 category in the UK, you can fly over strangers (but not crowds) without needing any extra permission.

There is a chance that the Mini 5 Pro's new tech (bigger camera, LiDAR sensors, perhaps new motors) have pushed it over the weight limit, but I'd be surprised if that was the case.

The whole selling point of the Mini series has been that they offer advanced camera tech in a bundle that's barely bigger than a camera lens, and also ducks under that crucial weight barrier. I love my DJI Mini 3 Pro for all of those reasons, so a new model that doesn't tick all of the same boxes wouldn't have the same appeal.

Unfortunately, it appears the Mini 5 Pro has been postponed from its previously rumored launch date, so we may have to wait a while longer to see it – and to find out if it'll actually go on sale in the US.

You might also like

• DJI Mini 5 Pro leaks give us our first glimpse of the beginner drone and a potential release date – here’s what we know
• I tested DJI’s impressive first 360 camera, but Insta360 still has the edge
• DJI’s first robot vacuum gets launch date in official teaser – and it looks like the Romo is arriving sooner than expected

• Areas of Russia are experiencing mobile internet blackouts
• These are said to be necessary to combat Ukrainian drone attacks
• Many believe the blackouts may be a crack down on digital rights

Amidst the Russian invasion of Ukraine, technology has been a crucial point of leverage for both sides, with technological infrastructure like telecoms and mobile internet services targeted in disruptive offensives aimed at weakening the other side - even leading to Russia shutting down its own internet services in the name of national security.

A new report by non-profit Russian Internet Protection Society has outlined a record breaking 2,000 mobile internet service shutdowns in July, pointing to a dramatic increase in digital restrictions - and rights groups say that many of the blackouts don’t appear to be related to any threat.

Russian authorities have blocked the advertising of VPNs and Cloudflare subnets in a large-scale crackdown on digital rights. Platforms like Twitch, TikTok, YouTube, and even Duolingo are all inaccessible without a VPN.

Drone attacks blamed

These blackouts have a serious impact on life in parts of Russia and cut off access to things like maps, banking applications, buying fuel, or even communicating with loved ones - with talks of blocking WhatsApp suggesting the state pushing citizens onto highly monitored, government controlled messaging services.

Russia has cited a need to prevent and defend against Ukrainian drone attacks, but not everyone is convinced by this explanation. The economic losses from these disruptions are estimated by watchdogs to be around 26 billion rubles ($290 million) in July alone - and rural areas are left isolated.

Dmitry Gudkov, a former MP and co-founder of the Center for Analysis and Strategy in Europe told Le Monde that authorities are ‘gradually restricting freedoms’ through these outages;

"They are acclimating society to life without the internet. The day they need to cut the mobile network, for instance to stifle protests, they'll know they can do it."

Via: The Record

You might also like

• Take a look at our picks for the best firewall software around
• Check out our choice for best endpoint protection software to keep you safe
• Russia blocks almost 200 VPN services, but the Kremlin still wants to use them

• Samsung's Flip 7 went through an extreme durability test
• Its outer screen is hardy, but the inner screen is easily scratched and burnt
• It couldn't be snapped in half by hand, however

Our phone testing is very rigorous, but we focus on what a regular person would get out of the gadget and if it’s better or worse than the competition at the same price; we don’t routinely set fire to phone screens, bathe gadgets in dirt, or attempt to scratch every surface with a knife – but that’s what one tester has done to the Samsung Galaxy Z Flip 7.

In an extreme durability test, which you’ll find several of on his YouTube channel, JerryRigEverything seriously put Samsung’s new foldable through its paces – a follow-up to a similar video involving the Galaxy Z Fold 7 about a week earlier.

In the just under 9-minute video, Samsung’s phone is tortured, but we do learn some useful details.

The outer screen is pretty hardy, standing up to significantly harder materials than the inner screen – which a particularly forceful fingernail scratch is able to create grooves in – and also withstanding direct heat from a lighter’s flame. The inner screen is left with a couple of permanent green damage marks after about 15 seconds of fire exposure.

It then gets drenched in dirt, including getting a healthy heaping dropped straight into its hinge. Opening and closing the phone with dirt covered did allow some soil to enter the mechanism – you can hear it crunch when opening and closing after the fact – but at least during the video, no other permanent harm is noticeable.

Lastly, JerryRigEverything does his best to snap the phone in half with his hands, but it survives his attempts.

You should absolutely go watch the full video – especially if, like me, you’re too cautious with your tech to let the intrusive thoughts win – because it highlights the extreme lengths you have to go to actually tear these foldables apart.

Tougher than you think

Foldables are hardier than we give them credit for (Image credit: Peter Hoffmann)

I’ve used a foldable as my daily driver for about a year. I started with the Galaxy Z Fold 6 and recently upgraded to the Samsung Galaxy Z Fold 7.

Because I get those phones through my job, I’m admittedly a bit more laissez-faire with how I treat them (meaning I’ll buy a cheap screen protector, but not a case), but both phones have been holding up very, very well.

The Z Fold 6 has a few bruises at the edges, but nothing serious, and my Z Fold 7 is close to immaculate. Neither has any issues with their hinges, and while you can see their creases, it’s not super noticeable on either phone when you’re using one.

This brings me to a point I made on our Samsung special podcast. While that IP48 dust and waterproof rating means the foldables can let fine dirt in (anything smaller than 1mm), and their inner screen isn’t as durable as a regular phone display, most normal people probably won’t have any serious issues.

I’ve been to a sandy beach two times in the past year, and visited one desert, without my foldable being affected. I was careful, sure, but no more so than I am with a non-foldable.

If you regularly work at the beach or in, say, a dusty woodwork shop, or in a super-harsh environment, a foldable might not be ideal. But for most folks, these extreme and my year-long tests prove they’re surprisingly durable if you’re kind to them. If durability was the reason you’ve been put off buying one, it might not be as big an issue as you imagine.

You might also like

• The Samsung Galaxy S26 could come with multiple AI assistants
• A YouTuber is folding and unfolding the Galaxy Z Fold 7 200,000 times
• This rumored Samsung Galaxy S26 upgrade will finally make contactless payments as easy as they are on the iPhone

• Kamala Harris says she won't use wireless headphones for security reasons
• Vulnerabilities have been found in many big-name products
• Security risk is very very low, but it's not zero

Everybody loves wireless headphones – but perhaps everyone shouldn't. That's what Kamala Harris suggests, and she knows more than most of us: speaking to Stephen Colbert last week, she explained why her preference for wired earbuds over wireless ones wasn't because she's old-fashioned.

"I know I've been teased about this," she said. "But I like these kinds of earpods that have [a cable] because I served on the Senate Intelligence Committee."

As she explained: "I have been in classified briefings and I'm telling you, don't be on the train using your earpods [she meant AirPods here] thinking someone can't listen to your conversation."

This isn't tinfoil hat territory

With very few exceptions, most wireless headphones connect via Bluetooth – and earlier this year, German IT security analysts identified vulnerabilities in a whole host of headphones from very big names including Bose, Jabra, JBL, Marshall and Sony. The system-on-a-chip manufacturer that those firms use has since updated its SDK to deal with the issue, though we'd be foolish to assume it's the only one.

The presence of a vulnerability in a chip doesn't mean it has been or is going to be exploited. But it's a possibility. Last year BlueSpy demonstrated how to take advantage of such a vulnerability to record audio from "a particular high-end headset". And this year it emerged that a Bluetooth vulnerability could enable eavesdropping on in-car calls.

The risks from wireless devices aren't just interception. For example, the UK's Ministry of Justice notes that Bluetooth devices can be used to track someone's location. It also tells staff that while Bluetooth is fine for people working with "Official"-level material, it may not be safe enough for "Official-Sensitive or higher material".

The likelihood of the average pair of headphones or earbuds being intercepted by anyone is vanishingly small. And that's probably just as well, because 3.5mm audio jacks are vanishing too – although of course you could always use a USB-C adapter or DAC to go wired. But most of the best headphones today are wireless ones, especially for wearing outdoors – and doubly so if you want active noise cancellation. Wired ANC earbuds are nearly extinct, frustratingly for the security

For the average listener like you or I, the security risk is virtually zero. But if like Kamala Harris you don't exactly have an average job, there are plenty of people willing to invest serious time, talent and tech in trying to eavesdrop on your communications. And that could mean that the best wired earbuds are the smart move – going wireless could be going reckless.

You might also like

• Sony, JBL and Bose headphones all affected by major Bluetooth vulnerability
• Bluetooth security flaws could affect thousands of cars
• The best wired headphones 2025 for politicians and for everyone else too

• Sam Altman posts on X about new products over the next couple of months
• Warns of “capacity crunches” and that things “may be slightly choppy"
• ChatGPT-5 possibly delayed as it didn't launch at the start of the month

OpenAI CEO Sam Altman has posted on X.com that, “we have a ton of stuff to launch over the next couple of months – new models, products, features, and more”, but so far there is no sign of the widely expected ChatGPT-5.

Many news outlets have predicted that ChatGPT-5 was going to arrive in August, and it was hotly tipped to be released at the start of the month. However, as the calendar flicked over to August, all we got was the tweet from the OpenAI CEO.

The start of the month would have been the ideal time to launch ChatGPT-5, and the fact that Altman mentions the “next couple of months” means we might have to wait a little while yet for the release of OpenAI’s most impressive model.

we have a ton of stuff to launch over the next couple of months--new models, products, features, and more.please bear with us through some probable hiccups and capacity crunches. although it may be slightly choppy, we think you'll really love what we've created for you!August 2, 2025

At the same time, Altman warned that ChatGPT users could be in for some disruptions over the next couple of months: “Please bear with us through some probable hiccups and capacity crunches. Although it may be slightly choppy, we think you'll really love what we've created for you!”, he said.

Capacity cruches

As if on cue, a ChatGPT user has complained that the service has started to cap the number of image uploads that can be done a day to 30, without any notice.

User Rhizopus_Nigrians says, “This cripples the workflow for serious users. And I am one. I’ve paid for Plus since the beginning. I love it. I’ve promoted this tool to friends, peers, even students”, he posted.

Most of the other Reddit users on the thread point out that this is probably just a temporary bug, rather than an actual policy change by OpenAI. To test his theory I uploaded 40 images today, and there was no problem, however that doesn’t rule out that it could also be part of the “capacity crunches” or “hiccups” that Altman warned about in his tweet.

When will ChatGPT-5 arrive?

ChatGPT 5 is expected to be the first OpenAI model that will automatically route your request to its most appropriate model version, or research method, instead of you having to choose which confusingly-named model version to use. It’s also reported to have key improvements in accuracy, speed, and resource efficiency.

While it’s frustrating that OpenAI has missed the start of the month for a release of ChatGPT 5, there’s still plenty of time left in the month for an August release. What will be interesting to see is whether the release is only for Plus users or if free ChatGPT users will also get access to the new model.

When ChatGPT 5 is released, demand to use it will likely be at an all-time high, especially if it contains any improvements to the image generator, which is why I think Altman is keen to warn people that we may be in for some disruption now.

You might also like

• Rumors of GPT-5 are multiplying as the expected release date approaches
• Apple is ready to catch up in the AI market, CEO Tim Cook says
• Microsoft's up to its old tricks, this time pushing Copilot AI in your face when searching on Bing - and this is getting tiresome

She might be the executive producer of new Hallmark+ miniseries Providence Falls, but when we think of JoAnna Garcia Swisher, we think of Sweet Magnolias. Season 4 dropped on Netflix in February this year, and from the looks of the last episode, everything could be about to change in the fictional town of Serenity. Well, that’s if everyone actually chooses to stay there, with our lead trio’s storylines pulling them in different directions.

For Maddie (Swisher), that means leaving Serenity entirely. She’s been offered a marketing job in New York she just can’t say no to, meaning her weekly margarita nights with Dana Sue (Brooke Elliott) and Helen (Heather Hadley) can be no more. Of course, we don’t actually know if Maddie will really follow through with her relocation, but that’s not the narrative most fans currently have on their minds.

Thanks to an interview with US Weekly, Swisher suggested Maddie could be pregnant in future episodes, sending fans spiraling about whether her family with Cal Maddox (Justin Bruening) could be expanding. "I actually just talked about this yesterday with Justin because I’m like, 'we never got a honeymoon,'" she told the outlet. "I don’t know if that will involve a little bun in the oven – or at least an opportunity.” But is there truth to this? I asked Swisher outright, and there definitely won’t be babies in Sweet Magnolias season 5 by her account.

It’s official: Maddie won’t be pregnant in Sweet Magnolias season 5

The cast of Sweet Magnolias, where nobody is pregnant. (Image credit: Netflix)

“I'm not wearing a pregnancy suit this season,” Swisher clarifies when I ask her about the pregnancy chatter online. “But there is so much that happens, and there's a lot of beginnings and new possibilities that are just like I am glowing about.”

But she doesn’t stop there. “There's also endings to things that are maybe unexpected. But they lead to very exciting things.

“I'm directing for Sweet Magnolias and to wear both hats at the same time [as Providence Falls], it takes a lot of brain power. So, I have a lot of support right now at Sweet Magnolias.”

As Swisher is a dual actor and director for Sweet Magnolias, she really could be talking about anyone. “Endings” could refer to Maddie’s time in Serenity (although this seems too obvious), Dana Sue’s potential investor for her cooking kitchen, or Helen accepting Erik’s (Dion Johnstone) proposal.

This could even extend to the possible end of Ty (Carson Rowland) and Annie’s (Anneliese Judge) relationship if Annie accepts an offer from an out-of-state college, which is probably going to see its struggles regardless of what Swisher means with her ominous response.

Sweet Magnolias season 5 doesn’t have a release date yet, meaning the earliest we’ll likely see it on Netflix is February 2026. Maddie and Cal’s future together is likely still unwritten at this stage, but boy is it going to be worth tuning in for.

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
• The Summer I Turned Pretty season 3 episode 4 has set up Conrad’s excruciating POV in next week’s episode, and I already can’t bear it
• I watched Netflix’s Too Much and calling it Lena Dunham’s new version of Girls doesn’t do it justice

• Honkai: Star Rail Version 3.5 launches on August 13
• Hysilens joins the game to bring the DoT archetype back from the dead
• Cerydra is also coming in 3.5, as well as new story, events, and more

When I started playing Honkai: Star Rail, one of the first teams I put together focused on the DoT (damage over time) archetype. But in the last year or so, that composition dramatically fell out of favor. But now, we’re so back.

The Honkai: Star Rail Version 3.5 update – ominously-titled ‘Before Their Deaths’ – will launch on August 13, introducing new story content, two new playable characters, a new map, a new boss, and several new events.

The main highlight, at least for me, is the introduction of Hysilens. As a five-star Physical character following the path of Nihility, her kit looks set to revive the struggling DoT archetype.

Her kit includes an ability that generates a zone in combat, which reduces the attack and defense of all enemies – but, more importantly, each instance of DoT on targets in the zone triggers additional DoT from her.

Add on top that all of her abilities have a chance to apply a random DoT effect to targets and you can see why all the signs point towards her being the saviour of my favorite playstyle.

Whether that is how things turn out in practice remains to be seen, but I’m hopeful that her introduction to the game should at least go some way to pick up the fortunes of DoT and its fellow enjoyers like me.

Image 1 of 5

Hysilens has a lot of abilities to improve DoT damage and apply more DoT effects in battle. (Image credit: Hoyoverse)Image 2 of 5

Cerydra is a clear hyper-carry support for characters such as Phainon. (Image credit: Hoyoverse)Image 3 of 5

The new area of Styxia, the City of Infinite Revelry, is added in version 3.5. (Image credit: Hoyoverse)Image 4 of 5

It's like we're going back to the early flash game era with the latest restaurant management mini-game. (Image credit: Hoyoverse)Image 5 of 5

The drink mixing returns for another round of colorful concoctions. (Image credit: Hoyoverse)

Anyway, there is more than this happening in the update. The second half will add Cerydra to the game: a five-star Wind element character who follows the path of Harmony.

Developer Hoyoverse has leaned heavily into the chess theming with this strategist and support character. As well as buffing the damage of your team in various ways, her main gimmick is the ability to duplicate the skill of an ally after she builds up enough charges. Naturally, players are keen to pair her with Phainon, but there’s clearly a lot of potential for experimentation in other team compositions that rely on skills.

The re-run banners for 3.5 have also been confirmed: the first half will feature Kafka and Silver Wolf will return in the second half. Both of these make sense considering they have both recently received buffs, while Kafka is the perfect partner to Hysilens in what will likely be the premium DoT team.

Perhaps time for me to invest in an eidolon, hmm?

As well as the next part of the Trailiblaze story in a new part of Styxia, the big event for Version 3.5 is the Chrysos Maze Grand Restaurant. Yep, we’re getting a restaurant management game. Looking like a cross between Farmville and Diner Dash, you grow and serve food while renovating the restaurant to give customers the best experience possible..

On a similar theme, the Old Brews & New Friends event takes us back to the Dreamjolt Hostelry, where we’ll be mixing up and serving fancy drinks to new and familiar faces. Sounds great, gimme them Stellar Jades.

Honkai: Star Rail Version 3.5 launches on August 13. The free-to-play RPG is available now on PC, mobile, and PlayStation 5.

You might also like...

• I learned tai chi from a bear, fought a man named Chubby Brother, then caught the bubonic plague in Where Winds Meet – and now I’m downloading this free PS5 game on day one
• If these PlayStation 6 leaks are to be believed, Sony's next-gen system might not be the dramatic leap you expect
• The best free games in 2025: gaming fun at no extra cost

Spoilers for The Gilded Age season 3 episode 7 ahead.

The Gilded Age is carrying the mantle of televised period drama greatness set up by Downton Abbey over a decade ago, and I mean that as the highest compliment. Instead of soap opera sensationalism, tension and high stakes comes in the form of complicated relationships, intellectual manipulation and ever-changing social dynamics. Nobody’s getting involved in nefarious love triangles that make no sense, or having punch-ups on the street simply because there’s nothing better to do. Each step is a calculated one, and it works.

But has The Gilded Age season 3 changed all that? Across episodes 6 and 7, the action has been teetering towards the over-produced. We’ve seen the fandom come alive over Larian’s engagement, a double-crossing at the Haymarket with Oscar (Blake Ritson) left as the victim, and two death scares that have seemingly come out of nowhere.

It’s the second of these scares that now hangs a huge question mark over The Gilded Age season 3 episode 8, and it’s got the power to change how we see the HBO Max show going forward. It’s clear that each new season is trying to do something different than the last, and now that season 4 has been officially renewed, following through with the threat of death seems like the perfect way to elevate The Gilded Age all over again.

Sorry, but George has got to go for good in The Gilded Age season 3 episode 8

Yep, that’s right. For reasons currently unknown to us, George’s (Morgan Spector) life hangs in the balance in the closing scene of season 3 episode 7. He’d previously won out against Clay (Patrick Page) and Mr. Sage (Peter McRobbie), getting the business back on track in spite of their rivalry. However, this comes at a price. A courier soon visited George’s office, pulling a gun on his secretary before aiming it straight at George himself. The gun goes off, the screen goes black and we’ve got no idea whether he’s alive or died.

As far as The Gilded Age goes, this is a much more stereotypical cliffhanger than the show usually opts for. But it’s a signal that once again, it’s looking to change. Just like Carrie Coon told TV Line: “It started with how we kicked things off in the Wild West [in the season premiere]. It was signaling to the audience that this is not Season 2. We’re doing new things.”

So, what if season 3 is signaling not to expect the same again from season 4? The best way to keep things fresh is to do something completely unexpected, even if it goes against the structural fabric of what’s come before. If anything, George dying could serve as a total reset in The Gilded Age’s timeline, and who knows what could come from that. A time jump? Flashbacks? A hidden past?

If we can overlook the fact the HBO Max show has the potential to be sensationalized in season 4 (and fingers crossed it won’t be), such a brutal shift in tone could actually be really exciting. Sure, we lose George as collateral damage and poor John Adams (Claybourne Elder) was taken out in a freak carriage accident the week before. But the training wheels are off for what is and isn’t possible in this version of 1800s New York. Besides, even Spector himself is enthused by it.

“[It] just didn’t feel like the kind of thing that would happen on The Gilded Age,” he explained to TV Line, despite the historical accuracy compared to the real-life 1872 murder of Jim Fisk. “I’m thrilled the more we get to expand the world… [it’s] adding colors to our palette.”

You might also like

• HBO’s Harry Potter TV show is supposed to be a fresh start, but Nick Frost’s Hagrid says otherwise
• Superman's most surprising cameo has got me asking 3 big questions about James Gunn's next DCU project
• IT: Welcome to Derry – everything you need to know about the new Max horror series

• Researchers have identified vulnerabilities in China's Great Firewall
• The firewall attempts to block QUIC connections
• Blocking attempts leave the state exposed

Upgrades to China’s Great Firewall (GFW) have not gone as planned, and the resulting ‘critical flaw’ reduces the effectiveness of the firewall in moderating traffic loads, researchers have found. Attempts by China to censor a specific type of internet traffic in the country have left the state at risk and vulnerable to attack;

‘We [..] demonstrate that this censorship mechanism can be weaponized to block UDP traffic between arbitrary hosts in China and the rest of the world. We collaborate with various open-source communities to integrate circumvention strategies into Mozilla Firefox, the quic-go library, and all major QUIC-based circumvention tools.’

The paper was written by researchers from activist group Great Firewall Report, as well as Stanford University, University of Massachusetts Amherst, and the University of Colorado Boulder - and is titled ‘Exposing and Circumventing SNI-based QUIC Censorship of the Great Firewall of China’.

Internet censorship

The vulnerabilities stem from China’s attempts to block Quick UDP Internet Connections (QUIC) - a transport layer network protocol that is designed to replace Transmission Control Protocol (TCP) because of its built in security, flexibility, and fewer performance issues.

QUIC was invented by workers at Google back in 2012, and at least 10% of sites use the protocol - with many Google and Meta sites included. Both of these organizations are blocked by the GFW, so blocking QUIC connections seems to be an extension of this, although researchers note that not all QUIC traffic is blocked successfully.

The mechanism used to block QUIC connections is vulnerable to attacks that could block all open or root DNS resolvers outside of China from access from within the state, resulting in widespread DNS failures;

“Defending against this attack while still censoring is difficult due to the stateless nature and ease of spoofing UDP packets,” the paper explains. “Careful engineering will be needed to allow censors to apply targeted blocks in QUIC, while simultaneously preventing availability attacks.”

Via; The Register

You might also like

• Take a look at our picks for the best firewall software around
• Check out our choice for best endpoint protection software to keep you safe
• US government wants to ban Chinese technology in submarine cables