Agregátor RSS

OpenClaw, the AI digital assistant, is looking to establish itself as an essential tool for users — but if it is as profligate in its use of tokens as its creator, Pete Steinberger, then users could be in for some huge usage bills.

Steinberger, now working for OpenAI, spent $10 million on AI tokens there last month, he told Cloudflare CEO Dane Knecht at the Cloudflare Connect 2026 Conference in San Francisco.

The conversation was reported by The Register, which noted Knecht’s remark that Steinberger had “a virtually unlimited token budget, which not all of us in the room have,” and Steinberger’s response, “What people don’t always understand is not all of that is useful.”

Steinberger went on to question the adequacy of companies’ infrastructure, musing, “Why is this still using my computer? Shouldn’t this all be in the cloud?”

Adding cloud costs into the equation, as well as shoving privacy issues to one side would not reassure the audience.

This was scarcely a conversation to reassure those CFOs who were getting palpitations about the real cost of AI tokens, whether this was the best way to introduce AI into the system and whether there are better ways of managing costs. However, such dizzying heights of spending may be music to the ears of Steinberger’s employer as it prepares for its upcoming IPO.

This article first appeared on InfoWorld.

Raven nabízí alternativu k tradičnímu přístupu založenému na Wine. • Cílem je lepší kompatibilita, vyšší výkon a méně ručního nastavování. • Za projektem stojí Mirko Brombin, autor Bottles a spolutvůrce Vanilla OS.

The PHP security team has shared mitigation measures for a new remote code execution vulnerability that affects systems with non-default configurations and has no patch. [...]

A judge in New Mexico has fined Meta $25 billion for its flouting of data privacy laws between 2010 and 2021, including its controversial passing of data to Cambridge Analytica.

The action serves as a stark reminder to enterprises not to play fast and loose with data protection legislation, as the consequences could be highly damaging.

Meta faced the court action after it was revealed that data company Cambridge Analytica had improperly obtained personal data that Facebook was holding.

It could have been even worse for Meta. Under New Mexico law, the damages could have been as high as $219 billion, which would have exceeded its revenue for 2025.

Judge Francis Mathew reduced his demand to $25 billion without explanation, but was scathing in his judgment on the social media company. “The harm in this matter is difficult to overstate, and goes to the heart of Facebook’s basic business proposition,” he wrote in his order.

The company had committed more than 43 million violations of New Mexico’s consumer protection laws and had deceived its users by giving third-party app developers access to personal information without consent.  The company is set to appeal against the judgment.

It’s been a bad couple of months for Meta. In August, the company agreed to an $18 billion settlement over claims that Facebook and Instagram failed to prevent children from accessing its platforms. And in July, another judge in New Mexico fined the company just under $1 billion for, again, failing to protect minors.

In Europe, companies are also under pressure over data protection. Last year, breaches of the General Data Protection Regulation (GDPR) rose 22%, with a record €1.2 billion in fines issued, as legislators look to get tougher on big tech companies.

Internet je plný videí, ve kterých ukrajinské vrtulníky Mil vypouštějí neřízené střely ze svých raketnic po stranách. Doba ale pokročila, a tak je na některých strojích nahradily moduly s autonomními interceptory jako na čerstvém záběru, který se nyní šíří sociálními sítěmi. Ve videu vidíme ...

AI agents can be manipulated through malicious emails, websites, and plugins into executing commands, stealing data, or deleting files using their legitimate permissions. Gen Digital explains how its Agent Trust Hub helps protect agents through skill verification, secure connectivity, and runtime controls that intercept dangerous act [...]

There is no big red button or glass box hiding a lever.

In the wake of high-profile hacking incidents involving advanced artificial intelligence agents, some politicians and business executives have been calling for an AI kill switch: a safeguard that could shut down these AI systems in an emergency.

The idea is that, if a model were to behave in dangerous ways, a kill switch would allow humans to regain control over that system by metaphorically flipping a switch from “on” to “off.”

Several efforts at AI regulation in 2026 have focused on the notion of a kill switch. These include proposed congressional legislation such as the AI Kill Switch Act, and California Gov. Gavin Newsom’s executive order aimed at reining in AI. President Donald Trump has opposed the idea of an AI kill switch as unnecessary regulation, and to date there has been no national regulation that includes a kill switch.

Experts such as Dylan Baker, a research engineer at the nonprofit Distributed AI Research Institute, have explained that one regulatory problem with the metaphor of the kill switch is that today’s advanced AI models involve a complex infrastructure of distributed technologies, which is beyond the control of a single kill switch.

Additionally, the metaphor of the kill switch also obscures the human decisions involved in the responsibility and oversight of an out-of-control model. The problem is that the metaphor does not account for the technological or human aspects of AI.

Black Boxes and Kill Switches

The metaphor of the kill switch is not necessarily new. In fact, it is rooted in the same rhetoric as the black box metaphor, which similarly blurs the complex aspects of AI by framing inner workings as confined and unintelligible. Referring to AI as a black box frames these systems in terms of restricted incomprehensibility: a monolithic stranger to human control and understanding.

At the University of Cambridge, digital humanities researcher Eryk Salvaggio argues that “the black box of AI cultivates useful confusion about what can and cannot be known” about the complexity of frontier models. The idea of the kill switch cultivates a similar confusion by obscuring the distributed infrastructure and human influence that are important aspects of AI. While these metaphors have opposite meanings—resisting control and regaining control—they invoke the same implicit argument about the nuts and bolts of AI.

Both the black box and kill switch are what cognitive linguist George Lakoff and philosopher Mark Johnson call structural metaphors. They rely on simple physical concepts—for example, object, substance, up-down, and in-out. Structural metaphors use one clearly delineated concept to help make another concept clearer. For example, black box uses the concept of an opaque container to give a clear structure to the concept of hidden or inscrutable inner workings.

Additionally, Lakoff and Johnson say that this type of metaphor can also hide some aspects of a concept by focusing attention on others. For example, the kill switch metaphor hides the complexity of advanced AI models by focusing on the fact that a model can be on or off.

What this means is that framing the operation of AI in terms of a black box and kill switch relies on the straightforward ideas of “box” and “switch” to conceal the complex operations of AI. Reining in AI becomes difficult on a policy level when discussions about AI regulation fail to account for the infrastructural and human aspects of advanced models.

Centralization and ‘Rogue’ AI

There are two problems with the structural metaphor of an AI kill switch, both demonstrating that it misses the mark for a national safeguard. The first is that a kill switch makes the global infrastructure supporting AI appear centralized, when in reality it is a web of interconnected technologies spread across the globe.

The advanced models that companies like OpenAI and Anthropic are concerned about releasing rely on multiple computers and data centers that are spread out nationally. It is not as simple as pulling a plug, opening a box, or flipping a switch. The kill switch metaphor reduces the infrastructural entanglements of AI to outdated ideas about technology. For instance, in proposing his since-failed “AI Emergency Button Act” bill, Sen. John Kennedy compared the concept to kill switches in machines such as jet skis.

The second problem is that the kill switch idea keeps hidden the human decision-making involved in the training and deployment of frontier models. Rather than serving as a reminder that people are involved, the kill switch positions the human outside of the loop and ready to flip the switch during a catastrophic emergency. Human intervention is invoked as an afterthought, instead of a process that involves training, supervision, and approval.

For instance, accounts of how OpenAI’s model went “rogue” and hacked AI company Hugging Face depict the AI agents as though they operated on their own agenda. But experts have explained that not only did OpenAI personnel turn off safety mechanisms, but they also gave the model impossible tasks and left a door open to the internet. The role people played counters the “rogue” agent framing.

The kill switch metaphor is ill-suited to the challenge of regulating AI because there is no way to switch off an AI model without accounting for its distributed infrastructure and human influence. There is no big red button, emergency brake, or glass box hiding a lever. Effective national regulation of AI will not begin with any single switch or any one person’s finger.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

The post Why Calling for a ‘Kill Switch’ Is the Wrong Way to Think About Reining in AI appeared first on SingularityHub.

Cybersecurity researchers have disclosed details of a large-scale campaign that has targeted exposed network edge devices such as Fortinet FortiGate and Palo Alto Networks GlobalProtect firewalls. "Artifacts from this investigation also indicate that operators in this campaign have obtained access to numerous organizations worldwide, and have attempted to further compromise some victims,

Cybersecurity researchers have disclosed details of a large-scale campaign that has targeted exposed network edge devices such as Fortinet FortiGate and Palo Alto Networks GlobalProtect firewalls. "Artifacts from this investigation also indicate that operators in this campaign have obtained access to numerous organizations worldwide, and have attempted to further compromise some victims, Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Technologie SynthID vkládá do AI obsahu neviditelné digitální vodoznaky. • Detektor rozpozná obrázky z ChatGPT a Gemini, u řady dalších generátorů ale selže. • Technologie je slibná, zatím však nejde o univerzální řešení pro odhalování AI obsahu.

Představena byla (𝕏) virtuální platforma FLYBENCH FPV, která nováčky provede zapojením a konstrukcí dronů.

Hewlett Packard Enterprise (HPE) released security updates to patch 72 vulnerabilities in the Aruba Operating System (ArubaOS), including eight maximum-severity authentication bypass and code execution issues. [...]

It seems clear by now that one of the most widely shared criticisms of the iPhone Duo picked up by many of those who got to take an early look at the device was a little unfair. 

They complained that there weren’t many third-party apps when they looked at the device. But what made that point such weak sauce was when those reviews were written.

I don’t think most developers build apps with the media reviews schedule for new hardware in mind. (Perhaps they’d get more publicity if they did.) Instead, they tend to target products at the point when people can buy them, which in the case of the iPhone Duo meant October 23.

The apps are on the way. Many landed today

After those first reviews appeared, Apple published a press release describing more than 50 apps it knows are coming for its folding smartphone. These include key streaming services Netflix, Amazon Prime, and Paramount+. They include Microsoft Office products, Zoom, Slack, and Instagram.

While not included in Apple’s release, the Omni Group has released its first iPhone Duo products, while the team behind Ulysses have already updated their much-loved pro writing app. 

“The response to iPhone Duo from our developer community has been incredible, and we’re inspired by how quickly they’ve embraced its larger display and new ways to interact,” said Susan Prescott, Apple’s vice president of Worldwide Developer Relations. While you would expect Apple’s developer lead to say something like that, it remains a polite rebuttal to some of the objections critics raised to the device.

Not every criticism is unfair

That’s not to say every criticism can be rejected. You’d expect people to find weaknesses in a product aimed at the mass market with a two-thousand-dollar price tag. It is fair to let people know that the apps aren’t available yet, but it is also appropriate to say they’re on the way.

Finding those weak spots and reporting honestly on them is exactly what gadget reviews are all about. Apple did have to make some compromises to get the product out the door. That means no Face ID and a slightly poorer camera — you won’t be purchasing the Duo for pro photography or making a movie.

What the iPhone Duo 2 might fix

You need to think about tech as a continuum. Even as you read the critics slamming the lack of Face ID support and camera choice, you can somehow already tell those limitations will become improvements Apple will introduce in just over 11 months when the iPhone Duo 2 hits stores. 

Bloomberg’s well-connected Apple reporter, Mark Gurman, claims Apple is already working on an under-display Face ID sensor, with a view to introducing it in the next version of the device. Wait for v.2 if you really want those things.

Knowing the limitations doesn’t seem to have dented the enthusiasm for the product. Crowds gathered outside Apple Stores to greet the new iPhone dawn as Apple’s iPhone Duo (briefly) hit stores worldwide. Everyone knew availability was limited, but still they came, hoping to get ahold of the device. Seeing what was coming, Apple increased its iPhone Duo manufacturing orders.

Meanwhile, Apple just seems to have sold millions of phones at an astonishing $2,000 each in the middle of a cost-of-living crisis that has knocked around 20% off the size of the global PC and smartphone markets. How on earth did it manage to achieve that?

How Apple sold millions of Duos at $2,000 each

I think it’s a mix of things.

Apple’s customer loyalty helped. Good trade in deals and Apple’s shrewd introduction of Apple Upgrade earlier this year also helped. Carriers also seem to be offering decent deals on iPhone Duo; partly because they are feeling the pain of smartphone sales slowing down and must maintain customer sign-ups to bolster revenue. 

Excitement at the new product plus the well-covered perception that Apple had few units to sell probably stimulated interest; as did the fact that the people with $2,000 to spend on a first-generation device are unlikely to be feeling the same squeeze on household budgets as the rest of us. Apple’s brand certainly helped, as did the fact we’ve been talking about foldable iPhones since around 2011.

The market was primed. The market was ready. The apps arrived when the phone did.

Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core, or follow me on BlueSky, LinkedIn, or Mastodon.

Vivo X500 Pro Max je fotomobil pro náročné • Už samotné rozbalení fotografické edice zabere dost času • Kromě telefonu dostanete kryt, poutko, grip a hned dva přídavné telekonvertory

SAP professionals can now renew their certifications through AI-based roleplays. The new service builds on SAP’s revamp of its certification program last year, shifting the focus from rote learning of facts towards real-world use cases.

“The evolution of Stay Certified reflects a broader shift already taking place across SAP Certification: moving beyond assessing what someone knows toward validating how they can apply that knowledge. AI-powered roleplays bring that principle into the annual certification renewal experience,” the company said in a blog post announcing the new option.

The AI-based scenarios across more than 100 certifications simulate customer interactions with a defined knowledge goal. “The scenarios are grounded in content reflecting the latest SAP product releases, helping learners demonstrate applied, up-to-date knowledge,” SAP said. Immediate feedback is provided to highlight what went well during the roleplay, and what needed improvement.

However, the use of the AI isn’t mandatory: IT pros can still choose to complete a standalone assessment quiz when renewing their certifications.

The process

Users have a 90 day renewal window before certification expiry to complete their recertification. To do so, they must have a subscription to the SAP Learning Hub, at $1452 per user per year, which provides expert-led live sessions to help prepare them, access to hands-on practice systems, four SAP certification exam attempts per year and access to “enablement and assessment to keep certifications valid.” Alternatively, SAP said, users can purchase and retake a certification exam each year.

Yet another option, being piloted at SAP TechEd in Berlin, will be an event-based approach giving eligible users the ability to fulfill their Stay Certified requirement via on-site, hands-on sessions.

A move in the right direction

Analysts are guardedly impressed with the revised approach.

“I think SAP is moving in the right direction, and the ERP industry could have addressed this sooner,” said Robert Kramer, managing partner at KramerERP. “With cloud, AI and automation advancing at their current pace, traditional certification models are struggling to keep up. Knowing the software is one thing. Understanding how to apply it to a customer’s business is another.”

He likes the idea of AI-powered roleplays, because they move certification closer to the application. “Evaluating how someone handles a customer conversation, understands a business problem and recommends a solution is more meaningful than testing whether they memorized product features,” he said.

Frank Dickson, principal analyst at Dickson Research, agreed. The quiz-based approach to renewals was inconsistent, he pointed out: “A credential earned by doing should not be maintained by answering trivia. Airline pilots do not stay current by retaking a written test. They go back into the simulator. That is the right model for cloud software that changes constantly.”

Valence Howden, an advisory fellow at Infotech, added, “SAP’s revised approach to certification is a positive step and a good idea, although it’s not a new or novel approach. In-room and virtual confirmation of the ability to apply training beyond the exam has existed in the industry for a long time, from Cisco certifications a few decades ago to ITIL and other frameworks where top-level certification created scenarios to be solved and addressed in order to validate the ability to apply learning.”

So, he said, “SAP has taken an old but valuable method that will improve the value of their certification, the value of those who achieve it, and provide some assurance around their skills for those who hire them.”

Sanchit vit Gogia, principal analyst at Greyhound Research, agreed. “The AI roleplay is not new to SAP certification,” he said. “It arrived in first-time exams a year ago. What is new is that it has reached renewal, across more than a hundred certifications, and renewal is where SAP’s installed base lives.”

But, he said, “The deeper question is who comes to own the definition of competence, because SAP supplies the AI that helps the consultant, plays the customer, and assesses the exchange. That is the part history will weigh.”

Questions remain

While Kramer approves of the model, he still has reservations about the new testing regime. “A successful roleplay doesn’t necessarily make someone a capable consultant,” he said. “ERP implementations are rarely textbook exercises. They involve decades of customer processes, customizations, industry requirements, questionable data and competing business priorities. Sometimes the best recommendation is not to implement the latest capability, but to address the customer’s readiness first.”

And, he added, “there’s also the question of governance and accountability. Who decides what counts as a correct response? How are assessments validated? What happens when the AI gets it wrong? These questions matter when professional credentials and customer confidence are at stake.”

Dickson, too, has concerns. “The umpire is now an algorithm. Every roleplay is graded by AI, so the strike zone has to be the same for every batter,” he said. “SAP has not said how scoring is calibrated, audited, or appealed. Two consultants giving the same answer should get the same result. Staying certified also requires an SAP Learning Hub subscription, according to SAP’s own FAQ. SAP sells the training and grades the test. That makes transparency on scoring more important, not less.”

SAP is right to test whether certified people can still apply what they know, he noted, and the roleplay is a good simulator. “But a simulator still needs a fair, transparent grader,” he said.

He also has questions about the testing methodology. “Last fall’s exams allowed candidates to use AI tools,” he said. “The announcement does not say what is permitted during a roleplay. If AI assistance is allowed, the roleplay partly measures how well someone works with an assistant. That may be perfectly fine. SAP should simply say so.”

Kramer, too, is looking for more detail. “I support the direction, but I would recommend SAP provide metrics that AI-powered certification produces better professionals, not just a more efficient certification process,” he said. “If it does, this could help close the ERP adoption gap.”

This article first appeared on CIO.

Linus Torvalds is embracing AI to write Linux kernel code, and some people are having fits. He has a message for them. So do I. Roy Schestowitz, editor of Tux Machines, summed up his reaction to the news as "Linus Torvalds Turns to Slop." In the podcast Pivot to AI, podcaster David Gerard agreed: "Linux kernel goes full AI slop code." Meanwhile, in the Tom's Hardware comments section, commenter "LordVile" wrote: "This is a publicity stunt from his AI overlords [The Linux Foundation] who pay his wage." They followed this by saying this is the "same reason he's flipped-flopped on AI." I've known Linus for a few decades, and I don't take his use of AI as evidence that he's suddenly abandoned his standards. And the very idea that he'd do anything he didn't want to do with Linux on behalf of the Linux Foundation, or anyone else, is ludicrous. I get it, though. There are reasons to be suspicious of AI. I've written about many of them. For instance, AI model collapse, which happens when AI engines learn from other AI models and their ties to the truth fray. It's a very real problem. Not, mind you, that LLMs find the grist for their answer mills from very reliable sources in the first place. Even when AI gets good data, as it does in programming, its owners pay little to no attention to the copyrights and licenses protecting the open source code it picked up its coding tricks from. As for paying for the information it runs on, please! The very idea that Big AI must pay writers, artists, you name it, for using their work is anathema to them. All that said, when it comes to programming, the ship has sailed. AI is becoming part of the programming toolkit, whether its critics like it or not. Why? Because while you still can't vibe code your way into delivering production-quality software, when someone who knows what they're doing, like Torvalds, uses AI, it can be very useful. Anyone who says otherwise is just making the same mistake the Luddites did. Who now, some of you may ask? They were textile workers who attacked machinery in the early 19th century in protest against changes threatening their wages and livelihoods. You may have noticed that while people still weave cloth by hand, industrial production is mechanized. The same is happening now with AI and programming. I think we're about halfway through the process, and by 2030 or so, people programming by "hand" will be about as common as weavers doing it by hand are today. In short, get over it. I don't love it, either. I write for a living, and my business is being crunched too. At least in programming, quality work still counts, no matter how it's done. With writing, though, true AI slop is increasingly the rule, not the exception. Back to Linux: at the Linux Kernel Maintainers Summit in Prague the other week, Sasha Levin, a Linux kernel developer, announced that the maintainers had decided to "deprecate the Assisted-by Tag." That means you no longer need to say if you used AI in your code submissions. That doesn't mean you can turn in slop, though. That code won't pass muster. As Torvalds told me in a note the other day: "I think the whole 'don't use AI for commit messages' kind of argument … Maybe it's well-intentioned, but I think it's entirely misguided. I really think we want competent people to use AI as a tool." Competent is the key word there. Torvalds also wrote: "The 'clueless people using AI' problem is different. I do think we want to discourage that. Not discourage the people who use AI and truly want to learn and truly want to become useful members of the kernel community – but discourage the drive-by patches, and discourage people who just want to feel like they sent a patch and never really wanted to become anything more. We all start clueless, but some people spend the time and effort to become something more." Exactly so. As Torvalds said at Open Source Summit Europe: "I think [vibe-coding is] a wonderful way to find joy in programming. It makes you feel like you're doing something relevant. It allows you, as a new programmer, to do things that you would otherwise have a really hard time with." He added: "I love the concept of AI as a gateway drug" to programming. Besides, AI has already proven very useful for improving Linux code. As long-term Linux kernel maintainer Greg Kroah-Hartman told me back in March: "Months ago, we were getting what we called 'AI slop,' AI-generated security reports that were obviously wrong or low quality." But then, "something happened a month ago, and the world switched. Now we have real reports." It's not just Linux, he continued. "All open source projects have real reports that are made with AI, but they're good, and they're real." True, the maintainers are getting buried in good bug reports now, but as Kroah-Hartman and discussed in Prague, AI bug finders might eventually exhaust the readily discoverable problems, easing the pressure on maintainers. Another AI win is that you can use AI to reverse-engineer Windows drivers for Linux more easily. Christian F.K. Schaller, Red Hat's director of software engineering, created hw-harness while working on a Linux driver for a slide scanner. This open source project provides "tooling for working out how a piece of hardware behaves so that you can write a Linux driver for it." By automating some seriously slow, tedious ways of figuring out what a binary-only driver was doing with the hardware, Schaller has made building Linux drivers from Windows drivers much easier. I know, because I did this myself back in the mid-2000s, and those manual methods are still what people use today. This could go a long way toward fixing one of those lingering annoyances of the Linux desktop: poor or entirely missing hardware drivers. In the meantime, Torvalds told me: "Well, in all fairness, we do clearly have an 'AI problem.' Or rather, we have two independent AI problems: we have a large codebase that AI finds real bugs in, and that adds a lot of work on already fairly overworked maintainers. And then we have people who use AI to generate patches without actually understanding the code in question." He continued: "But I'd argue that these aren't problems with AI itself; they are just problems that are exposed by AI. They are very much both very real problems, and I understand that people can really come to dislike – and even hate – AI due to them, but I do think the underlying issue isn't AI itself." It's how people misuse AI. In the right hands, as the top Linux programmers are showing us, AI can be useful. ®

PHP has warned of another way an attacker could run code on a web server, two days after it shipped emergency fixes for the critical PHP2Shell flaws. The new issue, assigned CVE-2026-115560, affects only web applications configured in an unusual way, and PHP considers it an unlikely risk.The project is not issuing an emergency fix for it. Instead, it recommends turning off a feature called

PHP has warned of another way an attacker could run code on a web server, two days after it shipped emergency fixes for the critical PHP2Shell flaws. The new issue, assigned CVE-2026-115560, affects only web applications configured in an unusual way, and PHP considers it an unlikely risk.The project is not issuing an emergency fix for it. Instead, it recommends turning off a feature called Swati Khandelwalhttp://www.blogger.com/profile/[email protected]

The Russian state-sponsored hacking group Sandworm teamed up with another threat actor to carry out a destructive wiping attack targeting a Ukrainian logistics company last year. According to ESET, the attack involved UAC-0099 (aka Earth Sirrush) obtaining initial access to the target on July 16, 2025, via a spear-phishing email bearing a malware-laced attachment that led to the deployment of