Coincidentally after writing this morning wondering about what's going on with AerynOS (formerly known as Serpent OS), contributors "NomadicCore" and "Ermo" have issued a new project update outlining what's been going on the past few months...

Thousands of Norwegians briefly believed they had won massive Eurojackpot prizes after a manual coding error by Norsk Tipping mistakenly multiplied winnings by 100 instead of dividing. The Register reports: Eurojackpot, a pan-European lottery launched in 2012, holds two draws per week, and its jackpots start at about $12 million with a rollover cap of $141 million. Norsk Tipping, Norway's Eurojackpot administrator, admitted on Friday that a "manual error" it its conversion process from Eurocents to Norwegian kroner multiplied amounts by 100 instead of dividing them. As a result, "thousands" of players were briefly shown jackpots far higher than their actual winnings before the mistake was caught, but no incorrect payouts were made. Norsk Tipping didn't disclose how large the false jackpots were, but math suggests the improper amounts were 10,000x times higher. Regardless, it seems like a lot of people thought they were big winners, based on what the company's now-former CEO, Tonje Sagstuen, said on Saturday. "I have received many messages from people who had managed to make plans for holidays, buying an apartment or renovating before they realized that the amount was wrong," Sagstuen said in a statement. "To them I can only say: Sorry!" The incorrect prize amounts were visible on the Norsk Tipping website only briefly on Friday, but the CEO still resigned over the weekend following the incident. While one of the Norsk Tipping press releases regarding the incident described it as "not a technical error," it still appears someone fat-fingered a bit of data entry. The company said it will nonetheless be investigating how such a mistake could have happened "to prevent something similar from happening again."

Read more of this story at Slashdot.

While Q2 is drawing to an end in the coming hours, Intel software engineers this evening just released the Intel FFmpeg Cartwheel 2025Q1 update that provides all of their latest patches around Intel GPU/video acceleration for this widely-used, open-source multimedia library that have yet to be upstreamed into FFmpeg proper...

An anonymous reader quotes a report from Tom's Hardware: Microsoft EVP Yusuf Mehdi said in a blog post last week that Windows powers over a billion active devices globally. This might sound like a healthy number, but according to ZDNET, the Microsoft annual report for 2022 said that more than 1.4 billion devices were running Windows 10 or 11. Given that these documents contain material information and have allegedly been pored over by the tech giant's lawyers, we can safely assume that Windows' user base has been quietly shrinking in the past three years, shedding around 400 million users. This is probably why Microsoft has been aggressively pushing users to upgrade to Windows 11 after the previous version of the OS loses support -- so that its users would install the latest version of Windows on their current system (or get a new PC if their system is incapable of running the latest version). Although macOS is a threat to Windows, especially with the launch of Apple Silicon, we cannot say that those 400 million users all went and bought a MacBook. That's because, as far back as 2023, Mac sales have also been dropping, with Statista reporting the computer line, once holding more than 85% of the company revenue, now making up just 7.7%. The shrinking Windows user base can be attributed to a combination of factors -- a major one being the global move toward a mobile-first world, where smartphones and tablets are increasingly replacing traditional PCs for everyday computing needs. At the same time, Microsoft's strict hardware requirements for Windows 11 have alienated users with perfectly functional older machines, prompting some to stick with unsupported versions or abandon Windows entirely. Additionally, many users find Windows 11 less intuitive than its predecessor and are frustrated by Microsoft's push toward data collection and Apple-style design changes.

Read more of this story at Slashdot.

Oracle has signed a landmark $30 billion annual cloud deal -- nearly triple the size of its current cloud infrastructure business -- with revenue expected to begin in fiscal year 2028. The deal was disclosed in a regulatory filing Monday without the customer being named. Bloomberg reports: "Oracle is off to a strong start" in its fiscal year 2026, Chief Executive Officer Safra Catz said in the filing. The company has signed "multiple large cloud services agreements," she said, adding that revenue from Oracle's namesake database that runs on other clouds continues to grow more than 100%. The $30-billion deal ranks among the largest cloud contracts on record. That revenue alone would represent nearly three times the size of Oracle's current infrastructure business, which totaled $10.3 billion over the past four quarters. A major cloud contract awarded in 2022 from the US Defense Department, that runs through 2028 and could be worth as much as $9 billion, is split among four companies, including Oracle. That award was a shift after an earlier contract worth $10 billion was awarded to Microsoft and was contested in court.

Read more of this story at Slashdot.

Automattic has put its plan to migrate Tumblr's backend to WordPress on hold, with CEO Matt Mullenweg citing a shift in focus toward features users are actively requesting. "I still want to do it," Mullenweg says. "It's just cleaner. But right now, we're not working on it." The Verge reports: The decision to halt the change also appears to mean that Tumblr posts won't be available in the fediverse in the near future. WordPress.com currently offers an ActivityPub plug-in, so Tumblr moving onto WordPress would theoretically let people bring Tumblr posts to the fediverse. "That would've been a free way to get it," Mullenweg says. "And so that was one of the arguments for migrating everything to WordPress." In the meantime, however, "I think if there was a big push to implement fediverse, we would just do it on the Tumblr code base," according to Mullenweg.

Read more of this story at Slashdot.

HP is developing a "PCFax" system similar to CarFax for used cars that securely collects and stores detailed device usage and health data at the firmware level to extend the life of PCs and reduce e-waste. A team of HP experts outlines the system in a recent IEEE Spectrum report: The secure telemetry protocol we've developed at HP works as follows. We gather the critical hardware and sensor data and store it in a designated area of the SSD. This area is write-locked, meaning only authorized firmware components can write to it, preventing accidental modification or tampering. That authorized firmware component we us is the Endpoint Security Controller, a dedicated piece of hardware embedded in business class HP PCs. It plays a critical role in strengthening platform-level security and works independently from the main CPU to provide foundational protection. The endpoint security controller establishes a secure session by retaining the secret key within the controller itself. This mechanism enables read data protection on the SSD -- where telemetry and sensitive data are stored -- by preventing unauthorized access, even if the operating system is reinstalled or the system environment is otherwise altered. Then, the collected data is recorded in a timestamped file, stored within a dedicated telemetry log on the SSD. Storing these records on the SSD has the benefit of ensuring the data is persistent even if the operating system is reinstalled or some other drastic change in software environment occurs. The telemetry log employs a cyclic buffer design, automatically overwriting older entries when the log reaches full capacity. Then, the telemetry log can be accessed by authorized applications at the operating system level. The telemetry log serves as the foundation for a comprehensive device history report. Much like a CarFax report for used cars, this report, which we call PCFax, will provide both current users and potential buyers with crucial information. The PCFax report aggregates data from multiple sources beyond just the on-device telemetry logs. It combines the secure firmware-level usage data with information from HP's factory and supply chain records, digital services platforms, customer support service records, diagnostic logs, and more. Additionally, the system can integrate data from external sources including partner sales and service records, refurbishment partner databases, third-party component manufacturers like Intel, and other original equipment manufacturers. This multi-source approach creates a complete picture of the device's entire lifecycle, from manufacturing through all subsequent ownership and service events.

Read more of this story at Slashdot.

Freelance workers using AI tools are earning significantly more than their counterparts, with AI-related freelance earnings climbing 25% year over year and AI freelancers commanding over 40% higher hourly rates than non-AI workers, according to new data from Upwork. The freelance marketplace analyzed over 130 work categories and tracked millions of job posts over six months, finding that generative AI is simultaneously replacing low-complexity, repetitive tasks while creating demand for AI-augmented work. Workers using AI for augmentation outnumber those using it for automation by more than 2 to 1. Freelancers with coding skills comprising at least 25% of their work now earn 11% more for identical jobs compared to November 2022 when ChatGPT launched.

Read more of this story at Slashdot.

Apple must face the U.S. Department of Justice's lawsuit accusing the iPhone maker of unlawfully dominating the U.S. smartphone market, a judge ruled on Monday. From a report: U.S. District Judge Julien Neals in Newark, New Jersey, denied Apple's motion to dismiss the lawsuit accusing the company of using restrictions on third-party app and device developers to keep users from switching to competitors and unlawfully dominate the market. The decision would allow the case to go forward in what could be a years-long fight for Apple against enforcers' attempt to lower what they say are barriers to competition with Apple's iPhone.

Read more of this story at Slashdot.

An anonymous reader shares a report: Sen. Ted Cruz (R-Texas) has a plan for spectrum auctions that could take frequencies away from Wi-Fi and reallocate them for the exclusive use of wireless carriers. The plan would benefit AT&T, which is based in Cruz's home state, along with Verizon and T-Mobile. Cruz's proposal revives a years-old controversy over whether the entire 6 GHz band should be devoted to Wi-Fi, which can use the large spectrum band for faster speeds than networks that rely solely on the 2.4 and 5 GHz bands. Congress is on the verge of passing legislation that would require spectrum to be auctioned off for full-power, commercially licensed use, and the question is where that spectrum will come from. When the House of Representatives passed its so-called "One Big Beautiful Bill," it excluded all of the frequencies between 5.925 and 7.125 gigahertz from the planned spectrum auctions. But Cruz's version of the budget reconciliation bill, which is moving quickly toward a final vote, removed the 6 GHz band's protection from spectrum auctions. The Cruz bill is also controversial because it would penalize states that regulate artificial intelligence. Instead of excluding the 6 GHz band from auctions, Cruz's bill would instead exclude the 7.4-8.4 GHz band used by the military. Under conditions set by the bill, it could be hard for the Commerce Department and Federal Communications Commission to fulfill the Congressional mandate without taking some spectrum away from Wi-Fi.

Read more of this story at Slashdot.

Apple is considering using AI technology from Anthropic or OpenAI to power a new version of Siri, according to Bloomberg, sidelining its own in-house models in a potentially blockbuster move aimed at turning around its flailing AI effort. From the report: The iPhone maker has talked with both companies about using their large language models for Siri, according to people familiar with the discussions. It has asked them to train versions of their models that could run on Apple's cloud infrastructure for testing, said the people, who asked not to be identified discussing private deliberations. If Apple ultimately moves forward, it would represent a monumental reversal. The company currently powers most of its AI features with homegrown technology that it calls Apple Foundation Models and had been planning a new version of its voice assistant that runs on that technology for 2026. A switch to Anthropic's Claude or OpenAI's ChatGPT models for Siri would be an acknowledgment that the company is struggling to compete in generative AI -- the most important new technology in decades. Apple already allows ChatGPT to answer web-based search queries in Siri, but the assistant itself is powered by Apple.

Read more of this story at Slashdot.

NUC 14 : Pro, AI, Lunar Lake, mais pas lunaires En 2023, nous avions jaugé du NUC 13 Pro Desk Edition, dernier modèle commercialisé par Intel — et peut être le plus beau dans sa robe immaculée. Deux ans plus tard, nous remettons le couvert, cette fois avec le NUC 14 Pro AI+ et le NUC 14 Pro AI tout court — avant, très prochainement, le NUC 15 Pro Plus, qui aura droit à son propre test... [Tout lire]

TorrentFreak spotlights VP.net, a brand-new service from Private Internet Access founder Andrew Lee (the guy who gifted Linux Journal to Slashdot) that eliminates the classic "just trust your VPN" problem by locking identity-mapping and traffic-handling inside Intel SGX enclaves. The company promises 'cryptographically verifiable privacy' by using special hardware 'safes' (Intel SGX), so even the provider can't track what its users are up to. The design goal is that no one, not even the VPN company, can link "User X" to "Website Y." Lee frames it as enabling agency over one's privacy: "Our zero trust solution does not require you to trust us - and that's how it should be. Your privacy should be up to your choice - not up to some random VPN provider in some random foreign country." The team behind VP.net includes CEO Matt Kim as well as arguably the first Bitcoin veterans Roger Ver and Mark Karpeles. Ask Slashdot: Now that there's a VPN where you don't have to "just trust the provider" - arguably the first real zero-trust VPN - are trust based VPNs obsolete?

Read more of this story at Slashdot.

Automattic CEO Matt Mullenweg said he regrets telling the media that "WordPress.org just belongs to me personally" during a new interview about his company's legal dispute with hosting provider WP Engine. The comment has been "taken out of context so many times" and represents "the worst thing ever," Mullenweg said in a new podcast interview with The Verge. The dispute began when Mullenweg accused WP Engine of "free-riding" on WordPress's open-source ecosystem without contributing adequate resources back to the project. Mullenweg filed a lawsuit against WP Engine while cutting off the company's access to core WordPress technologies. WP Engine countersued, and Automattic was forced to reverse some retaliatory measures. The controversy triggered significant internal upheaval at Automattic. The company offered "alignment" buyouts to employees who disagreed with the direction, reducing headcount from a peak of 2,100 to approximately 1,500 people. Mullenweg said this was "probably the fourth big time" WordPress has faced such community controversy, though the first in the current media landscape. WordPress powers 43% of websites globally. Mullenweg said he wants to return to "the most collaborative version of WordPress possible" but noted the legal proceedings continue with both sides spending "millions of dollars a month on lawyers."

Read more of this story at Slashdot.

China's transition to digital payments has reached the point where physical cash has nearly vanished from daily commerce, with WeChat and Alipay now handling transactions from supermarkets to public transportation across the world's second-largest economy. Many businesses no longer maintain traditional cash registers and instead scan QR codes presented by customers, while numerous taxis refuse cash payments entirely. The widespread adoption has given tech giants Tencent and Alibaba immense power over routine financial transactions, prompting China's central bank to develop a competing digital yuan currency.

Read more of this story at Slashdot.

Microsoft's new AI diagnostic system achieved 80% accuracy in diagnosing patients compared to 20% for human doctors, while reducing costs by 20%, according to company research published Monday. The MAI Diagnostic Orchestrator queries multiple leading AI models including OpenAI's GPT, Google's Gemini, Anthropic's Claude, Meta's Llama, and xAI's Grok in what the company describes as a "chain-of-debate style" approach. The system was tested against 304 case studies from the New England Journal of Medicine using Microsoft's Sequential Diagnosis Benchmark, which breaks down each case into step-by-step diagnostic processes that mirror how human physicians work. Microsoft CEO of AI Mustafa Suleyman called the development "a genuine step toward medical superintelligence."

Read more of this story at Slashdot.

Avantare writes: Microsoft Authenticator houses your passwords and lets you sign into all of your Microsoft accounts using a PIN, facial recognition such as Windows Hello, or other biometric data, like a fingerprint. Authenticator can be used in other ways, such as verifying you're logging in if you forgot your password, or using two-factor authentication as an extra layer of security for your Microsoft accounts. In June, Microsoft stopped letting users add passwords to Authenticator, but here's a timeline of other changes you can expect, according to Microsoft: July 2025: You won't be able to use the autofill password function. August 2025: You'll no longer be able to use saved passwords.

Read more of this story at Slashdot.

Big Bro... AI

La vision par ordinateur, domaine de R&D super cool où les chercheurs et les ingénieurs aident à construire des voitures autonomes et à créer des moteurs de recherche d’images ? Des scientifiques montrent que la plupart des travaux se focalisent sur des thématiques de surveillance.

La vision par ordinateur (en anglais, computer vision) est souvent présentée aux étudiants comme un domaine fun et qui peut rendre service dans la vie de tous les jours. Par exemple, la présentation du livre (reconnu dans le milieu) « Computer Vision: Algorithms and Applications » de Richard Szeliski affirme qu’il décrit « des applications concrètes stimulantes où la vision est utilisée avec succès, à la fois dans des applications spécialisées telles que la recherche d’images et la navigation autonome, mais aussi pour des tâches amusantes et grand public que les étudiants peuvent appliquer à leurs propres photos et vidéos ».

Pourtant, l’actualité montre qu’elle est fréquemment utilisée pour la surveillance, notamment couplée à des caméras.

• Vidéosurveillance algorithmique : le gouvernement insiste pour prolonger jusqu’en 2027
• Toulouse s’équipe en vidéosurveillance algorithmique alors qu’elle est encore interdite

Une équipe de cinq chercheurs et chercheuses en informatique montre justement, dans un article publié dans la revue Nature, que la plupart des travaux dans ce domaine servent à mettre en œuvre des technologies de surveillance.

Des racines historiques dans la surveillance

« La vision par ordinateur fait référence à l’IA qui se concentre sur la mesure, l’enregistrement, la représentation et l’analyse du monde à partir de données visuelles telles que des images et vidéos », expliquent-ils. Ces chercheurs et chercheuses affirment qu’un flou est entretenu dans le domaine sur les réelles applications des recherches mises en place. Robotique, véhicules autonomes mais aussi design de nouvelles protéines, arts créatifs et modélisation sur le changement climatique sont souvent mis en avant.

Mais ils rappellent que les « racines historiques » du domaine se trouvent « dans la surveillance militaire et carcérale, où elle a été développée pour identifier des cibles et recueillir des renseignements dans des contextes de guerre, d’application de la loi et d’immigration ». Ils s’appuient notamment sur le livre « The Birth of Computer Vision » publié récemment par James E. Dobso.

Avant tout de l’analyse de données biométriques humaines

Mais où en est-on ? En analysant plus de 19 000 articles scientifiques du domaine et 23 000 brevets qui découlent de la recherche publiée entre 1990 et 2021, ils montrent qu’une part très importante est utilisée pour la surveillance.

Ainsi, la plupart des données analysées dans les articles de recherche font référence à des corps humains dans leur ensemble (37 %) ou à des parties du corps (31 %, notamment de l’analyse faciale). Et 17 % autres données se réfèrent à des informations sur des espaces utilisés par les humains (comme des maisons, des bureaux, des rues.

Sur le suivi du corps humain au sens large, les articles scientifiques et brevets ciblent « souvent des personnes en pleine activité quotidienne (par exemple, en marchant, en faisant des achats ou en participant à des événements de groupe), et les objectifs cités comprenaient la détection, le suivi et le comptage des corps, ainsi que la surveillance de la sécurité et la reconnaissance de l’activité humaine », expliquent les chercheurs.

En eux-mêmes, ces chiffres ne sont pas des preuves d’une utilisation massive de ce champ pour de la surveillance des humains. En effet, on peut argumenter, par exemple, que, dans le cas du véhicule autonome, la détection de piéton est importante. Mais l’est-elle beaucoup plus que la détection d’un autre véhicule ou d’un animal ?

Une augmentation de la proportion des brevets dans cette thématique

Mais l’analyse des brevets amène d’autres chiffres. On y voit par exemple une augmentation significative de la proportion des brevets permettant la surveillance depuis 1990. Mais aussi un changement de vocabulaire. Ainsi, les termes « identification » et « person » sont beaucoup plus utilisés dans les titres des articles publiés dans les années 2010 que dans les années 1990.

Dans leur article, les chercheurs expliquent qu’il y a une normalisation de la surveillance par IA dans le domaine et que celle-ci se fait notamment par l’influence des différents financeurs de la recherche.

En premier lieu, on voit que Microsoft est l’entité qui dépose le plus de brevets découlant de recherche dans le domaine. Et les trois quarts concernent des applications de surveillance. On retrouve un peu près la même proportion pour les dix plus importantes institutions qui financent le domaine.

Sur la répartition géographique, on peut constater que les institutions étasuniennes (entreprises comme universités ou état) sont les principaux financeurs de ce genre de brevets. La Chine apparait en deuxième position.

Big Bro... AI

La vision par ordinateur, domaine de R&D super cool où les chercheurs et les ingénieurs aident à construire des voitures autonomes et à créer des moteurs de recherche d’images ? Des scientifiques montrent que la plupart des travaux se focalisent sur des thématiques de surveillance.

La vision par ordinateur (en anglais, computer vision) est souvent présentée aux étudiants comme un domaine fun et qui peut rendre service dans la vie de tous les jours. Par exemple, la présentation du livre (reconnu dans le milieu) « Computer Vision: Algorithms and Applications » de Richard Szeliski affirme qu’il décrit « des applications concrètes stimulantes où la vision est utilisée avec succès, à la fois dans des applications spécialisées telles que la recherche d’images et la navigation autonome, mais aussi pour des tâches amusantes et grand public que les étudiants peuvent appliquer à leurs propres photos et vidéos ».

Pourtant, l’actualité montre qu’elle est fréquemment utilisée pour la surveillance, notamment couplée à des caméras.

• Vidéosurveillance algorithmique : le gouvernement insiste pour prolonger jusqu’en 2027
• Toulouse s’équipe en vidéosurveillance algorithmique alors qu’elle est encore interdite

Une équipe de cinq chercheurs et chercheuses en informatique montre justement, dans un article publié dans la revue Nature, que la plupart des travaux dans ce domaine servent à mettre en œuvre des technologies de surveillance.

Des racines historiques dans la surveillance

« La vision par ordinateur fait référence à l’IA qui se concentre sur la mesure, l’enregistrement, la représentation et l’analyse du monde à partir de données visuelles telles que des images et vidéos », expliquent-ils. Ces chercheurs et chercheuses affirment qu’un flou est entretenu dans le domaine sur les réelles applications des recherches mises en place. Robotique, véhicules autonomes mais aussi design de nouvelles protéines, arts créatifs et modélisation sur le changement climatique sont souvent mis en avant.

Mais ils rappellent que les « racines historiques » du domaine se trouvent « dans la surveillance militaire et carcérale, où elle a été développée pour identifier des cibles et recueillir des renseignements dans des contextes de guerre, d’application de la loi et d’immigration ». Ils s’appuient notamment sur le livre « The Birth of Computer Vision » publié récemment par James E. Dobso.

Avant tout de l’analyse de données biométriques humaines

Mais où en est-on ? En analysant plus de 19 000 articles scientifiques du domaine et 23 000 brevets qui découlent de la recherche publiée entre 1990 et 2021, ils montrent qu’une part très importante est utilisée pour la surveillance.

Ainsi, la plupart des données analysées dans les articles de recherche font référence à des corps humains dans leur ensemble (37 %) ou à des parties du corps (31 %, notamment de l’analyse faciale). Et 17 % autres données se réfèrent à des informations sur des espaces utilisés par les humains (comme des maisons, des bureaux, des rues.

Sur le suivi du corps humain au sens large, les articles scientifiques et brevets ciblent « souvent des personnes en pleine activité quotidienne (par exemple, en marchant, en faisant des achats ou en participant à des événements de groupe), et les objectifs cités comprenaient la détection, le suivi et le comptage des corps, ainsi que la surveillance de la sécurité et la reconnaissance de l’activité humaine », expliquent les chercheurs.

En eux-mêmes, ces chiffres ne sont pas des preuves d’une utilisation massive de ce champ pour de la surveillance des humains. En effet, on peut argumenter, par exemple, que, dans le cas du véhicule autonome, la détection de piéton est importante. Mais l’est-elle beaucoup plus que la détection d’un autre véhicule ou d’un animal ?

Une augmentation de la proportion des brevets dans cette thématique

Mais l’analyse des brevets amène d’autres chiffres. On y voit par exemple une augmentation significative de la proportion des brevets permettant la surveillance depuis 1990. Mais aussi un changement de vocabulaire. Ainsi, les termes « identification » et « person » sont beaucoup plus utilisés dans les titres des articles publiés dans les années 2010 que dans les années 1990.

Dans leur article, les chercheurs expliquent qu’il y a une normalisation de la surveillance par IA dans le domaine et que celle-ci se fait notamment par l’influence des différents financeurs de la recherche.

En premier lieu, on voit que Microsoft est l’entité qui dépose le plus de brevets découlant de recherche dans le domaine. Et les trois quarts concernent des applications de surveillance. On retrouve un peu près la même proportion pour les dix plus importantes institutions qui financent le domaine.

Sur la répartition géographique, on peut constater que les institutions étasuniennes (entreprises comme universités ou état) sont les principaux financeurs de ce genre de brevets. La Chine apparait en deuxième position.

Jon Seager as the VP of Engineering at Canonical today publicly announced the formation of their "Debcrafters" global team to help ensure the health of the Ubuntu Archive...