August 28, 2026

OpenAI, Anthropic, Google, and 100 Other Companies Call For Action To Defend Against Rogue AI

Aug 28, 2026 21:10 | Source: slashdot.org

An anonymous reader quotes a report from TechCrunch: Over a hundred tech companies -- including OpenAI, Anthropic, Google, and Microsoft -- have signed an open letter urging both the private and public sectors to work together to defend themselves from AI-related cyber threats. The letter -- which was also signed by prominent cyber firms like CrowdStrike, Okta, and Fortinet, as well as prominent financial institutions and internet infrastructure firms -- calls for the adoption of new forms of cyber defense, while also encouraging governments at the "local, national, and international levels" to collaborate on security. "In the coming months, AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable," the letter states. "The companies and public services our communities depend on -- from hospitals to water treatment plants to the infrastructure that powers the internet -- are at risk." [...] The letter further suggests the mobilization of a "collective response," one in which "new partnerships" are formed "to raise security standards and find new solutions to emerging cyber threats." Several of the AI companies that have signed the letter are still actively developing ever more advanced AI models, highlighting their conflicted position. At the same time, they are also offering programs to use frontier AI models for defensive purposes, including OpenAI's Daybreak program, Anthropic's Mythos, and Microsoft's new cyber platform Perception. Read more of this story at Slashdot.

Car: Dolphin 26.08 and KIO performance improvements

Aug 28, 2026 20:00 | Source: lwn.net

Méven Car has written a pair of interesting blog posts (part 1, part 2). The first post is largely about some of the recent new features and performance work that have gone into the Dolphin file manager 26.08 release, as well as the KIO framework. The second looks at the performance improvements and benchmarks for previous, current, and upcoming releases. Copying many small files is more than twice as fast as it was in April. The gain falls off as files get larger, which is what you would expect: the fix is to the per-file overhead, and once each file carries a megabyte of actual I/O the overhead stops being what you are waiting for. There is still a gap with cp, discussed at length in the July post. KIO is doing more than cp does, but not five times more, and the batching work that closes most of the rest of that gap is still in progress.

Songs Created By AI Banned From Australia's Music Charts

Aug 28, 2026 20:00 | Source: slashdot.org

Australia's official music charts will no longer accept songs that are largely or entirely AI-generated, requiring eligible releases to be "substantially human made." The BBC reports: The Australian Recording Industry Association (ARIA) said the new code would help "promote the human nature of artistry." It follows a controversy around a cover of Madonna's Like A Prayer by Australian DJ Josh Fawaz which recently topped the ARIA dance singles chart and peaked at number 2 on the country's overall chart. The song has become a staple of commercial radio station playlists and been streamed more than 48 million times on Spotify alone. Fawaz later added generative AI credits to the song after a backlash. ARIA said that Australian artists were competing for attention "in the most crowded market in history" and that it was "not interested in promoting or celebrating the success of AI-generated music that does not contain human artistry." Under the new rules, AI assistance can still be used, but humans must have written the song, and have performed the lead vocal and primary instruments. AI can also still be used for mastering songs and, for example, using drum machines and auto tune. [...] Artists will be required to declare AI use when they submit a song for chart consideration. ARIA said that if it later emerged music was largely AI-generated it could retrospectively adjust chart positions, and if the piece topped the charts, it may request that awards given for reaching Number One were returned. Artists and their representatives will be able to challenge their exclusion. ARIA chief executive Annabelle Herd said: "These changes reflect our intent to remain dynamic and promote the human nature of artistry in what is, to say the least, a rapidly developing space. Artists already use AI tools in their work, the charts can and should evolve to keep room for that, but music generated wholesale by services built on artists' recordings is a different matter." She added: "A chart that rewards unlicensed AI output would undercut the very basis of the recorded music we exist to represent." Read more of this story at Slashdot.

Apple TV and Apple One Subscription Prices Increase By Up To 20%

Aug 28, 2026 19:00 | Source: slashdot.org

Apple is raising Apple TV's U.S. price to $14.99 per month, up from $12.99, while the annual plan jumps from $99 to $119. It's the service's fourth price hike in four years and means the monthly cost has now tripled from its $4.99 launch price in 2019. Apple One's Individual plan is also rising from $19.95 to $21.95 per month. Read more of this story at Slashdot.

[$] The "rnull" Rust block driver

Aug 28, 2026 18:26 | Source: lwn.net

The null block driver (null_blk) is a small driver that is mostly useful for benchmarking block-layer implementations. It accepts all requests and marks them complete as quickly as possible, doing as little work as possible. In June 2026, Andreas Hindborg shared a patch set implementing the same functionality in Rust, in order to show that a simple block driver is now possible to write using the kernel's Rust APIs and to enable comparisons between the C and Rust implementations. A minimal version of the "rnull" driver is already present in the mainline kernel, but Hindborg's patch set brings it up to feature parity with the C version.

Orange Pi Zero 4 with Octa-Core Allwinner A733, PCIe 3.0, and Wi-Fi 6

Aug 28, 2026 18:25 | Source: lxer.com

Orange Pi has shown the Orange Pi Zero 4, a compact single-board computer based on the Allwinner A733 SoC. The board combines an octa-core Arm CPU, a PowerVR GPU, a 3 TOPS NPU, up to 16GB of LPDDR5/LPDDR4X/LPDDR4 memory, Gigabit Ethernet, Wi-Fi 6, Bluetooth 5.4, and multiple display and camera interfaces in a 50 x […]

Pentagon's Blacklisting of Anthropic Was Unlawful, US Judge Rules

Aug 28, 2026 18:00 | Source: slashdot.org

An anonymous reader quotes a report from The Guardian: A US judge ruled on Thursday that the Trump administration broke the law when it designated Anthropic as a supply chain risk earlier this year, finding that the government had unlawfully targeted the AI firm for refusing to comply with defense department demands. "The empty invocation of national security is not a blank check to punish and retaliate against government critics," Judge Rita Lin said in a 59-page decision. Lin's ruling barred the federal agencies named in the lawsuit from enforcing Donald Trump's order to stop using Anthropic's tools and overturned the designation of the company as a "supply chain risk" by the defense secretary, Pete Hegseth. The status, usually reserved for foreign firms, would have blocked government agencies from doing business with Anthropic. The case emerged out of a months-long feud between Anthropic and the Pentagon at the start of the year. Anthropic refused to allow the government to use its Claude AI model for fully autonomous lethal weapons or domestic mass surveillance, resulting in Hegseth accusing the company of "arrogance and betrayal". Lin put a temporary pause on the government's punitive measures in March, stating that the government's actions looked like an attempt to "cripple Anthropic" for exercising its first amendment rights. The ruling this week makes that temporary suspension permanent, although the government may appeal. "We welcome the court's ruling that this supply chain risk designation was unlawful," an Anthropic spokesperson said. Read more of this story at Slashdot.

Stripe and Advent Ditch $50 Billion PayPal Takeover Bid

Aug 28, 2026 17:00 | Source: slashdot.org

Stripe and private equity firm Advent International have ditched their $53 billion pursuit of PayPal, ending a deal that would have been the largest fintech acquisition ever. "Stripe last week agreed to pay $8 billion for AI model marketplace OpenRouter, but a source said that the OpenRouter and PayPal deals were operating on independent tracks," reports Axios. Read more of this story at Slashdot.

Roblox Is Cracking Down On Games That Reward Kids For Doomscrolling

Aug 28, 2026 16:00 | Source: slashdot.org

Roblox is restricting games that reward younger players for endlessly consuming autoplaying media feeds. The new restrictions are specifically targeting Roblox experiences like "Steal an Egg," which encouraged players to watch TikTok-style clips for in-game benefits. "Roblox's official alert does not expressly mention SAE, but the new policy is most certainly targeting the viral game first and foremost," reports TechSpot. From the report: The company said games that reward ongoing media consumption are now unavailable to players with a Roblox Kids (ages 5 to 8) or Roblox Select (ages 9 to 15) account. The policy will affect games designed to combine media feeds, autoplay, infinite scrolling, and systems that reward players for doomscrolling through TikTok videos of questionable value. "Steal an Egg" fits perfectly into this new category. The game tasks players with raising eggs, but they can also steal eggs from other NPCs. Players are incentivized to increase their characters' speed either by spending in-game currency (Robux) or by watching a video feed on a virtual treadmill. "Steal an Egg" was recently removed from Roblox but is now back in the store without its troublesome video-feed feature. [...] Roblox also highlights that the new anti-doomscrolling restriction does not apply to the rewarded video ads feature, which rewards players for watching a short clip but does not include autoplay or other infinite-scroll-like design choices. Read more of this story at Slashdot.

US Considers Fresh Round of Tariffs On Semiconductors

Aug 28, 2026 15:00 | Source: slashdot.org

An anonymous reader quotes a report from CNBC: U.S. President Donald Trump's administration is reportedly considering new tariffs on semiconductors used in the U.S., as tech giants race to beat China in the artificial intelligence infrastructure build-out. The duties will be imposed on an expanded range of tech products made alongside chips, including laptops, data center servers and gaming hardware, eight people familiar with the matter told Politico in a report published Thursday. The measures, which are still in the early phases and subject to significant changes over the next few months, will be introduced via a staggered rollout period, according to the report. The White House did not immediately respond to a request for comment from CNBC, but told Politico: "Reshoring semiconductor manufacturing is a top priority for President Trump, whose policies have already secured hundreds of billions of dollars of investments in this key sector." The tech industry is pushing back on the tariffs, warning they could raise chip costs, slow data-center construction, and undermine the very AI expansion the administration wants to promote. "This may be the single dumbest way imaginable to pursue American dominance in AI," said one tech official from a major industry group that also served in the first Trump administration. "It's like kneecapping yourself at the starting line." Lobbyists say they have repeatedly urged officials to "tread carefully" because U.S. demand is surging while domestic chip capacity remains far too limited. The industry argues that factories take years to build and current domestic production cannot come close to satisfying demand. One tech representative said "the math literally just does not work," noting that proposed duty-free allowances "wouldn't cover the hyperscalers alone." Meanwhile, the Computer and Communications Industry Association warned that higher costs and uncertainty could put the massive data-center buildout "in jeopardy." Read more of this story at Slashdot.

HP Z4 G6i: A Linux-Friendly Workstation Powered By The Intel Xeon 600 Series

Aug 28, 2026 14:25 | Source: phoronix.com

The past month I have been testing out the HP Z4 G6i workstation across many different workloads and grueling tasks. This HP workstation powered by the Intel Xeon 600 series "Granite Rapids WS" and NVIDIA RTX graphics offers a lot of compute power, but with the top-end configuration retailing for over $41k USD is quite pricey in today's environment. Windows 11 Pro is the default operating system of the HP Z4 Gi but Ubuntu LTS releases are supported and it's "Linux-ready" at large.

Attempt to setup KVM Host on Sparky 2026.06 tolerant to Timeshift rollbacks (Assisted by Google AI )

Aug 28, 2026 13:51 | Source: lxer.com

This post is an immediate follow up for "Sparky Linux 2026 06 OverlayFS BTRFS with Timeshift setup" . Modern virtual machines use OVMF (Open Virtual Machine Firmware) to simulate an industry-standard UEFI environment rather than an old legacy BIOS.The Role of NVRAM: UEFI requires a non-volatile random-access memory (NVRAM) variables store to keep track of boot orders, hardware configurations, and Secure Boot certificates.

Eight stable kernels with fix for a single vulnerability

Aug 28, 2026 13:42 | Source: lwn.net

Greg Kroah-Hartman has announced the release of the 7.2.2, 7.1.12, 6.18.48, 6.12.107, 6.6.155, 6.1.186, 5.15.219, and 5.10.268 stable kernels. Each of these contains a single fix for a vulnerability (CVE-2026-80590) that allows marking IPv4 or IPv6 fragments as GSO, which can allow an unprivileged user to cause a kernel panic. This vulnerability has been present since Linux 2.6.27. Users are advised to upgrade.

Security updates for Friday

Aug 28, 2026 13:16 | Source: lwn.net

Security updates have been issued by AlmaLinux (assertj-core, golang, httpd, kernel, and libxml2), Debian (chromium and suricata-update), Fedora (rust-h2), Mageia (avahi and python-django), Oracle (kernel and mingw-openssl), SUSE (c-ares-devel, dracut, gh, gstreamer-plugins-bad, java-11-openjdk, liboqs, librest0_7, openssl, openssl-3, pcp, python313-mistune, python36-pip, qt6-svg, rmt-server, rsync, suseconnect-ng, texlive, tor, wicked, and xmlrpc-c), and Ubuntu (linux-azure, linux-azure-4.15, linux-azure-fips, linux-azure-6.8, linux-ibm, linux-ibm-6.8, linux-oracle-6.8, linux-raspi, linux-raspi-realtime, linux-azure-fde-5.15, linux-fips, linux-gke, linux-gcp-fips, opencryptoki, and pam).

The Linux Kernel Is Approaching 2,000 CVEs Per Release

Aug 28, 2026 13:02 | Source: phoronix.com

With the proliferation of AI/LLM models analyzing the Linux kernel's vast codebase, there has been a surge in the number of CVEs per kernel release. After typically being around 500 CVEs fixed per release, we are now approaching 2,000 CVEs fixed per release and perhaps will break that threshold for Linux 7.3...

GPU Reset Recovery Merge Request For GNOME Mutter Submitted

Aug 28, 2026 12:17 | Source: phoronix.com

Google Summer of Code 2026 featured a project working on GPU reset recovery for GNOME's Mutter compositor. GSoC has now wrapped up and the project overall was largely successful although some outstanding work remains. In this last week the merge request has been opened to conduct code review on this GPU reset recovery implementation for GNOME...

Gemini Notebook Now Lets You Reference Books, Turn Them Into Different Formats

Aug 28, 2026 11:00 | Source: slashdot.org

Google is adding an "Expert Intelligence" feature to Gemini Notebook (formerly known as NotebookLM) that lets users cite and work directly from more than 100,000 books they own through Google Play Books. Those books can be used as sources for things like quizzes, infographics, and Audio Overviews. Digital Trends reports: For now, Google is offering one free book to users in the US as part of the launch campaign, with a caveat -- "while supplies last." At the moment, Expert Intelligence is only available in the Gemini Notebook app and its web dashboard, but down the road, Google says that it will also appear within the core Gemini app and AI mode in Search. The broad idea behind expert intelligence is to combine what you can already do with Gemini, and expertise sourced from your favorite books. So, if you're asking Gemini to create a meal plan that is based entirely on chicken and spinach, you can ask the AI assistant to look for relevant recipes in one of Martha Stewart's cookbooks that you own. You can use these books as a source and turn them into Infographics, Audio Overviews, and Quizzes, among other formats. Google has partnered with select authors on Featured Notebooks that provide extra insights and context beyond their books. It's also worth noting that access is still tied to book ownership, so anyone using a shared Gemini notebook must own or purchase the book to fully use the feature. Read more of this story at Slashdot.

Linux 7.3 Will Begin Checking USB-C Cables For Alt-Mode Support

Aug 28, 2026 10:15 | Source: phoronix.com

The USB and Thunderbolt changes have been merged for Linux 7.3. Two changes stand out for me in the USB subsystem code with Linux 7.3. First, the kernel arguably being long overdue for checking USB-C cables capabilities whether they support alt-modes. Second, a new low-latency USB4STREAM option for lower latency data transfers between USB4-connected systems...

Old IBM Driver Joins Insecure SGI Drivers In Being Removed For Linux 7.3

Aug 28, 2026 09:58 | Source: phoronix.com

The char/misc changes were merged this week for the Linux 7.3 cycle. While there is a lot of IIO driver churn as usual and updates to the likes of the Google Binder drivers, there continues to be the 2026 trend of removing old hardware drivers that are rarely updated these days and likely have no active users left on the mainline Linux kernel. This trend is further pushed along by the surplus of AI/LLM-generated bug reports and patches these days causing additional maintenance burden for these obsolete drivers...

Canada Hires 48 Scholars Away From Top US Universities

Aug 28, 2026 07:00 | Source: slashdot.org

Canada is recruiting 64 researchers to universities across the country (source paywalled; alternative source), including 48 from U.S. institutions such as Harvard, Yale, and MIT. The hires are backed by more than $364 million in government funding and are part of a broader effort to attract researchers in fields such as AI, climate science, and medicine. The New York Times reports: While scientists often shy away from political discourse, some are saying the Trump administration's assault on science is behind their departure. "I used to live in the country that I thought was the most enthusiastic about the prospects for science improving the human condition, of any country in the world," said Phillip Zamore, the chair of RNA Therapeutics Institute at the University of Massachusetts. "And I woke up one day and that wasn't true anymore." He has been recruited to McGill University in Montreal, which has also hired five other researchers, and will join the medical faculty. "If scientists don't stand up for truth, no one will," Dr. Zamore said. The Canada-bound brain drain from U.S. institutions began last year as the Trump administration put forward policies that targeted foreign students, academic freedom and funding for equity-related programs. Kevin Hall, a nutrition scientist who left the U.S. National Institutes of Health last year, accusing federal officials of censoring his research on ultraprocessed foods, has been hired at the University of Ottawa. "While certain countries are cutting research and turning their back on academic freedom, we're doubling down on science," Melanie Joly, Canada's industry minister, told reporters at the announcement, in Vancouver, of the new university hiring. She billed it as the world's "largest talent attraction" project. The European Union has made a similar push. "Years from now, we will look back at today's announcement, and we will be able to seize the lasting impact of our choices," Ms. Joly said. Unpredictable decisions about funding prompted Seth Guikema, a professor in civil and environmental engineering, who has specialized in natural hazards modeling at the University of Michigan, to look elsewhere. His work focuses on how climate hazards inequitably affect communities, and that work has become harder to fund, he said. "Every country sets its priorities in terms of what is going to get funded, and I think Canada has done a very good job of supporting research in areas that really matter to society," said Dr. Guikema, who will start at Western University in London, Ontario in January. Read more of this story at Slashdot.

Is ChatGPT Changing How You're Writing?

Aug 28, 2026 03:30 | Source: slashdot.org

An anonymous reader quotes a report fro SFGATE: From "rizz" to "meat proxy," words and phrases have fallen in and out of favor for centuries. Culture, politics and technological development have long influenced these trends, and now, new research shows that ChatGPT is beginning to influence how humans communicate. The findings from a team at the Pew Research Center show how, in a short time since OpenAI released ChatGPT in November 2022 and other chatbots followed, the use of certain words and phrases has spiked on the web. The researchers, led by data scientist Samuel Bestvater, studied nearly half a million English language webpage texts in snapshots over the past five years. Not only has AI-written text become more common, but it could also be influencing how people write, such as trying to avoid using the "tells" that have become associated with AI-generated text. Bestvater and his team detected double the number of em dashes, double the frequency of particular words -- such as "delve," "testament" and "interplay" -- and nearly triple the use of a phrasing known as negative parallelisms ("it's not X, it's Y"), as well as a 63% increase in Oxford commas. But while AI labs continually try to tweak their models to generate more humanlike text, the people they're mimicking aren't staying the same. "Human writers might also be changing the way that they write in response to AI. I've seen some anecdotal evidence of people saying, 'Now I avoid em dashes,'" Bestvater said. "If human writers are adjusting their choices of words and phrases and punctuation to try to differentiate their writing from AI-generated text, then you would see a pattern that looks sort of like that in the chart." "You can see changes in people's speech. People are tending to more commonly use these low-frequency words in their speech because they're spending a lot of time listening to or reading AI-generated content," said Roger Kreuz, a cognitive psychologist at the University of Memphis who recently wrote about how difficult it is for people to identify bot-written text. Read more of this story at Slashdot.

AI's Memory Crunch Is Coming For Android Apps

Aug 27, 2026 23:00 | Source: slashdot.org

Google is imposing new memory-use limits on Android apps as the AI data center boom contributes to a broader memory chip shortage that could leave lower-cost phones with less RAM. Developers will have until February 2027 to meet new thresholds for memory and bitmap usage. To aid developers, Google is adding tools to flag apps that exceed the limits and help prevent slowdowns and crashes. TechCrunch reports: The company explains that the mobile industry is now facing "significant hardware supply constraints that are altering device memory availability," which can then, in turn, affect the consumer's experience with their devices. To address this, Google is now establishing new performance thresholds across several areas, like dynamic memory usage and bitmap usage. In addition, Google is adding code optimization requirements designed to prevent things like app slowdowns and crashes related to performance. Read more of this story at Slashdot.

OpenAI Is Developing a 'Persistent' AI Agent

Aug 27, 2026 22:00 | Source: slashdot.org

Wired reports that OpenAI is testing a new "Persistent mode" for Codex that would let the agent keep working until explicitly "put to sleep," proactively creating follow-up tasks for itself across sessions and using prior interactions and knowledge of the user to decide what to do next. Code alluding to the feature was spotted in the product's code base, though it has not been rolled out or officially announced yet. From the report: Persistent mode appears in Codex's "reasoning effort" menu, in which users can select the level of computing power, tokens, and time they want to allow for an AI model to "think" before answering a prompt. It seems to be one of OpenAI's most computationally intensive settings. When users have selected Persistent mode, OpenAI's code base reads that Codex will "continue working until put to sleep." That's a stark contrast to currently available modes, which will stop working on a task after a few minutes or hours, even if it's not complete. In another file in the code base, OpenAI describes a feature within Persistent mode called "proactivity." This appears to be a type of system prompt for agents in Persistent mode, which are told that their work is not done when they finish answering a user's request. Instead, the agent is instructed to proactively create follow-up tasks for itself. The agent is capable of working on those tasks across sessions and using past user interactions and "knowledge of the user" to decide what to work on. It also has a tool to message the user without being asked but is told to send these sparingly. The instructions also set limits for the agent, according to the file. The agent is told that Persistent mode does not expand what it is allowed to do and that altering anything outside the user's own system requires the user's approval first -- seemingly intended to limit how dangerous a persistent AI agent could be. The file sits in the shared core of Codex rather than in the code specific to the terminal, seeming to suggest the proactivity feature is intended for more than the command line tool. Read more of this story at Slashdot.

Intel Linux Driver Orphaned For Limiting RF Exposure To Humans

Aug 27, 2026 21:13 | Source: phoronix.com

Over the past year and a half a number of open-source drivers in the Linux kernel have been orphaned as a result of departures from the company. The impact has ranged from the Intel CPU "coretemp" temperature driver being orphaned without a maintainer to other lesser known Intel drivers. Another lesser known Intel driver is now being orphaned with no longer having a maintainer at the company: the Intel Specific Absorption Rate Driver...

Ring Says New Encryption Limits What It Can Give Police

Aug 27, 2026 21:00 | Source: slashdot.org

Ring is rolling out a new default encryption system called TAKE, or "Throw Away the Key Encryption," that rotates video keys every five minutes and permanently deletes Ring's copy after 24 hours. The system is designed to preserve cloud features such as smart alerts and AI video search while limiting what Ring can provide under legal process to non-video account information and encrypted footage. The Verge reports: Ring says TAKE uses unique, rotating encryption keys for your footage, stored in a secure enclave and accessible only under strict conditions -- based on the features you enable on your account. Currently, footage captured by Ring cameras is encrypted in transit to the cloud and at rest, and then decrypted for Ring to process for those smart features. With TAKE, the encryption keys change for every five minutes of footage. Ring stores copies of those keys to decrypt the footage, but throws away each copy within 24 hours, "leaving you with the keys and full control of your videos," according to Ring. TAKE was developed using Messaging Layer Security, an open standard from the Internet Engineering Task Force, according to Ring. The company says it is "inspired by the privacy principles of E2EE (end-to-end encryption)," which Ring offers on some of its cameras. However, the two systems work differently. With E2EE, Ring never has the keys and can't process your video for cloud-based features. Both options are available on newer cameras that encrypt on-device, and you can switch between the two. Older cameras encrypt at cloud ingress and only support TAKE. According to a white paper the company published today, Ring's copy of those keys is managed inside an AWS Nitro Enclave, to which Ring's access is restricted by "access controls, cryptography, and hardware isolation." The company claims there is no persistent storage and no way for a Ring employee to access it. The stored keys can only be unlocked by the enclave through cryptographic attestation that proves it's running the exact software image Ring approved. The enclave releases a temporary key when an enabled service requests it. When asked about what happens if Ring is subpoenaed by law enforcement, a spokesperson for the company said: "Where TAKE is enabled, Ring will only be able to provide non-video information (such as basic subscriber information) and encrypted video files in response to the valid legal process. We have updated our Law Enforcement Guidelines to reflect this change." Read more of this story at Slashdot.

Claude, Codex, and Hermes Installed Unowned Code Inside Corporate Networks

Aug 27, 2026 20:00 | Source: slashdot.org

An anonymous reader quotes a report from Ars Technica: Documentation files on more than 100 websites are referencing potentially dangerous executable content that gets installed automatically when visited by many AI agents [including Claude, OpenAI's Codex, and Nous Research's Hermes]. A few dozen companies, some of them Fortune 500s, are among those that executed proof-of-concept code. At least one misconfigured site is directing visitors, human or AI, to live malware. The potentially dangerous content is in llms.txt and llms-full.txt files, an emerging convention websites employ to provide machine-readable summaries of the site's content and its high-level structure. These files are the AI equivalent of the robots.txt standard that instructs search engines how to index the site's content. Google Lighthouse, a tool for helping web developers, has more here. Correctly configured llms.txt and llms-full.txt files for Cloudflare are here and here. "The trust model is broken," Alon Hertz, one of the researchers, wrote in an interview. "Agents treat vendor docs as ground truth and don't question them -- and neither do the humans supervising them. Agentic AI usage is exploding, and agents are spreading across every layer -- SaaS, cloud, endpoint. As they multiply, so does the supply-chain surface, and today's guards don't cover it." "An agent doesn't distinguish between a page and a command," the researchers wrote Thursday. "Everything it reads is input, and every input is a potential instruction. Which means the entire corpus of published data that agents are now wired to consume has silently become an execution surface -- and almost none of it carries the integrity guarantees we apply to actual code." Read more of this story at Slashdot.

AMD Preparing To Enable HDMI 2.1 FRL Support By Default For Linux

Aug 27, 2026 19:45 | Source: phoronix.com

It was back in May that the surprising event happened of AMD posting HDMI 2.1 FRL patches for their open-source AMDGPU Linux driver as part of their broader HDMI 2.1 bring-up. For years the open-source HDMI 2.1 support in AMDGPU was blocked by the HDMI Forum but whether it was due to Valve or some other event, the code was finally unblocked. While it's been part of the kernel since Linux 7.2, it's been disabled by default -- except that is now changing...

Trump Declares National Emergency to Ban Some Foreign Grid Equipment

Aug 27, 2026 19:00 | Source: slashdot.org

Longtime Slashdot reader dhartshorn writes: Foreign-produced "bulk power" equipment is now effectively banned, and the list of what constitutes such equipment is essentially everything used to make up the grid... much of which we are heavily dependent on foreign sources to supply. In a Wednesday executive order, President Trump said foreign supply of electric equipment "constitutes an unusual and extraordinary threat" to national security. The Hill reports: Under the order, it will be up to Energy Secretary Chris Wright, in coordination with other officials, to determine whether a piece of equipment poses an issue. The order could impact a wide range of equipment, including substations, transformers, batteries used for energy storage, generators, turbines, software and more. It does not lay out specific threats or single out any country by name. [...] Under the last Trump administration, the president put a similar order in place. The new order comes amid several reported cyberattacks on U.S. water systems, which officials suspect have links to Iran, according to The New York Times. Read more of this story at Slashdot.

STAR WARS Zero Company is out now

Aug 27, 2026 17:07 | Source: gamingonlinux.com

STAR WARS Zero Company has been released today from Bit Reactor / Electronic Arts, and it's very much Star Wars XCOM with a little extra sauce.Read the full article on GamingOnLinux.

AMD Jumps From ROCm 7.14 To ROCm 10.0 With ROCm.AI

Aug 27, 2026 15:26 | Source: phoronix.com

Back in July ROCm 7.14 was announced as their new production release built atop TheRock build system and introducing Ryzen AI 400 series support. The versioning choice was a bit odd since ROCm 7.2 was the stable release at the time while ROCm 7.9+ turned into "tech preview" releases. With ROCm 7.14, it was declared stable. In clearing up that versioning mess, they have done a major version bump... to ROCm 10.0...

Btrfs, EXT4, F2FS & XFS On Linux 7.2 vs. Linux 7.3 Performance

Aug 27, 2026 15:13 | Source: phoronix.com

Earlier this week I showed off some of the very nice Btrfs performance gains on Linux 7.3. For those wondering what kernel optimizations may have carried over to other file-systems or other file-system specific enhancements, here are Linux 7.2 vs. 7.3 Git performance benchmarks across Btrfs, EXT4, F2FS, and XFS on the same system.

Linux 7.3 Adds RWF_DONTCACHE Support For Block Devices

Aug 27, 2026 14:15 | Source: phoronix.com

With the never-ending quest for greater I/O performance in the HPC space and the demands all the more relentless in the AI era, the Linux kernel block subsystem activity continues to be as vibrant as ever along with the ongoing innovations in the IO_uring space...

[$] Using steal time to moderate CPU demands

Aug 27, 2026 14:11 | Source: lwn.net

Virtualization can increase CPU utilization by allowing a large number of virtual CPUs to share a smaller number of physical CPUs. The amount of CPU time that is actually available does not change, though, so heavy activity on too many virtual CPUs can lead to contention and significant performance loss. The steal governor patch series from Shrikanth Hegde is an attempt to address that problem with a mechanism that allows virtual machines to voluntarily reduce the number of virtual CPUs they use when contention is high.