As soon as I heard that Gemini Spark was now part of the Chrome browser I had to try it for myself.

Gemini Spark is an AI agent designed to take on longer-running tasks that continue in the background. It’s much smoother now thanks to an update that incorporates the platform directly into Google‘s Chrome browser. So now the AI can actually browse the web alongside you, doing its own tasks, instead of simply talking about it, even if you close your browser.

Laptop Intelligence: Google finally announced the much-talked-about project to blend Android and ChromeOS into a new, AI-driven “computing” experience earlier this year. Now, the first OEM Googlebook design has leaked ahead of the machine’s expected – and still unconfirmed – release date.

Google unveiled the Googlebook laptop in May, explaining that several OEMs would be joining in with their own take on the AI-ready machine. Thanks to Digital Citizen, we now have a closer look at the Googlebook Lenovo is building for a late-2026 release. The Lenovo Googlebook 15 carries some design choices expected to become standard across the new platform, while other hardware decisions set minimalism aside in favor of better connectivity.

The leak shows what seems to be a full press image set for Lenovo’s Googlebook 15. The PC giant has apparently built one of the very first Googlebook laptops to reach the market, a notable milestone that could help shape the early commercial fortunes of Google’s latest hardware push.

The Lenovo Googlebook 15 features some of the platform’s most distinctive design choices, starting with a Google “G” key at the bottom left of the keyboard. Much like Microsoft’s much-maligned Copilot key, the new G key is expected to serve as a shortcut to launch the Gemini chatbot.

The keyboard also includes 14 Function keys, with a dedicated accessibility key on the upper right that’s rarely seen in laptop designs. According to the leaked press images, the Lenovo Googlebook 15 comes packed with connectivity on both sides. Options include a full-sized HDMI port, a USB-C port, a 3.5mm headphone jack, and what appears to be a full-sized USB-A port alongside a second USB-C port.

The leaked images provide a rendered mockup of what should be Googlebook’s UI, with a rather standard desktop offering and a status bar taken straight from a modern Android iteration. Googlebooks use Aluminium OS, a new operating system Google has been working on for a while to bridge Android app compatibility with ChromeOS’ always-on environment.

The one thing the images don’t reveal is any hint of actual hardware specifications. Google has called the Googlebook its most advanced laptop yet, built to get the most out of the Gemini chatbot, with the company framing intelligence as the new spec that matters most. As PCWorld has noted, the heavy reliance on automation and agentic AI could make Googlebooks the first wave of “anti-personal computers,” devices designed to act on a user’s intent, leaving them no longer solely in control of the machine.

The state already hosts more than 500 data centers.

On Monday, Apple briefly removed the popular messaging app Telegram from the App Store for iOS devices after a report that a user was sharing illegal content. According to both Apple and Telegram CEO Pavel Durov, the app — which has more than 1 billion users — was restored on the App Store within hours. But Durov accused Apple of overreacting to a ransom attempt in a group chat.

In a Telegram post about the incident, Durov said a member of a public group chat edited a previously posted message and added what Durov called “AI-modified illegal content” to it in an attempt to get the group chat flagged, because its admins had presumably refused to pay a ransom. Instead, Apple removed the app from the Apple App Store entirely until the user was banned. Apple confirmed in an email to CNET that Telegram remained available in the separate Mac App Store and that users who already had Telegram downloaded were not affected.

“We briefly removed Telegram from the App Store after our review found content that violates our strict guidelines prohibiting child sexual abuse material,” Apple said in a statement. “The app was subsequently restored after the developer promptly removed the content and banned the user who posted it.”

Apple did not immediately respond to questions about why the Mac version of Telegram remained available for download while the mobile version was removed.

A representative for Telegram declined to comment directly, instead pointing to Durov’s post.

Telegram warns other app developers

Durov said that the extortionist manipulated Apple into an “overreaction” that led to the app’s removal, which he said occurred without warning. That could happen, he said, to other apps that include user content.

“If an app used by more than a billion people can be removed from the App Store without prior warning, any app can be,” Durov wrote.

Telegram’s official X account more pointedly criticized Apple’s actions in a series of posts and replies. In one, Telegram posted, “reports of my demise are greatly exaggerated,” followed by an apple emoji. In another that included a link to Durov’s post, Telegram wrote, “I’m sure this stance will be applied equally to all other apps in the store, in the future, right? @Apple.”

AI + ML

The House of Zen’s new Helios racks, Venice Epycs, may dent Nvidia’s dominance — if the bubble doesn’t pop first

AMD has posted strong second quarter results and forecast even better future financials once its Helios rack systems and Instinct MI400-series GPUs reach buyers.

“In data center AI, the growing number and scale of Helios and MI450-series deployments position the [datacenter] business for significant growth in the second half of the year, with growth accelerating in 2027,” CEO Lisa Su told investors on Tuesday during the chip design company’s Q2 earnings call. “We now expect data center segment revenue to more than double year over year in 2027,” she added.

Yet, despite reporting Q2 profits surging 163 percent year-over-year on revenues of $11.5 billion, and several multi-gigawatts worth of Helios commitments from the likes of OpenAI, Anthropic, and Meta in the bag, Wall Street isn’t buying it.

The company’s share plunged 10.5 percent after its results announcement, before settling 8.7 percent below opening price at the time of publication.

The apparent cause for concern: AMD’s growing exposure to the AI bubble. Much of the company’s growth potential across both CPUs and GPUs is tied to AI adoption by a handful of companies that are yet to prove they can operate profitably.

On Tuesday’s earnings call, Su attempted to assuage investor fears, but in the same breath she said the quiet part out loud. 

“When we talked about the large frontier-model companies, OpenAI, Anthropic, Meta, they will be consuming through a number of CSPs,” Su said. “There are additional customers or lots of customers who are interested in Helios at, let’s call it, a more regular scale than gigawatt scale.”

In other words, while AMD can sell plenty of GPUs, most are sold to a handful of customers. And while other entities have AMD on their shopping lists, they don’t buy in bulk.

Microsoft, another flagship customer for AMD’s latest generation of AI picks and shoves, serves both OpenAI and Anthropic, while Meta is reportedly looking to enter the GPU cloud biz itself.

Despite this, AMD remains optimistic about its prospects over the next few quarters.

It’s not hard to see why because since the launch of its MI300-series GPUs in late 2023, AMD has established itself as the most credible alternative to Nvidia.

At its Advancing AI event in San Francisco last month, the company showcased a new rack-scale compute platform called Helios packing 72 Instinct MI455X GPUs each with 432 GB of HBM4 memory on board.

On paper, the system meets or beats the performance of Nvidia’s Vera Rubin platform on most metrics. The Reg explored these chips in greater detail here.

AMD is also eager to cash in on demand for CPUs to power agentic AI sandboxes, which Su anticipates will be the biggest growth driver for Epyc sales before long.

Also unveiled at Advancing AI, AMD’s Venice Epycs will offer up to 256 cores and 512 threads with support for 16 memory channels at speeds of up to 1.6 TB/s per socket.

However, the CPU market isn’t the duopoly it once was and AMD now faces its traditional foe Intel, plus new competitors such as Nvidia, Arm, Qualcomm, AWS, Google, and Microsoft.

Nonetheless Su remains bullish about AMD’s prospects.

“Whether it’s server CPU, datacenter, AI, or our embedded business, and our PC business, we see them all benefiting from the AI tailwinds,” Su said.

It seems the AI tides will lift all chips, but let’s not forget that tides rise and fall. 

In any case, while AMD’s embedded division is doing quite well with revenues topping $977 million, an increase of 19 percent over this time last year, things aren’t looking so hot for the House of Zen’s gaming and client computing divisions. While client PC sales were up 23 percent YoY to $3.1 billion during the quarter, AMD warns that the ongoing RAMpocalypse is likely to cut into PC sales over the next few quarters.

Meanwhile in gaming, the end of a console sales cycle is hitting AMD hard on semi-custom processor orders, with gaming revenue down 31 percent during the quarter to $779 million

AMD’s datacenter and AI sales teams therefore get the job of delivering the company’s Q3 forecast which calls for revenues of $13 billion plus or minus $300 million. ®

If you’ve been dying to run macOS on iPad hardware natively, a jailbreak solution has emerged for M1 and M2 iPads running iPadOS 16, though it is extremely experimental.

There has been an endless debate surrounding the iPad since Apple debuted the first iPad Pro. It runs iPadOS, a branch of iOS, which is much more locked down compared to macOS.

Users that want to use macOS on an iPad can finally give it a try, but it’ll be a limited experience. The GitHub repository for the Virtual Mac on iPad software will provide you all of the information you’ll need.

Since it takes advantage of an exploit found in iPadOS 16 to iPadOS 16.3.1, you’re going to need a very specific device and setup. Even then, the macOS and Xcode versions available are limited.

An iPad Pro with 1TB or more storage is recommended, though this works on iPad Air or other iPad Pros with M1 or M2 chips. Users can install anything from macOS Monterey to macOS Golden Gate, but it is best to stick to macOS Ventura to macOS Sequoia for stability.

If you’re wanting Xcode, install the correct version for the given OS you’ve installed.

While a Magic Keyboard isn’t required, it is recommended. macOS isn’t designed for touch, and you’ll be using some odd workarounds with the virtual keyboard otherwise.

The GitHub is open and asking for contributions to the active projects. Goals include adding support for iPadOS 15 and finding ways to support the feature in later iPadOS versions.

Not for the novice tinkerer

Since this is a jailbreak and an unofficial install of macOS, you won’t be able to sign into your Apple ID. Expect other aspects to be odd or broken as well.

The iPad has always been a computer whether anyone was confused about that or not

Jailbreaking can lead to problems, and the Virtual Mac on iPad software is experimental at best, so do this at your own risk. Or, if you’re really that curious, perhaps watch someone on YouTube do it for you.

I’d highly recommend just running Sidecar for a similar effect. Sure, you wouldn’t actually be running macOS on iPad, but you’d also not be ruining a perfectly functional iPad either.

I very seriously doubt we’ll ever see macOS on an iPad, not even via an official hybrid system introduced by Apple. The company has been very clear about keeping the hardware categories separate.

Of course, Apple has changed its mind before. This software is a fun proof of concept for something we already knew was possible — macOS can run on the iPad.

Alternatively, you could just wait for the rumored touchscreen MacBook Pro. Whether or not we actually want or need macOS on iPad remains up for debate.

OpenAI and Anthropic have confirmed that their AI models were involved in separate, newly disclosed third-party cybersecurity testing incidents that resulted in a real website being breached and social engineering attacks against people outside the intended testing boundaries.

These incidents are unrelated to the previously disclosed Hugging Face breach, in which OpenAI models hacked the AI platform and used exposed credentials to breach accounts at four other third-party services during another cybersecurity evaluation.

OpenAI disclosed the two new incidents on Tuesday, saying they occurred during evaluations conducted by the UK AI Security Institute and cybersecurity testing company Irregular.

Spear-phishing attacks on GitHub project maintainers

The UK AI Security Institute, commonly known as AISI, is a government research organization that evaluates the capabilities and risks of advanced AI models.

During a recent cyber-range evaluation, AISI says agents powered by Anthropic’s Claude Mythos 5 and OpenAI’s GPT-5.6 Sol took unsanctioned actions on the public internet while trying to complete simulated hacking challenges.

Across 122 evaluation attempts, AISI identified 19 unsanctioned actions on the live internet in 10 runs. Seventeen involved Mythos 5 and two involved GPT-5.6 Sol.

AISI says the attempts were unsuccessful and that it found no resulting real-world harm.

“These attempts were unsuccessful, and our investigations have not evidenced any resulting real-world harm,” AISI said in a separate advisory.

“But this is the first time we have seen risks around autonomy and deception manifest this clearly, without specific prompting, in the real-world.  “

AISI intentionally enabled open internet access and disabled the model providers’ cyber classifiers to measure the models’ underlying capabilities. However, the agents were only authorized to attack the simulated cyber range and were not explicitly told how they could use their internet access or instructed to avoid interacting with real people and systems.

Anthropic confirmed to BleepingComputer that AISI was testing a version of Claude Mythos 5 but said it is still investigating and cannot yet confirm all of the technical details described in AISI’s report.

The company said it was notified on Monday and is working with AISI to obtain the evaluation transcripts needed to conduct its own review.

“We’re grateful to the UK AISI for their leadership on this incident, which underscores the need for a broader conversation about how to safely evaluate increasingly capable AI agents,” an Anthropic spokesperson told BleepingComputer.

“As we shared after disclosing our own incident last week, the field needs stronger, shared standards for how evaluation environments are built and secured. We look forward to partnering with the UK AISI to learn more about this incident as we conduct our own investigation.”

Anthropic noted that AISI tested Mythos 5 without its standard cyber safeguards enabled and that this is not the same configuration available to customers. AISI says its evaluation design and configurations may have contributed to the behavior, but did not anticipate how the agent would show “signs of novel, potentially deceptive behaviours”.

AISI says a Mythos 5 agent searched the internet for terms related to the cyber challenge and mistakenly concluded that an unrelated public GitHub repository was connected to the test.

The agent then attempted a supply-chain attack by submitting malicious code to the real open-source project, believing that compromising the software could provide a path into a machine within the simulated range.

Most concerning, the agent researched the project’s maintainers, created multiple fake GitHub identities, and used those accounts in social engineering attacks to push the maintainer into approving a malicious pull request.

When a human reviewer warned that the pull request contained malware, the agent denied the accusation and used other fake accounts it controlled to pressure the maintainer and create the appearance that independent users had reviewed and approved the changes.

The agent continued its social engineering attacks by hiding its identity using Tor and proxy services and creating disposable GitHub accounts.

It sent five targeted emails to the developers, with some containing malware and others attempting to persuade them to approve the code change.

The agent also posted bug reports containing malicious code and hidden prompt-injection instructions designed to trick AI coding assistants into taking unintended actions.

When a real person questioned one of the posts, the agent edited it to conceal the malicious content. In another attempt, it signed off in Danish to appear more convincing to a Danish-speaking maintainer.

AISI researchers found that agents coordinated with each other across different evaluation runs, using a shared GitHub repository as a message board to leave instructions for one another.

“This is the first time AISI has seen deception of this severity that was targeted at a real person, unprompted, in the real world,” the institute said.

AISI says it remains unclear when, or whether, the model understood that it was interacting with real people rather than simulated targets.

OpenAI model hacked real website during simulated CTF

The second newly disclosed incident occurred during Capture-the-Flag evaluations conducted by Irregular that were supposed to be isolated from the internet.

In one test, the fictional target’s name happened to match a real domain, causing the model to exploit a real website that it believed was part of the simulated challenge.

While Irregular’s testing environment was supposed to be isolated from the internet, a misconfiguration allowed OpenAI models to access the public internet and target the real website.

“Based on Irregular’s investigation, the model also found and used credentials to operate that same site,” OpenAI said.

OpenAI says the model exploited a basic vulnerability rather than using a zero-day or escaping its testing environment.

OpenAI says Irregular has not discovered any impact beyond the affected site’s own data, but its investigation remains ongoing. OpenAI says the company is preparing a white paper on containment and securely conducting cyber evaluations.

Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.

The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.

Get the whitepaper

Presented by Rezolve Ai

Most brands know something is shifting in how consumers find and choose products. What most don’t know is how much of that shift has already taken place, where it’s happening, or whether they’re on the right side of it. That uncertainty is the problem. And the analytics stack most brands rely on isn’t built to resolve it.

The decision layer has moved

In 2014, 82% of digital commerce started on a brand’s website. By 2024 that had fallen to 38%, according to Salesforce research. The journey that used to begin at a brand’s front door now begins somewhere else. Increasingly, it begins with a question asked of an AI platform and ends with an answer that shapes the purchase decision before any brand-owned touchpoint is engaged.

Consumers are asking AI where to shop, what to buy, and which product is right for them. Bain research shows that four in five consumers rely on zero-click results at least 40% of the time. That means the shortlist a consumer receives from an AI answer engine is, in many cases, the only shortlist they consult. Adobe Analytics recorded over 800% year-over-year growth in AI-driven traffic to retail sites, a signal of how rapidly AI platforms are inserting themselves between brands and their customers.

This is a structural shift, not a trend. And it has created a category of commercial loss that most analytics tools are architecturally incapable of detecting.

What you can’t see is costing you

The gap is this: a brand can have strong onsite conversion metrics and still be losing significant ground in the market, because the customers who never arrived aren’t captured in any dashboard. There’s no “AI excluded you” event in a session log. There’s no abandoned cart entry for a shopper who was told by an AI assistant that a competitor was the better fit.

This is different from the SEO problem brands have managed for two decades. With traditional search, absence had a visible signal. You could see your ranking, audit the gap, and act on it. With AI answer engines, absence is invisible by default. The surface doesn’t show you what it didn’t show the consumer.

Sixty percent of searches now end without a click, according to Semrush’s 2025 zero-click study. For AI-mediated discovery, that number is structurally higher. The answer is the destination. If a brand isn’t in the answer, it isn’t in the consideration set, and its analytics will never surface that fact.

The metric that isn’t being measured

The commerce industry has developed sophisticated instrumentation for the journey from landing page to purchase. It has essentially no instrumentation for the journey from consumer intent to brand discovery, the layer where AI is now operating.

Brands that want to understand their actual competitive position in an AI-mediated market need to ask a different set of questions: How does my brand appear when consumers ask AI for recommendations in my category? What language does AI use to describe my products? Where am I present, where am I absent, and where am I being described in ways that don’t reflect my positioning?

These aren’t marketing questions. They’re infrastructure questions. And answering them requires a different kind of audit than anything in the current commerce or marketing toolkit.

Rezolve Ai commissioned research across 1,500 US consumers in January 2025 that found the majority of shoppers who use AI for product research make purchase decisions directly from those AI-generated recommendations, without returning to a search engine or brand site to verify. The implication for brands is significant: by the time a consumer reaches a brand’s owned properties, the decision may already have been made, or unmade, somewhere else.

What comes next

The brands that will maintain commercial relevance as AI mediates more of the discovery layer are those that develop visibility into it, not just presence on their own platforms. That means treating AI discoverability as a measurable discipline, not an assumption, and building the infrastructure to understand, track, and influence how AI systems represent them to consumers.

The tools to do that are emerging. The measurement frameworks are not yet standardized. But the brands that begin building that visibility now will have a structural advantage as the market continues to shift.

AI answer engines are already forming preferences. Every day without visibility is a day those preferences solidify without you.

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.

Big anniversary, big screens? Apple might bump up the display size of the iPhone Pro and iPhone Pro Max next year as the company celebrates the 20th anniversary of the iPhone, according to an insider.

Digital Chat Station, an anonymous but well-regarded source of tech leaks, shared on Chinese social media platform Weibo that the iPhone 20 Pro could feature a 6.4-inch display, while the iPhone 20 Pro Max could come with a 7-inch screen. Although the account holder’s identity has never been publicly confirmed, Digital Chat Station has built a reputation for providing reliable information about upcoming devices.

Apple is expected to release those devices in September 2027, three months after the company celebrates the 20th anniversary of the iPhone’s launch on June 29, 2007. (For a blast from the past, check out CNET’s review of that legendary device.) Apple has not announced any official launch dates.

A representative for Apple didn’t immediately respond to a request for comment.

The latest versions of the iPhone Pro and Pro Max are a tad smaller — 6.3 inches for the iPhone Pro and 6.9 inches for the iPhone Pro Max.

The iPhone 18 Pro and iPhone 18 Pro Max, expected to launch next month, will also have 6.3- and 6.9-inch screen sizes. According to our roundup of reports and rumors about Apple’s upcoming products, the company’s next lineup could also include its first foldable iPhone, the Apple Watch Series 12, and the Apple Watch Ultra 4

Glass all around?

In addition to the larger screen sizes for the iPhone 20 Pro and Pro Max, the Digital Chat Station report also said the devices might have a quad-curved design and no bezel showing. If true, when you look at the front of those phones, you’ll only see glass that wraps around all four edges. You won’t see a bezel, the border that surrounds the screen, which is typically found on most smartphones.

Digital Chat Station said that, even with the bigger screens, the phones will maintain the same aspect ratio.

The last time Apple increased screen size was two years ago. The iPhone 16 Pro screen size increased to 6.3 inches (up from 6.1), and the Pro Max to 6.9 inches (up from 6.7).

Increasing screen sizes “makes sense” as people consume more content on their iPhones, says Austin Evans, who has been testing tech devices for years on his YouTube channel, which has nearly 6 million subscribers. He also likes the idea of the wraparound glass look on the front display.

“I’m excited to see what they put together for the 20th anniversary,” Evans said. “If they’re able to ship an iPhone with effectively zero bezels at all, it could really set the tone for what an iPhone can be for the next decade.”

What else have we heard?

According to the rumor mill, there could be several other upgrades and changes for the iPhone 20 lineup.

There might not be physical buttons for powering on, adjusting volume or controlling the camera. Instead, solid-state buttons might be integrated into the frame itself, enabling consumers to initiate actions by pressing lightly or firmly.

There might be a Face ID sensor under the main screen, according to some reports. The Dynamic Island might shrink, with the front camera reduced to pinhole size, although these changes might arrive sooner with the iPhone 18.

The iPhone 20 might also have a brighter, thinner OLED panel to reduce glare, increase brightness and use less power.

Internally, the iPhone 20 might be fitted with a second-generation 2nm chip for better efficiency and performance, as well as a pure-silicon anode battery.

Ever since being admittedly fascinated by the Cambridge coffee webcam from the 1990s, I’ve written about VPNs, the NFL, smartphones, living wages, over/unders and everything in between.

See full bio