By Omair Khaliq Sultan The Beats Solo 4 is down to $119.99 at Amazon as part of the Big Spring Sale, a $80 saving off its $199.95 list price. That’s 40% off a pair of on-ear Bluetooth headphones that offer 50 hours of battery life, a lightweight build, and a tuning that’s more balanced than anything Beats has put […]

Source:: Digital Trends

By Callum Turner Interactive content now generates 52.6% higher engagement than static formats, with users spending significantly longer interacting with dynamic media and showing higher recall for brands that use it. In practical terms, that shift may have transformed expectations around how digital content should be produced, especially in commerce and B2B environments, where attention is often a […] This story continues at The Next Web

Source:: The Next Web

By Alina Maria Stan The current wave of investment in artificial intelligence reflects one of the largest capital shifts in modern technology, yet questions around financial return remain central to how this growth is being interpreted. According to a report, global venture capital investment in AI firms reached over $258 billion in 2025, accounting for 61% of all global […] This story continues at The Next Web

Source:: The Next Web

By Hisan Kidwai It’s no secret that 2025 was the golden age of budget phones. A year where you…
The post OPPO K14 5G Review: Big Battery, Decent Performance, But Is It Enough? appeared first on Fossbytes.

Source:: Fossbytes

By Deepti Pathak The story of Sung Jinwoo in Solo Leveling is characterized by his special power of creating…
The post Solo Leveling: Ranking All Sung Jinwoo Shadows by Power appeared first on Fossbytes.

Source:: Fossbytes

By Pranob Mehrotra Researchers have developed a new silicon-carbon battery design that can store up to nine times more energy while staying stable over time.

Source:: Digital Trends

Anthropic didn’t intend to introduce Mythos this way. Details of what it calls its most capable AI model yet surfaced through a data leak in its content management system (CMS), revealing a LLM with sharply improved reasoning and coding skills.

The data leak, which was the result of the company’s staffers inadvertently exposing material about the LLM, including a draft blog post about it, via a publicly accessible data repository, was first identified by independent security researchers last week.

Following disclosure of the issue, Anthropic restricted public access to the data store, only to later attribute the exposure to a configuration error in its CMS and confirm the existence of the model to Fortune, which was the first to report the leak.

Apple-focused leaker M1Astra also flagged the exposure, archiving a copy of a draft Anthropic blog post about Mythos on X before access was restricted.

In that draft, Anthropic itself struck a cautious tone, signaling concern about the model’s potential implications on cybersecurity.

“In preparing to release Claude Mythos, we want to act with extra caution and understand the risks it poses — even beyond what we learn in our own testing,” the company wrote, adding that it is particularly focused on assessing near-term cybersecurity risks.

The blog further stated that Anthropic wants to seed Mythos across enterprise security teams first and has already been testing the model’s cybersecurity prowess with a “small number of early access customers.”

The rationale seems straightforward: if today’s models can already identify and even help exploit software vulnerabilities, a more capable system like Mythos could significantly accelerate both discovery and misuse — raising the stakes for defenders and attackers alike.

Pareekh Jain, principal analyst at Pareekh Consulting, says Mythos could cut both ways for CISOs and enterprise security teams, compressing the gap between cyber offense and defense.

While at one end, models like Mythos could transform security by automating vulnerability discovery, continuous red-teaming, faster triage, and large-scale threat hunting areas, on the other hand, it could make cyberattacks easier by letting AI agents act autonomously with high skill, Jain said.

That risk for CISOs is not theoretical, Jain added, as earlier-generation models were quickly repurposed into tools for developing malware.

The risk is even higher with Mythos because of its capabilities like “recursive self-fixing,” Vladimir Belomestnov, senior technical specialist at HCLTech, wrote in a post on LinkedIn.

“The leaked files highlight a capability for the AI to autonomously identify and patch vulnerabilities in its own code. Even if this is currently limited to assisted exploitation, it suggests a narrowing gap between human and machine software engineering,” Belomestnov wrote.

However, Anthropic appears to be some distance from a full release of the model.

“Mythos is also a large, compute-intensive model. It’s very expensive for us to serve, and will be very expensive for our customers to use. We’re working to make the model much more efficient before any general release,” the copy of the draft blog post reads.

What is clear, however, is that the company is already planning a phased rollout targeting cybersecurity use cases.

“We’ll be slowly expanding access to Claude Mythos to more customers using the Claude API over the coming weeks. Since we’re particularly interested in cybersecurity uses, that’s where we aim to expand the EAP initially,” the company wrote in the draft blog post.

There is another copy of the blog post, which also names the model as Capybara. Anthropic hasn’t made it clear what the final name of the model will be.

The indecision over the model’s name, though, didn’t stop it from rattling markets last week. Shares of cybersecurity vendors, including CrowdStrike, Palo Alto Networks, Zscaler, and Fortinet, fell as investors assessed what more capable models within Claude Code Security could mean for the competitive landscape.

However, Avasant’s research director, Gaurav Dewan, was more optimistic about Mythos’ impact on vendors: “Powerful models will not replace cybersecurity platforms”.

Rather, Dewan sees vendors increasingly embedding frontier models from Anthropic and OpenAI and others into their stacks for vulnerability discovery, code and cloud posture management, and threat investigation and response automation.

“One can expect partnerships and controlled integrations, not disintermediation. Vendors that already own telemetry, workflows, and enforcement will benefit most,” Dewan added.

The article originally appeared in CSO.

Source:: Computer World

Microsoft has reportedly had a hard time convincing corporate customers to pay premium prices for its enterprise AI products, particularly as their employees prefer ChatGPT or Gemini. The Information reported in December that Microsoft quietly slashed sales targets for some of those products — a report that Microsoft denied. Nevertheless, by Microsoft’s own admission, just 3.3% of its vast Microsoft 365 user base has a paid Microsoft 365 Copilot license.

Customers’ reluctance to pay top dollar for Microsoft’s AI tools likely reflects the pace at which new AI alternatives have raced to market. With so many services to choose between and the industry at such a febrile and inventive stage, customers want to dance between the options as they seek those they prefer, rather than being in thrall to one provider — while recent Microsoft 365 price increases show the extent to which sector dominance can leave customers exposed to price hikes. Now bitten, customers are shy to commit too much to one provider.

Spoiled for choice

Why would they want to, given that there are so many alternatives to choose from? AI services are like streaming services, except you don’t need to subscribe to them all: they pretty much all offer the same thing, though some are better for some tasks.

Apple understands this. By its actions, it is showing us that that AI models are destined to become commodities, which is why the company is resolutely focused on making sure its systems become the best platforms on which to run the models.

This recognition means Apple Intelligence is likely to only ever become a selection of hand-picked on-device assistants most of us will use some of the time, with additional tasks supported by third-party providers. We know that Apple plans to use Google Gemini to help it fast-track development of additional Apple Intelligence services, but we also know it intends to support multiple AI services. 

Apple’s App Store for AI

I think the best way to look at this is as an App Store for AI. You’ll be able to do a certain amount using on-device AI and Siri, and you’ll be able to choose between third-party AI services to handle other tasks. That plan means the existing exclusive arrangement with OpenAI’s ChatGPT will be abandoned as the company opens Siri up so its customers can choose which third-party AI services to use. It isn’t clear yet if this will extend to use of on-prem AI systems, which will be a particularly attractive proposition to regulated industries and privacy advocates using Mac minis to run independent LLMs. 

An App Store model also gives Apple a chance to offer up APIs to AI developers to enable sophisticated AI applications that do not devour personal privacy. That seems a very Apple-like approach to these things.

What is clear is the extent to which Apple the hardware company now understands that AI doesn’t replace platforms, but depends on them. Apple as a hardware and operating systems provider just needs to focus on providing the best available ecosystem on which to build and run AI systems, with a user experience to match.

Pop goes the weasel

Asymco’s Horace Dediu notes the significance of such a shift: “If foundation models are heading toward commodity status, then the strategic value shifts to whoever controls the integration layer and the user relationship,” he wrote. 

Apple’s 2-billion-plus ecosystem gives it the edge in distribution, while its tried and tested App Store approach helps validate and optimize the user relationship. The idea that AI services become apps to be bought and sold on Apple’s platforms isn’t far-fetched. Bloomberg suggests that Apple is building tools to let chatbot apps installed via the App Store work with Siri and other Apple Intelligence features. 

Reflecting that some AI services are better at some tasks than they are at others, Apple will also make it possible for customers to choose which AI service handles each request. Apple will probably take a slice from any AI services subscription sales made via its platform as part of this plan, just as it makes bank from every other fee-based app.

I do wonder if this could end up with a weird Catch 22-like situation for Microsoft Office users, in which everything they do on their Apple device is handled by their chosen AI service, except when using an Office app when they may find themselves trapped with Copilot. 

More to come at WWDC?

Summing up, it looks very much as if after the filth and the fury of the first stage of AI evolution, the song remains the same — you still need solid platforms to run this stuff on. Which is, of course, where Apple’s powerful Apple silicon-powered devices have so much to bring.

The company is expected to introduce this with its 27-series of OS updates, the first glance of which we will gain at WWDC in early June. Once we see what kind of system Apple is putting together, we’ll have a much better understanding of what the future of the AI industry is going to be. From where I sit, it seems obvious: while AI may change the world, it’s likely to do so while running on an Apple product.

You can follow me on social media! Join me on BlueSky,  LinkedIn, and Mastodon.

Source:: Computer World

By Varun Mirchandani AI-powered smart glasses are raising new concerns as users reportedly exploit them for exam cheating, alongside ongoing privacy fears.

Source:: Digital Trends

By Varun Mirchandani New research suggests AI can match average human creativity, but true originality and top-level creativity remain uniquely human.

Source:: Digital Trends

By Hisan Kidwai We’ve just witnessed a stellar BGIS Grand Finals tournament, filled with ups and downs for many…
The post BGIS Grand Finals Day 3 Highlights: Soul Crowned the Champions appeared first on Fossbytes.

Source:: Fossbytes

By Hisan Kidwai BGIS Grand Finals just wrapped up, and we’ve had a nail-biting competition that went down to…
The post BGIS Grand Finals 2026 Standings After Day 3: Soul Wins the Championship appeared first on Fossbytes.

Source:: Fossbytes

By Ana-Maria Stanciuc A week that spanned semiconductor physics, orbital logistics, defence interceptors, and carob-based chocolate tells you something about the breadth of European and European adjacent capital right now. The dominant thread is not a single sector but a single instinct: back the infrastructure layer, whether that means chip interconnects, satellite transfer vehicles, or the AI plumbing […] This story continues at The Next Web

Source:: The Next Web

By Shikhar Mehrotra With FCC filings now surfacing for products codenamed Scriber and Blazer, Meta’s push into the prescription eyewear space looks less like a rumour and more like an imminent launch.

Source:: Digital Trends

By Varun Mirchandani New research reveals generative AI is making fraud faster and more scalable, turning cybercrime into a $400 billion global problem.

Source:: Digital Trends

By Alina Maria Stan Every co-founder Elon Musk recruited to build xAI has now reportedly left the company. Manuel Kroiss, who led the pretraining team, told people this month that he was departing. Ross Nordeen, described by Business Insider as Musk’s “right-hand operator,” left on Friday. They were the last two of eleven co-founders, all of whom have exited […] This story continues at The Next Web

Source:: The Next Web

By Allison Steffens Herrera Meta is preparing to launch two new Ray-Ban smart glasses models designed specifically for prescription wearers, according to a Bloomberg report published on Thursday. The models, codenamed Scriber and Blazer, were first spotted in Federal Communications Commission filings and are expected to reach consumers as early as next week. They do not represent a new […] This story continues at The Next Web

Source:: The Next Web

By Hisan Kidwai Day 2 of the BGIS Grand Finals was chaotic in the best way possible. A variety…
The post BGIS Grand Finals Day 2 Highlights: Soul Finish Strong, Genesis Dominate appeared first on Fossbytes.

Source:: Fossbytes

By Hisan Kidwai We said day two of the BGIS Grand Finals would be the day of comebacks. While…
The post BGIS Grand Finals 2026 Standings After Day 2 appeared first on Fossbytes.

Source:: Fossbytes

Microsoft is finally blocking a long-since retired program that it said led to “abuse and credential theft,” yet remained widely trusted for years.

Beginning in April, Redmond will remove trust for kernel drivers that haven’t been vetted through its Windows Hardware Compatibility Program (WHCP). The company is specifically targeting kernel drivers signed by the now defunct cross-signed root program. 

But while this closes a security hole, Microsoft acknowledges that it could impact some legacy applications and use cases. To balance security with compatibility, the company will initially roll out the policy in “evaluation mode” with its April 2026 Windows 11 and Server update. It will also provide some leeway for older, widely-used and reputable drivers, and admins in certain situations will have the ability to override the new policy altogether.

“Essentially, Microsoft is closing a 20-year-old critical security hole in its OS,” said David Shipley of Beauceron Security. “Device drivers get to touch the OS kernel and can abuse that supreme-level access to do fun things like disable anti-virus and endpoint monitoring tools.”

Program put platforms at risk

Microsoft introduced the cross-signed root program in the early 2000s to give driver providers Windows-trusted code signing certificates with, it said, “varying degrees of partner vetting.” But it offered zero assurances around the security and compatibility of that kernel code.

The program was administered by third parties who stored private keys associated with the certificates, which, Microsoft says, “led to abuse and credential theft that put our customers and their platforms at risk.”

As a result, the program was deprecated in 2021, and all certificates have since expired. However, third-party drivers signed by the program are still “broadly trusted,” Microsoft says.

The new kernel trust policy will apply to Windows 11 24H2, 25H2, 26H1, and Windows Server 2025, and all future versions will enforce it, because, Peter Waxman, a group program manager at Microsoft, writes in a blog post, “drivers are a critical part of the Windows ecosystem, and ensuring their integrity is essential to providing a secure and trustworthy environment.”

However, in its initial evaluation mode, Microsoft will monitor and audit driver loads to test for compatibility issues should cross-signed drivers be blocked. Systems will remain in evaluation mode until they meet specific runtime (100 hours) and boot-start (2-3 restarts) scenarios. If all drivers loaded during the evaluation period are trusted, the policy activates, but if any cross-signed drivers are audited that would not pass the new kernel trust policy, the system remains in evaluation mode until those drivers are no longer audited.

Once activated, the kernel trust policy will automatically block untrusted drivers. Only drivers that have been passed and signed by WHCP, which scans all driver submissions for malware and tests whether the device is compatible with Windows hardware and operating system, will be loaded by default. Those drivers are then trusted across the Windows ecosystem.

Microsoft calls the certification program a “rigorous driver signing process” that helps ensure its partners are “continuously vetted” and comply with its latest security and compliance requirements.

Workarounds in certain scenarios

Even as it enforces this new policy, Microsoft is taking steps to minimize disruption. The tech giant will maintain an explicit allow list so that the kernel can load old, but “widely-used and reputable drivers” previously vetted through the retired program. This exception list is based on two years of real-world data around enterprise use of older drivers.

Further, admins will have the ability to override the policy via Application Control for Business. This can be particularly useful in scenarios where enterprises are loading custom drivers built for internal use.

“This gives enterprises the ability to run privately-signed drivers on enrolled systems without degrading security,” Microsoft notes. However, these bypass policies must be granted specific authority in the device’s cryptographic key to ensure it is applicable only to a specific enterprise environment. 

It’s a win for security — but there are tradeoffs

On the whole, this is great news for security, analysts note.

“There seems to be no single breach, instance, law, or sudden trigger that is prompting this step,” said Thomas Randall, a research director at Info-Tech Research Group. “Instead, this is part of a long-running cleanup for avoidable security risks.”

But, noted Beauceron’s Shipley, with great security comes some “usability or convenience trade-offs.” For instance, disabling device drivers, even those whose certificates have expired, might render highly-customized internet of things (IoT) devices obsolete. This could also impact critical medical equipment like x-ray machines.

Microsoft is being smart by baking in the two safety valves during evaluation (the 100 driver loads and three restarts) to make sure “they don’t brick something vital,” said Shipley.

Does this add more friction to malware creators? Yes. Is it a silver bullet? No, he said. “Endpoint detection and response-killing malware is still going to happen, but the blast radius just got smaller, so that’s a win.”

What enterprises should keep in mind

While Microsoft does extend some exceptions to its new policy, enterprises taking advantage of them shouldn’t consider themselves immune to ultimate driver blockage.

“Even those permissible older drivers will be on borrowed time,” Randall said. “We can expect Microsoft to eventually phase those out, too.”

The main risk for organizations is that they may be using older drivers without realizing it, he noted. For instance, factories or medical facilities relying on older hardware, specialized equipment, long-running business systems, or internal tools that were built years ago but never fully updated are all at risk.

“The device may still work perfectly well, but it may rely on an older driver, and nobody had a strong reason to replace it until now,” said Randall. “If one of those older drivers is not on Microsoft’s allowed list and no replacement exists, a device or application may stop working properly once a machine begins enforcing the new rule.”

To prepare for the policy change, Randall advised enterprises to make a list of hardware and software that install Windows drivers, then test machines that are in heavy, day-to-day use, or that serve highly important functions (like emergency medical equipment); ask hardware and software providers whether they have a roadmap for updated drivers; and identify any internal-only drivers that may require company-managed exceptions.

Further, “organizations should not assume Microsoft’s exception list will save them, because Microsoft says that list will be limited,” Randall noted.

Source:: Computer World