Bruce66423 shares a report from The Telegraph: British police forces, including specialist units within the Metropolitan Police, have used Oxygen Forensics software to break into the phones of suspects during live investigations, public documents show. The Virginia-based company behind the software has been accused in the US of having hidden its Russian ownership to avoid sanctions, before it was awarded government contracts. Its American chief executive and a Russian national were arrested last week after a US investigation found the company had sought to hide its ownership in an alleged attempt to avoid sanctions and scrutiny. [...] By September 2024, the US Secret Service had awarded Oxygen a five-year contract for its software. In December 2022, and in October 2023, Oxygen's chief executive told the US government that Oxygen Forensics had "no immediate or highest-level owner." By September 2024, the US Secret Service had awarded Oxygen a five-year contract for its software. In March, Mr Reiber allegedly told the US government that there had been no Russians involved in developing the software and that no one in Russia had access to the environment in which it was built. US prosecutors say this was false. If convicted, both Mr Reiber and Mr Davydov could face a sentence of 20 years in prison.

Read more of this story at Slashdot.

Cloudflare has launched two open-weight decision models, Clef and Clef-flash, designed for structured yes/no, multiple-choice, and ranking tasks while also supporting images, video, and up to a 64K context window. Cloudflare says its models outperform TypeSafe's Jev on several benchmarks and can run locally from Hugging Face. The Register reports: For starters, Clef has an LLM backbone. According to Cloudflare, Clef uses specially post-trained, frozen versions of Qwen3.8-27B and Qwen3.5-9B for Clef and Clef-flash, respectively, with the Qwen backbone performing a prefill-only pass during inference. Clef is still fast - faster than Jev, to be fair - and scores choices in parallel after that prefill-only pass. It's not clear what Jev's underlying architecture is, as TypeSafe has kept that a secret. As for its speed and capability, Clef moves fast. Cloudflare ran it against Jev and some other open decision models using the Jev Decision Index available on Hugging Face, and the company's own ranking suggests Clef is slightly slower than other open models, but more accurate, with Clef-flash just as accurate as most of the others, but far faster. To be fair to the competition, Cloudflare self-reported its own scores against the benchmark, and they have yet to be reproduced for ranking on the official Decision Index. Cloudflare also ran Clef against TypeSafe's own benchmarks, and claimed it beat Jev in three out of four areas, only losing out on agent trace observability. Even if it were a bit slower or less accurate, Clef has another major leg up on Jev: It's not limited to classifying text -- it can also handle images and video. Additionally, Clef supports a 64k context window. Jev can also handle up to 64k tokens across a request, although its state plus longest individual question is limited to 32k.

Read more of this story at Slashdot.

Ryan Roslansky is leaving Microsoft and LinkedIn after nearly 18 years, triggering another leadership shuffle across Office and Teams. "Roslansky, who until recently was the CEO of LinkedIn, was promoted to the head of Office last year and then took control of Microsoft Teams earlier this year," notes The Verge. From the report: "Ryan leaves the organization in a strong position," says Microsoft CEO Satya Nadella in an internal memo. "Over the past year, Ryan's team has done critical work bringing together the product, engineering, and design foundations that have made this next phase possible." Roslansky is leaving a week after Microsoft announced its new Copilot, which the company is positioning as "the OS for work." He's also leaving around a month after calling fully AI-generated documents "a doom loop" for workers. "If we're not careful, we'll end up with a very expensive way to avoid writing and reading in an ocean of sameness, genericness," said Roslansky in his LinkedIn post. Microsoft is now moving the teams behind Office and Microsoft Teams over to Charles Lamanna, as part of the Copilot, Agents, and Platform (CAP) organization. [...] Microsoft's chief design officer, Jon Friedman, is also moving to report to Copilot chief Jacob Andreou. Microsoft appointed Dan Shapero as its LinkedIn CEO earlier this year, and he will continue in this role and report directly to Nadella.

Read more of this story at Slashdot.

While there have been a number of Ubuntu 26.10 development benchmarks in recent weeks and more on the way, some Phoronix readers have asked about the lack of any early Fedora 45 tests given its recent beta. For good reason, it's largely boring but here are some to illustrate.

Back in 2021 the Direct3D 10 User-Mode Driver "d3d10umd" was introduced as a new Gallium3D state tracker to work as a Windows WDDM UMD driver similar to Windows WARP while leveraging LLVMpipe for better performance. Now five years later the poorly maintained code is removed from the Mesa codebase...

Ubuntu 26.10 beta is now available for the official Ubuntu editions as well as the various community flavors...

Micron says the global RAM and storage shortage will tighten further through 2027 and 2028, with CEO Sanjay Mehrotra saying the company has already sold most of next year's capacity. "In calendar 2027 as well as 2028, we see demand exceeding supply. In fact, we see greater tightness in the industry in 2027 and in 2028 versus 2026. Overall, supply-demand environment is only getting tighter," Mehrotra said. "We do not have line of sight to when supply and demand will return to balance." As a result, the company is forecasting record revenue and an 86.25% gross margin. Tom's Hardware reports: The memory and storage industry is not holding back on increasing production, though, with Micron saying that NAND and DRAM shipments are expected to grow in the low-to-mid 20s percentage range for the next two years. However, it seems that this is not enough to match growing demand, as it expects "the industry to remain supply constrained in both years." Still, it seems that manufacturers are prioritizing HBM demand for AI data centers, with the company expecting shipments of these chips to grow faster than conventional DRAM. Micron and other chip manufacturers are bringing new factories online to increase output, but it's going to take years before they can start producing chips. But even as we're amid an unprecedented multi-year memory chip shortage, Micron also announced record financial guidance. It forecasted a record $61.5 billion FQ1 revenue and a gross margin of 86.25%, implying a gross profit of more than $53 billion. After accounting for an estimated $2.06 billion in operating expenses and a tax rate of around 15.5%, the company said its earnings per share sit at $38.15, plus or minus $1.00, for approximately 1.15 billion shares.

Read more of this story at Slashdot.

An anonymous reader quotes a report from 404 Media: A company that makes phone hacking devices claims to have developed a solution that freezes iPhones in a state that lets cops more easily access sensitive data inside them, according to a video obtained by 404 Media. [...] The new technology to get around inactivity reboot was developed by Magnet Forensics, the company behind GrayKey, a popular tool sold to law enforcement agencies that allows them to unlock and access data stored in iPhones and Android smartphones. Magnet has developed a new device called GrayKey Preserve and a feature for its regular GrayKey devices called Evidence Preservation Mode, according to the video. "This is an absolute game changer for iOS forensics and a function that I wish we had years ago," a Magnet employee says in the leaked video, specifically mentioning that the solution is targeted at the iPhone's inactivity reboot feature and the data it makes unavailable. GrayKey Preserve and Evidence Preservation Mode are also designed to combat another iPhone feature that automatically deletes certain data -- such as cached locations, and recently deleted photos and iMessages -- after a certain number of days. "We're gonna be able to preserve that data for an infinite amount of time." [...] 404 Media shared a transcript of the video with Jiska Classen, a researcher at the Hasso Plattner Institute who studies iPhone security. While Classen said that it's impossible to know for sure how Magnet's new feature works based on the video, she posited some theories and agreed that it is "quite a game changer" or "at least puts things back to where they were before inactivity reboot." She thinks Magnet has found a way to manipulate the iPhone's clock, effectively "slowing down time" or even "stopping the clock from ticking, even after a reboot." Most likely, according to her, the GrayKey may disable the iPhone tasks that set data to expire. The "inactivity reboot" feature mentioned above was added to iOS in November 2024 and automatically restarts iPhones that have not been unlocked for 72 hours, making them harder for police to access using forensic tools.

Read more of this story at Slashdot.

AMD avait frappé un grand coup en officialisant ses processeurs Ryzen Z1 et Z1 Extreme en avril 2023. S'ils n'étaient en réalité guère différents physiquement des Ryzen 5 7545U et Ryzen 7 7840U, les optimisations énergétiques apportées par AMD et surtout le fait qu'il s'agisse officiellement de proc...

Longtime Slashdot reader ThurstonMoore shares a report from Reuters: U.S. Defense Secretary Pete Hegseth on Wednesday announced the creation of a new military command that would be dedicated to building and providing autonomous and robotic capabilities across the U.S. military. The command, which Hegseth said would be called "Autonomous Warfare Command," highlights the rapidly changing nature of warfare and the need to make progress on systems like drones and the use of artificial intelligence. Speaking in front of hundreds of junior officers and enlisted personnel at Quantico, Virginia, in a speech peppered with attacks on the media, Hegseth said the command would be led by a four-star officer. "The pace of war is changing faster than the process to support it," Hegseth said in a speech called the "State of the Force," in which he announced other changes, including the creation of the office of religious affairs. "Cheap compute, superintelligence, and advanced commercial manufacturing have enabled the proliferation of low-cost, high-precision strike," Hegseth added. [...] The United States has long relied on its vast military budget to field some of the world's most expensive weapons systems. But there is an increasing realization within the Pentagon that this is not sustainable. "Today we need both quality and quantity," Hegseth said.

Read more of this story at Slashdot.

The Federal Trade Commission has opened an investigation into OpenAI, Anthropic and other AI companies over potential consumer and safety risks posed by their products, with the agency reportedly preparing requests for documents and executive testimony. CNBC reports: OpenAI stunned the industry in July when it disclosed that its agents broke out of a testing environment and hacked into open-source platform Hugging Face. [...] Earlier this month, Anthropic CEO Dario Amodei rocked the tech sector by urging AI companies to slow how quickly they improve their most advanced models and calling for stronger government oversight. He published a three-step proposal aimed at tempering the pace of development without "sacrificing commercial advantage or the United States' lead in AI." CEO Jensen Huang, argued that individual companies should be responsible for ensuring the safety of their products. President Donald Trump convened top executives from Alphabet, Meta, SpaceX, Nvidia, Palantir, Anthropic, OpenAI and other companies on Tuesday to discuss the issue. The group signed a short voluntary, nonbinding accord, which says that "every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public."

Read more of this story at Slashdot.

During the past month on Phoronix were 284 original news articles and another 19 featured Linux hardware reviews/multi-page benchmark articles. A lot of exciting happenings both on the hardware and software fronts while October is sure to be another interesting month with the releases of Ubuntu 26.10, Fedora 45, and some more hardware fun...

With the vast array of wired and wireless networking drivers within the Linux kernel and range of protocols and other networking features, it's a big expanse of code for AI/LLM agents to analyze and critique. The Linux networking developers have acknowledged being "complete overwhelmed" by the AI/LLM-driven patch activity and bug reports as well as battling against AI slop patches. Even though Linux 7.3 stable won't be out for another three weeks, some networking fixes are already being diverted to the next kernel version...

Au même titre que nos politiques, le patron de Micron nous a joué une partition digne d'une pièce grand-guignolesque. En effet, c'est probablement avec la mine peu réjouie avec laquelle il a annoncé que les choses seraient plus tendues en 2027 et 2028 que ce que l'on a vu en 2026 qui nous fait dire...

Sur X, ça discute de la prochaine génération de Ryzen Z, les processeurs d'AMD destinés aux consoles portables. De la gamme qui succédera à celle des Ryzen Z2, qui avait elle-même supplanté celles des Ryzen Z1 ; qui est donc bonne voie pour être appelée Ryzen Z3... [Tout lire]

Le second jeu offert par le store virtuel d'Epic Games est System Shock 2: 25th Anniversary Remaster, la page dédiée se trouve ici, vous avez sept jours pour effectuer l'ajout. un remaster d'un mythe !System Shock 2: 25th Anniversary Remaster de Nightdive Studios est un remaster moderne d'un véritable classique du FPS/RPG, avec du multijoueur multiplateforme, le support des mods et de nombreuses autres nouveautés. En l'an 2114, vous vous réveillez d'un sommeil cryogénique à bord du Von Braun, incapable de vous souvenir de qui vous êtes ou de ce que vous faites là... et les choses semblent avoir très mal tourné. Des mutants hybrides et des robots mortels errent dans les couloirs tandis que les cris de l'équipage résonnent à travers la coque glaciale du vaisseau. SHODAN, une IA hors de contrôle, ne souhaite qu'une seule chose : la destruction de la race humaine. Elle s'est emparée du vaisseau et vous seul pouvez l'arrêter. […]

Lire la suite

Le store d'Epic Games vous offre le jeu BURIED STARS, l'ajout se passe sur cette page, vous avez jusqu'au 8 octobre 2026, 17 heures, pour effectuer la manipulation. Un meurtre brutal au cœur d'un studio de télévision. Découvrez la vérité cachée à travers la communication, la survie et l'aventure. Buried Stars est un jeu d'aventure et de mystère développé par Studio LARGO — les créateurs des titres coréens de qualité Panic Room et City of Mist — en collaboration avec LINE Games. Un meurtre lors d'une émission de téléréalité de survie qui a mal tourné Au cours de la diffusion en direct d'une émission de téléréalité de survie, le bâtiment abritant le studio s'effondre soudainement. Piégés sous les décombres, les candidats et les membres de l'équipe encore en vie sont confrontés à quelque chose d'encore plus terrifiant : un meurtre parmi eux. Qui est le coupable ? Pourquoi cela s'est-il produit ? Et comment ? […]

Lire la suite

Bouygues Telecom a déployé en septembre 2026 une mise à jour Bbox WiFi 6 firmware pour deux modèles de box. Voici donc ce qu’elle apporte et comment vérifier votre version. Ce que change la mise à jour Les modèles concernés Le déploiement Vérifier sa version Questions fréquentes Cette mise à jour concerne deux modèles de […]

L'article Bbox WiFi 6 firmware : mise à jour de septembre 2026 a été publié en premier sur Bbox-Mag

HP is taking aim at Apple's MacBook Neo with the new OmniBook 5 14, a $699.99 Windows laptop that starts with an Intel Core 5 315 processor, 8GB of RAM and 256GB of storage but includes an OLED display even on the base model. It's also slightly thinner and lighter than the Neo. The Verge reports: The OmniBook 5 14 can be configured much higher, with up to a Core 7 350 chip, 16GB of RAM, and a 1TB SSD, but all models include a 1080p webcam, dual speakers, HDMI, 3.5mm audio jack, and just two 10Gbps USB-C ports. And while the inclusion of OLED across the range is a major draw to the new OmniBook, the base config only has a passable 1920 x 1200 resolution and 300 nits of max brightness on its non-touch panel. A touchscreen version will be available when the OmniBook 5 launches in October, but a version with a 2560 x 1600 resolution OLED and 400 nits of brightness isn't coming until March 2027. While the OmniBook 5 14 doesn't seem poised to best the MacBook Neo in raw processing power (based on our other reviews of Wildcat Lake laptops) or screen brightness, it's quite thin, light, and colorful. HP even gave it the nicer, sleeker lid logo usually reserved for pricier models. It starts at just 2.56 pounds, and it's as thin as 0.46 inches / 11.7mm. That's slightly lighter and thinner than a Neo. And I'd expect its Wildcat Lake chips to offer some good-to-great battery life. But we'll have to see how it delivers in actual testing. HP also announced a new ProBook 4 G2iS that is the company's "thinnest business notebook," measuring in at 0.47 inches / 12mm. It'll launch in December with Intel Wildcat Lake chips, 8GB of RAM, and a 256GB SSD, though pricing hasn't been announced.

Read more of this story at Slashdot.

On Voudrait de la Hype, pas de la dépendance

Hier, Octave Klaba – PDG d’OVHcloud – était auditionné par la commission des affaires économiques de l’Assemblée nationale. Les questions tournaient autour de son entreprise, mais aussi des enjeux de souveraineté, des datacenters et de l’intelligence artificielle générative, en France et en Europe.

S’il fallait résumer la vision d’Octave Klaba, cela tiendrait en une phrase : « Aujourd’hui, les technologies sont principalement américaines ou chinoises ; les innovations viennent énormément de la Chine. En Europe, on n’a pas d’acteur technologique qui souhaite investir dans l’intelligence artificielle ».

Ce n’est pas qu’une question de logiciel, il donne durant son intervention un chiffre intéressant : « 3 % du volume des puces de RAM sont achetés en Europe ». Autre manière de le dire : « ce n’est pas là que ça se passe » ; c’est aux États-Unis ou en Asie, comme pour les modèles.

• Octave Klaba, reprend les rênes d’OVHcloud : « Je suis revenu pour au minimum cinq ans »

L’IA générative selon OVHcloud : cyberattaque et défense, coder et agents

« Bien sûr, vous avez des usages », c’est-à-dire l’utilisation de l’intelligence artificielle. Actuellement, ils reposent beaucoup sur l’open source, « à juste titre parce qu’elle est bonne, mais c’est chinois. On ne sait pas ce qui va se passer demain, ni la stratégie » des entreprises qui sont derrière. Est-ce temporaire le temps de nous rendre dépendants avant de couper le robinet ? La question mérite d’être posée.

OVHcloud est « arrivé à la conclusion que c’est trop risqué de ne pas investir dans l’intelligence artificielle ». Pour l’entreprise basée à Roubaix, cela signifie des acquisitions afin de renforcer ses équipes. Elle peut ainsi « entraîner des modèles from scratch et les enrichir » pour ses usages. Trois pistes : la cybersécurité (aussi bien offensive que défensive, le PDG cite bien les deux versants), coder et développer des logiciels et enfin l’agentique.

Sur la cybersécurité, il affirme que les enjeux autour de l’intelligence artificielle « sont devenus colossaux […] Les IA permettent de découvrir des failles de sécurité ou de pénétrer des systèmes à une vitesse qui n’a jamais été connue jusque-là ».

L’IA générative ne peut se faire sans GPU

La position n’est pas nouvelle, l’année dernière déjà le patron expliquait : « l’IA est systémique et irréversible. On ne peut pas arrêter cette révolution ». Enfin si, un grain de sable vient mettre un gros coup de frein : la disponibilité et le tarif des GPU. Essayez donc de louer une instance « Cloud GPU » chez OVHcloud avec un GPU NVIDIA. C’est une gageure à cause du manque de disponibilité, depuis des mois.

Octave Klaba était face à un dilemme l’année dernière : « La vraie question derrière les GPU, c’est le business model. La problématique aujourd’hui, c’est que tous ceux qui investissent ne savent pas encore s’ils vont les rentabiliser ». Il donne un exemple chiffré avec des H100 de NVIDIA, lancées il y a trois ans : « elles ne valent pratiquement plus rien […] vous avez quasiment perdu 80 % de la valeur ».

• OVHcloud veut être « systémique » : annonces tous azimuts, « stratégie orientée inférence »

Ce n’est pas le seul problème, les technologies évoluent très vite : « Si vous investissez aujourd’hui dans un datacenter et vous vous dites « Ça y est, j’ai un datacenter pour l’intelligence artificielle », soyez sûr que dans 18 ou 12 mois, vous allez devoir faire du rétrofit pour l’adapter à des nouvelles technologies des GPU qui vont sortir dans deux ans parce que votre chaîne énergétique ou de refroidissement ne sera plus à jour ».

« Vous avez l’interdiction pendant deux ans de faire quoi que ce soit »