AMD a lancé un bundle alléchant comme on les aime. En effet, pour tout achat d'une RX 7700 XT ou d'une RX 7800 XT à partir du 11 juin et jusqu'au 20 juillet, vous pourrez récupérer via le site AMDReward deux jeux au choix parmi la liste suivante :• Avatar: Frontiers of Pandora (qui revient puisqu'il...

An anonymous reader quotes a report from the New York Times: The news was featured on MSN.com: "Prominent Irish broadcaster faces trial over alleged sexual misconduct." At the top of the story was a photo of Dave Fanning. But Mr. Fanning, an Irish D.J. and talk-show host famed for his discovery of the rock band U2, was not the broadcaster in question. "You wouldn't believe the amount of people who got in touch," said Mr. Fanning, who called the error "outrageous." The falsehood, visible for hours on the default homepage for anyone in Ireland who used Microsoft Edge as a browser, was the result of an artificial intelligence snafu. A fly-by-night journalism outlet called BNN Breaking had used an A.I. chatbot to paraphrase an article from another news site, according to a BNN employee. BNN added Mr. Fanning to the mix by including a photo of a "prominent Irish broadcaster." The story was then promoted by MSN, a web portal owned by Microsoft. The story was deleted from the internet a day later, but the damage to Mr. Fanning's reputation was not so easily undone, he said in a defamation lawsuit filed in Ireland against Microsoft and BNN Breaking. His is just one of many complaints against BNN, a site based in Hong Kong that published numerous falsehoods during its short time online as a result of what appeared to be generative A.I. errors. Mr. Fanning's complaint against BNN is one of many. The site based published numerous falsehoods during its short time online.Credit...Paulo Nunes dos Santos for The New York Times BNN went dormant in April, while The New York Times was reporting this article. The company and its founder did not respond to multiple requests for comment. Microsoft had no comment on MSN's featuring the misleading story with Mr. Fanning's photo or his defamation case, but the company said it had terminated its licensing agreement with BNN. During the two years that BNN was active, it had the veneer of a legitimate news service, claiming a worldwide roster of "seasoned" journalists and 10 million monthly visitors, surpassing the The Chicago Tribune's self-reported audience. Prominent news organizations like The Washington Post, Politico and The Guardian linked to BNN's stories. Google News often surfaced them, too. A closer look, however, would have revealed that individual journalists at BNN published lengthy stories as often as multiple times a minute, writing in generic prose familiar to anyone who has tinkered with the A.I. chatbot ChatGPT. BNN's "About Us" page featured an image of four children looking at a computer, some bearing the gnarled fingers that are a telltale sign of an A.I.-generated image. "How easily the site and its mistakes entered the ecosystem for legitimate news highlights a growing concern: A.I.-generated content is upending, and often poisoning, the online information supply," adds The Times. "NewsGuard, a company that monitors online misinformation, identified more than 800 websites that use A.I. to produce unreliable news content. The websites, which seem to operate with little to no human supervision, often have generic names -- such as iBusiness Day and Ireland Top News -- that are modeled after actual news outlets. They crank out material in more than a dozen languages, much of which is not clearly disclosed as being artificially generated, but could easily be mistaken as being created by human writers."

Read more of this story at Slashdot.

According to The Verge, Best Buy conducted another round of layoffs and job restructurings to "right size" the business in response to declining sales post-pandemic. Further layoffs and changes are expected throughout the year. From the report: The layoffs appeared to have mostly targeted in-home sales roles called designers, who would go to customers' homes to help identify products that would work in their space. It's not clear how many were let go, but designers who weren't laid off have been moved into a different, largely in-store role. Also, pay scales for a similar, existing in-store "consultant" position were revamped. Best Buy confirmed the layoffs in an email to The Verge but declined to share how many people were let go or how pay was changing. "Many of our team members were moved to new areas or roles where our customers need it most," Best Buy spokesperson Ryan Furlong told The Verge. He said some employees in Best Buy's "Design and Consult workforce" -- the collection of roles with in-store workers (called consultants) and in-home field sales positions (called designers) -- will be transitioned into a new "Premium Designer role." Best Buy has been drastically restructuring in recent months, responding to factors like falling sales after the pandemic spiked consumer electronics spending. Best Buy CEO Corie Barry told investors in February that they should expect layoffs this year, and two months ago, mass layoffs of Geek Squad employees were reported. Barry repeated similar things during the company's first quarter earnings call in May, saying that many of Best Buy's moves to "right size" its business "are being implemented throughout this year."

Read more of this story at Slashdot.

The extensible scheduler "sched_ext" code has proven quite versatile for opening up better Linux gaming performance, more quickly prototyping new scheduler changes, Ubuntu/Canonical has been evaluating it for pursuing a more micro-kernel like design, and many other interesting approaches with it. Yet it's remained out of tree but that is now changing with the upcoming Linux 6.11 cycle...

Brazil's government is partnering with OpenAI to use AI for expediting the screening and analysis of thousands of lawsuits to reduce costly court losses impacting the federal budget. Reuters reports: The AI service will flag to government the need to act on lawsuits before final decisions, mapping trends and potential action areas for the solicitor general's office (AGU). AGU told Reuters that Microsoft would provide the artificial intelligence services from ChatGPT creator OpenAI through its Azure cloud-computing platform. It did not say how much Brazil will pay for the services. AGU said the AI project would not replace the work of its members and employees. "It will help them gain efficiency and accuracy, with all activities fully supervised by humans," it said. Court-ordered debt payments have consumed a growing share of Brazil's federal budget. The government estimated it would spend 70.7 billion reais ($13.2 billion) next year on judicial decisions where it can no longer appeal. The figure does not include small-value claims, which historically amount to around 30 billion reais annually. The combined amount of over 100 billion reais represents a sharp increase from 37.3 billion reais in 2015. It is equivalent to about 1% of gross domestic product, or 15% more than the government expects to spend on unemployment insurance and wage bonuses to low-income workers next year. AGU did not provide a reason for Brazil's rising court costs.

Read more of this story at Slashdot.

New York has launched its mobile ID program, "giving residents the option to digitize their driver's license or non-driver ID," reports The Verge. From the report: Beginning today, the New York Mobile ID app is available from Apple's App Store and Google Play. The app can be used for identity verification at airports. A physical license, permit, or non-driver ID is required to activate a mobile ID; you'll need to take a photo of the front and back with your phone during the enrollment process. The news was announced during a media briefing at LaGuardia Airport on Tuesday that included New York's and Transportation Security Administration federal security director Robert Duffy, among other speakers. Their pitch is that mobile IDs "will revolutionize the way New Yorkers protect their identities and will significantly enhance the way they get through security at airports across the nation." State officials are also emphasizing that it's a voluntary option meant for convenience. "When you offer your mobile ID to TSA or anyone else who accepts it, you are in full control of sharing that information. They can only see the information they request to see," Schroeder said. "If you only need to prove your age, you can withhold other information that a verifier doesn't need to see." The app is designed so that your phone remains in your possession at all times -- you should never freely hand a device over to law enforcement -- and shows a QR code that can be scanned to verify your identity. Any changes to your license status such as renewals or suspensions are automatically pushed to the mobile version, and the digital ID also mirrors data like whether you're an organ donor. For now, acceptance of mobile IDs by businesses (and the police) is completely voluntary -- and there's no deadline in place for compliance -- so it's definitely too soon to start leaving your physical one at home. But bars and other small businesses can start accepting them immediately if they install the state's verifier app. The New York Mobile ID app can be used "at nearly 30 participating airports across the country including all terminals at LaGuardia and John F. Kennedy airports," according to a press release from Governor Kathy Hochul. New York joins a small list of states that have rolled out mobile driver's licenses, including Arizona, Colorado, Delaware, Georgia, Florida, Iowa, Louisiana, Maryland, Mississippi, Missouri, and Utah.

Read more of this story at Slashdot.

An anonymous reader quotes a report from the Mercury News: Krista DeWeese has been laid off four times in the last eight years. She wakes up every morning feeling anxious. Will I lose my job today -- again? Will I have enough to pay the rent? Even though she's an educated, experienced marketing professional, worrisome thoughts trail the 47-year-old Fremont native's every waking moment. Currently a contract worker at a health science company, she has been struggling to find secure work that pays enough to keep up with the exorbitant cost of living in the Bay Area. She has a lot of company. The past year has been tough for the Bay Area, as thousands of layoffs skittered across the region. Even workers at Silicon Valley's tech titans -- including Meta, Apple and Google -- have faced job cuts. Since 2022, tech companies in the region have slashed roughly 40,000 jobs. And with each layoff, workers are entering a market that is less friendly to job seekers than it used to be. New research from tech advocacy organization Women Impact Tech, which examined job and salary data nationwide from 2020 to 2023, affirmed what many people already know: companies are tightening their belts -- slicing jobs and salaries alike -- and many people are struggling to find work that pays enough to live comfortably in the Bay Area. Despite having the highest tech salaries in the country, Silicon Valley has experienced the biggest drop in pay compared to other tech hubs, falling 15% from 2022 to 2023, according to Women Impact Tech. And with inflation, DeWeese and others are watching their spending power shrink. More than 10 years ago, she was earning over $100,000 in total compensation. That amount has dropped 15% since she was laid off from Yahoo in 2016, and has not increased since. "I feel like my career has been frozen in time," DeWeese said. "Things have been at a standstill." Paula Bratcher Ratliff, president of New York-based Women Impact Tech, said that the shrinking pay hits especially hard for women, given the continuing gender pay gap. "The Bay Area took one of the largest hits," Ratliff said. "Women make up about 28% of the entire workforce in tech. When you're seeing an overall decline at 15%, and for pay equity, women have not made much traction." [...] Despite the trend of shrinking salaries in the world's tech capital, Ratliff, with Women Impact Tech, doesn't believe it's necessarily a race to the bottom. "Today, about every company is a tech company, whether they're in retail, consumer goods or hospitality," Ratliff said. "There's so many opportunities in tech without having to focus on those jobs with the tech organizations alone. We're seeing great companies emerge." While it's still unclear where the light is at the end of the tunnel for DeWeese, she remains hopeful her situation will improve. "You have to have hope or else you're just going to live in fear of being let go, again and again," she said.

Read more of this story at Slashdot.

Systemd 256 is out today as the latest major feature update to this integral component to modern Linux distributions...

One of Qualcomm's indisputable strengths are its 5G modems – something which even Apple has yet to successfully ween itself from. And while Qualcomm is not integrating a modem into its first-generation Oryon-based Snapdragon X chips, the company is still looking to leverage that technology advantage via discrete modems that can be installed in Snapdragon X laptops.

To date, Qualcomm has won 23 laptop designs with its Snapdragon X Elite SoCs, and all of the leading PC vendors have introduced systems based on Qualcomm's Snapdragon X Elite processors. However, only some of them will be equipped with modems, the company detailed at Computex 2024.

While the technical rationale for this is very straightforward (not every vendor wants to buy and dedicate the space to modems), it's still a bit of a surprise in as much as Qualcomm has traditionally heavily pushed laptop vendors to include their modems. But as the Snapdragon X has entered the picture, the joint Qualcomm/Microsoft always connected PC (ACPC) initiative is taking a back seat – meaning modems are no longer being pushed nearly as hard. In its place, the two companies have pivoted hard to equipping mainstream systems with the hardware needed for local AI processing (i.e. NPUs), and with it, Microsoft's Copilot+ PC branding.

Laptop manufacturers, in the meantime, are breathing a sigh of relief, as this switch to emphasizing AI comes at a much lower hardware cost, since vendors don't need to buy additional discrete hardware. Qualcomm for its part has never fully disclosed the full cost of including a Snapdragon modem with a laptop, but the total cost adds up quickly. Besides buying a discrete modem, device manufacturers also need to buy and integrate a 5G-capable radio frequency front end module (RF FEM), as well as the all-important antenna. And mmWave support of any kind can add another wrinkle, as multiple antennas at different orientations are needed to get the best results.

And while not said out-loud, Qualcomm's premium positioning strategy for 8cx-based laptops has not paid significant dividends. Snapdragon X laptops are being priced much more competitively, as Qualcomm is aiming to capture a meaningful share of the PC market – and high-cost features like modems would drive up the final price tag.

Still, virtually all Qualcomm representatives I talked to at Computex were happy to argue that an integrated modem is a huge benefit for their PCs, as they can get fast connectivity almost everywhere in the world instantly and not depend on Wi-Fi or even their smartphones. So the dream of widespread 5G-capable laptops is not dead at Qualcomm; it may just be delayed. In the meantime, for laptop buyers that do need or want a 5G modem, there will still be at least a few premium laptop models on store shelves with the necessary hardware.

Yesterday, Apple made waves in the media when it revealed a partnership with OpenAI during its annual WWDC keynote. That announcement centered on Apple's decision to bring ChatGPT natively to iOS 18, including Siri and other first-party apps. During a followup interview on Monday, Apple executives Craig Federighi and John Giannandrea hinted at a possible agreement with Google Gemini and other AI chatbots in the future. 9to5Mac reports: Moderated by iJustine, the interview was held in Steve Jobs Theater this afternoon, featuring a discussion with John Giannandrea, Apple's Senior Vice President of Machine Learning and AI Strategy, and Craig Federighi, Senior Vice President of Software Engineering. During the interview, Federighi specifically referenced Apple's hopes to eventually let users choose between different models to use with Apple Intelligence. While ChatGPT from OpenAI is the only option right now, Federighi suggested that Google Gemini could come as an option down the line: "We think ultimately people are going to have a preference perhaps for certain models that they want to use, maybe one that's great for creative writing or one that they prefer for coding. And so we want to enable users ultimately to bring a model of their choice. And so we may look forward to doing integrations with different models like Google Gemini in the future. I mean, nothing to announce right now, but that's our direction." The decision to focus on ChatGPT at the start was because Apple wanted to "start with the best," according to Federighi.

Read more of this story at Slashdot.

British police have arrested two individuals involved in an SMS-based phishing campaign using a unique device police described as a "homemade mobile antenna," "an illegitimate telephone mast," and a "text message blaster." This first-of-its-kind device in the UK was designed to send fraudulent texts impersonating banks and other official organizations, "all while allegedly bypassing network operators' anti-SMS-based phishing, or smishing, defenses," reports The Register. From the report: Thousands of messages were sent using this setup, City of London Police claimed on Friday, with those suspected to be behind the operation misrepresenting themselves as banks "and other official organizations" in their texts. [...] Huayong Xu, 32, of Alton Road in Croydon, was arrested on May 23 and remains the only individual identified by police in this investigation at this stage. He has been charged with possession of articles for use in fraud and will appear at Inner London Crown Court on June 26. The other individual, who wasn't identified and did not have their charges disclosed by police, was arrested on May 9 in Manchester and was bailed. [...] Without any additional information to go on, it's difficult to make any kind of assumption about what these "text message blaster" devices might be. However, one possibility, judging from the messaging from the police, is that the plod are referring to an IMSI catcher aka a Stingray, which acts as a cellphone tower to communicate with people's handhelds. But those are intended primarily for surveillance. What's more likely is that the suspected UK device is perhaps some kind of SIM bank or collection of phones programmed to spam out shedloads of SMSes at a time.

Read more of this story at Slashdot.

An anonymous reader quotes a report from TechCrunch: A Finnish startup called Flow Computing is making one of the wildest claims ever heard in silicon engineering: by adding its proprietary companion chip, any CPU can instantly double its performance, increasing to as much as 100x with software tweaks. If it works, it could help the industry keep up with the insatiable compute demand of AI makers. Flow is a spinout of VTT, a Finland state-backed research organization that's a bit like a national lab. The chip technology it's commercializing, which it has branded the Parallel Processing Unit, is the result of research performed at that lab (though VTT is an investor, the IP is owned by Flow). The claim, Flow is first to admit, is laughable on its face. You can't just magically squeeze extra performance out of CPUs across architectures and code bases. If so, Intel or AMD or whoever would have done it years ago. But Flow has been working on something that has been theoretically possible -- it's just that no one has been able to pull it off. Central Processing Units have come a long way since the early days of vacuum tubes and punch cards, but in some fundamental ways they're still the same. Their primary limitation is that as serial rather than parallel processors, they can only do one thing at a time. Of course, they switch that thing a billion times a second across multiple cores and pathways -- but these are all ways of accommodating the single-lane nature of the CPU. (A GPU, in contrast, does many related calculations at once but is specialized in certain operations.) "The CPU is the weakest link in computing," said Flow co-founder and CEO Timo Valtonen. "It's not up to its task, and this will need to change." CPUs have gotten very fast, but even with nanosecond-level responsiveness, there's a tremendous amount of waste in how instructions are carried out simply because of the basic limitation that one task needs to finish before the next one starts. (I'm simplifying here, not being a chip engineer myself.) What Flow claims to have done is remove this limitation, turning the CPU from a one-lane street into a multi-lane highway. The CPU is still limited to doing one task at a time, but Flow's Parallel Processing Unit (PPU), as they call it, essentially performs nanosecond-scale traffic management on-die to move tasks into and out of the processor faster than has previously been possible. [...] Flow is just now emerging from stealth, with [about $4.3 million] in pre-seed funding led by Butterfly Ventures, with participation from FOV Ventures, Sarsia, Stephen Industries, Superhero Capital and Business Finland. The primary challenge Flow faces is that for its technology to be integrated, it requires collaboration at the chip-design level. This means chipmakers need to redesign their products to include the PPU, which is a substantial investment. Given the industry's cautious nature and the existing roadmaps of major chip manufacturers, the uptake of this new technology might be slow. Companies are often reluctant to adopt unproven technologies that could disrupt their long-term plans. The white paper can be read here. A Flow Computing FAQ is also available here.

Read more of this story at Slashdot.

The Raspberry Pi 5 features the "RP1" as the in-house silicon design for the southbridge to this single board computer. The RP1 driver maintained by Raspberry Pi is just found in their downstream kernel while a SUSE engineer is working to rework that driver so that it can be eventually mainlined in the upstream Linux kernel...

Four more U.S. states on Tuesday joined the Justice Department's lawsuit against Apple alleging the iPhone maker is monopolizing smartphone markets, the department said in a statement. From a report: The four states are Indiana, Massachusetts, Nevada and Washington, the Justice Department said. The original lawsuit was filed in March, and 15 states and the District of Columbia joined the lawsuit at the time. The lawsuit alleges that Apple uses its market power to get more money from consumers, developers, content creators, artists, publishers, small businesses and merchants. The civil lawsuit accuses Apple of an illegal monopoly on smartphones, maintained by imposing contractual restrictions on, and withholding critical access from, developers. The Justice Department has previously said Apple charges as much as $1,599 for an iPhone and makes a larger profit than any rival. Officials also said Apple imposes hidden charges on various business partners - from software developers to credit card companies and even rivals such as Alphabet's, Google, in ways that ultimately raise prices for consumers.

Read more of this story at Slashdot.

An anonymous reader shares a report: Since April, a hacker with a history of selling stolen data has claimed a data breach of billions of records -- impacting at least 300 million people -- from a U.S. data broker, which would make it one of the largest alleged data breaches of the year. The data, seen by TechCrunch, on its own appears partly legitimate -- if imperfect. The stolen data, which was advertised on a known cybercrime forum, allegedly dates back years and includes U.S. citizens' full names, their home address history and Social Security numbers -- data that is widely available for sale by data brokers. But confirming the source of the alleged data theft has proven inconclusive; such is the nature of the data broker industry, which gobbles up individuals' personal data from disparate sources with little to no quality control. The alleged data broker in question, according to the hacker, is National Public Data, which bills itself as "one of the biggest providers of public records on the Internet." On its official website, National Public Data claimed to sell access to several databases: a "People Finder" one where customers can search by Social Security number, name and date of birth, address or telephone number; a database of U.S. consumer data "covering over 250 million individuals;" a database containing voter registration data that contains information on 100 million U.S. citizens; a criminal records one; and several more. Malware research group vx-underground said on X (formerly Twitter) that they reviewed the whole stolen database and could "confirm the data present in it is real and accurate."

Read more of this story at Slashdot.

schwit1 shares a report: Bill Gates and his energy company are starting construction at their Wyoming site for a next-generation nuclear power plant he believes will "revolutionize" how power is generated. Gates was in the tiny community of Kemmerer Monday to break ground on the project. The co-founder of Microsoft is chairman of TerraPower. The company applied to the Nuclear Regulatory Commission in March for a construction permit for an advanced nuclear reactor that uses sodium, not water, for cooling. If approved, it would operate as a commercial nuclear power plant. The site is adjacent to PacifiCorp's Naughton Power Plant, which will stop burning coal in 2026 and natural gas a decade later, the utility said. Nuclear reactors operate without emitting planet-warming greenhouse gases. PacifiCorp plans to get carbon-free power from the reactor and says it is weighing how much nuclear to include in its long-range planning. The work begun Monday is aimed at having the site ready so TerraPower can build the reactor as quickly as possible if its permit is approved. Russia is at the forefront for developing sodium-cooled reactors.

Read more of this story at Slashdot.

A Chromebook refresh looms despite Google trying to extend the life of laptops by offering a decade of service updates for models sold since 2021. From a report: Sales of the hardware, which flew off the shelves during the pandemic, ran out of steam in 2022 after buyers had their fill. The US education market generally accounts for 70 to 80 percent of annual orders. The sharp downturn left some vendors holding excess inventory. Yet the refresh cycle may be starting again, according to HP boss Enrique Lores. "So we have started to see a pickup of demand in education, and this, especially in the US, is a Chromebook opportunity," he told an audience of investors at Bernstein's 40th Annual Strategic Decision Conference. He forecast a flurry of activity in 2025 for "many million of units" from education but downplayed the impact on HP's balance sheet because the company pulled back from the product line after the pandemic. Lores said: "We are going after these deals because we think it's good, but it's not like ... a huge impact on the company."

Read more of this story at Slashdot.

Un air de déjà-vu

Si les annonces d’Apple sur l’IA étaient attendues de pied ferme, les nouvelles versions majeures de ses plateformes ont d’autres nouveautés. C’est particulièrement vrai sur iPhone et iPad, avec un accent mis sur la personnalisation. Au point de rappeler furieusement Android par moments.

La présentation d’Apple s’est faite en deux temps. Une première partie listant les nouveautés classiques que l’on peut attendre de nouvelles versions majeures de ses systèmes. La seconde était consacrée à l’intelligence artificielle. L’exposé, relativement long, devait démontrer la vision intégrée d’Apple et son souci de la vie privée. En pratique, Apple est en retard, les fonctions n’arriveront qu’en bêta à l’automne, qu’en anglais et uniquement aux États-Unis.

• Apple Intelligence : une cascade de fonctions, une disponibilité très limitée

Même si Apple était attendue au tournant face à une concurrence prenant le large (particulièrement Microsoft, portée par les succès d’OpenAI), ses plateformes ont leurs nouveautés propres. Comme les années précédentes, c’est surtout vrai pour iOS et iPadOS. macOS récupère presque tous les apports et y ajoute quelques spécificités, comme iPhone Mirroring, qui permet d’afficher l’écran de son iPhone sur l’ordinateur. Nous y reviendrons dans un futur article, avec les autres plateformes.

iOS 18 fait un gros lâcher sur la personnalisation

Android ! Le nom survient seul, tant les nouveautés d’Apple sur la personnalisation de son iOS 18 rappellent le système de Google, surtout dans sa version « stock » (sans surcouche).

L’écran d’accueil, en particulier, est beaucoup plus personnalisable. On peut par exemple disposer les icônes où l’on veut sur la grille, et pas uniquement les unes à la suite des autres. On peut également changer la teinte des icônes pour leur imposer une couleur spécifique ou harmonisée avec le fond d’écran.

Ces réglages peuvent être appliqués sur une icône ou sur toutes. En outre, on peut supprimer les noms des applications et afficher certaines icônes dans un format plus grand.

Personnalisation également pour le centre de contrôle, qui en avait bien besoin. Premier gros changement, l’écran peut être scindé en plusieurs. On peut y regrouper les contrôles que l’on souhaite, selon ses besoins ou envies. Par exemple, pour regrouper tous les contrôles domotiques de la maison. On peut faire défiler ces écrans par le geste classique ou en prolongeant le glissement vers le bas depuis l’appel du centre. Deuxième changement, la possibilité de placer les boutons d’action (aussi ronds que ceux d’Android) où l’on souhaite. On peut également modifier leur taille.

Ces modifications s’étendent même aux deux boutons situés en bas de l’écran verrouillé : Lampe torche et Appareil photo. On peut désormais leur affecter d’autres fonctions, y compris provenant d’applications tierces. Dans sa démonstration, Apple montrait comment Appareil photo avait été remplacé par l’ouverture d’une nouvelle publication sur Snapchat. Il est aussi possible de supprimer ces deux boutons.

Apple casse un peu la sobriété de Messages

Nouvelle salve de nouveautés pour Messages. iOS 17 avait déjà apporté quelques « menus plaisirs », iOS 18 va un cran plus loin. Par exemple en colorant les réactions et en ajoutant – enfin – la possibilité de réagir à un message avec n’importe quel émoji.

Il est également possible d’ajouter des effets sur un mot dans un message pour lui donner une certaine emphase. On a le choix entre des effets comme « Grand », « Secousse », « Ondulation » ou encore « Explosion ». On peut sélectionner un mot et lui appliquer l’un de ces effets. Certains mots doivent être détectés automatiquement pour qu’iOS propose d’ajouter l’effet correspondant, mais la fonction ne semble pas encore disponible en français.

En plus de la gestion très attendue du RCS, iOS 18 permet de préparer à l’avance un message pour l’envoyer à la date et l’heure choisie. Par exemple, pour ne pas rater un prochain anniversaire au moment où on y pense.

Surtout, Messages pourra passer par satellites quand l’iPhone n’a pas de réseau. La fonction est disponible depuis les iPhone 14 et peut servir à envoyer des iMessage comme des SMS. Même les réactions et effets sont gérés. Apple précise que le chiffrement de bout en bout est préservé pour iMessage. Cette nouveauté sera d’abord réservée aux États-Unis, comme le positionnement par satellite à ses débuts.

Notez que nous faisons volontairement l’impasse sur toutes les fonctions liées à Apple Intelligence, résumées dans notre article dédié. On pourrait sinon aborder les envoies d’emojis personnalisés (Genmoji) ou encore la génération d’images.

Photos change d’approche

Apple a largement insisté sur la nouvelle version de Photos, qui aurait reçu sa « plus grosse évolution » depuis qu’iOS existe. Finie la vue fragmentée, qui avait d’ailleurs tendance à se montrer redondante. On trouve désormais un accueil unique, avec la grille en haut de l’écran et les collections en-dessous.

Ces collections sont une nouveauté et sont générées par IA, en fonction de thématiques se voulant plus intelligentes que celles créées automatiquement aujourd’hui. On peut épingler les collections les plus intéressantes et supprimer les autres. Si l’on « swipe » vers la gauche, Photos révèle un carrousel mettant en avant les souvenirs jugés comme les plus importants. Le carrousel est personnalisable. On peut également créer des souvenirs autour d’évènements avec un plus grand nombre d’options de personnalisation. Une musique « adaptée » est sélectionnée depuis Apple Music, avec possibilité d’en charger.

Photos comporte également des fonctions alimentées par l’IA et dont on ne sait pas encore quand elles seront réellement disponibles. Par exemple, Clean Up permet de sélectionner un élément dans un cliché pour le supprimer. Une fonction classique dans les applications de retouche et qu’Apple intègre donc dans iOS. Sans doute pour faire écho à Google, qui inclut ce type de fonctions dans sa propre application.

Vie privée : des applications à cacher et à verrouiller

iOS 18 introduit plusieurs fonctions importantes pour la vie privée dans la gestion des applications. Le système est ainsi capable de verrouiller n’importe quelle application, qu’elle soit interne ou tierce. Son ouverture réclame alors une authentification biométrique. En outre, cette application peut être cachée dans un nouveau dossier spécial, destiné à cacher tout ce qui s’y trouve et lui-même verrouillé.

Quand des applications sont verrouillées, ce qu’elles contiennent ne peuvent plus apparaître dans les recherches ou ailleurs. Les notifications ne sont plus affichées. L’idée est de pouvoir mettre son iPhone dans les mains d’une autre personne sans risquer que des informations sensibles soient affichées accidentellement, jusqu’à la présence d’une application spécifique.

Apple ajoute également la possibilité de ne sélectionner que certains contacts à partager avec une application. Dans le même esprit, la connexion d’un accessoire à une application peut se faire sans que celle-ci voir les autres appareils du réseau.

Mode jeu, Tap to Cash et le reste

On note plusieurs autres apports assez importants dans iOS 18. Par exemple, le Mode jeu, que l’on trouvait dans macOS Sonoma, débarque sur iPhone (et iPad). Le fonctionnement est le même : réduction de la priorité des tâches en arrière-plan et amélioration « considérable » de la latence pour les manettes et AirPods.

Tap to Cash est une extension d’Apple Pay. Comme son nom l’indique, la fonction autorise l’échange d’argent entre deux iPhone simplement approchés l’un de l’autre. Il faut malheureusement que le compte bancaire soit compatible Apple Cash… ce qui n’est possible qu’aux États-Unis. Dommage. L’application Cartes est aussi mise à jour, offrant une nouvelle vue pour les billets d’évènements, tels que les concerts.

Citons également l’affichage des Rappels dans Calendrier, la possibilité dans Maison de créer des accès invités pour leur laisser le contrôle de certains accessoires domotiques, une révision de la Fiche médicale pour mieux mettre en avant les informations cruciales, ou encore, pour les possesseurs d’iPod Pro de 2^e génération, la possibilité de hocher ou secouer la tête pour répondre « oui » ou « non » aux questions de Siri. Journal propose de son côté une nouvelle vue d’ensemble, peut compter le temps d’écriture, propose un widget et simplifie la recherche. Ajoutons, dans Calendrier, une vue mensuelle détaillée, permettant d’afficher les noms des évènements sans avoir besoin d’entrer dans la vue journée.

Enfin, les performances de Safari ont été améliorées (comme à chaque fois) et on note l’apparition d’une application dédiée pour la gestion des mots de passe. Un changement important sur lequel nous reviendrons dans notre article sur le nouveau macOS, Sequoia.

iOS 18 pour qui ?

La réponse est simple : tout iPhone actuellement compatible avec iOS 17 pourra installer iOS 18. Comme toujours, cela ne signifie que tout le monde sera servi de la même manière. On le sait déjà pour Apple intelligence, qui ne pourra fonctionner qu’avec au moins un iPhone 15 Pro. Mais d’autres fonctions peuvent ne pas être présentes sur des appareils anciens.

L’iPad, roi des maths ?

iPadOS 18 récupère évidemment toutes les nouveautés présentées dans iOS 18 et en ajoute quelques-unes « rien qu’à lui ».

La plus impressionnante est sans doute Math Notes. On écrit dans l’application avec le stylet comme on le ferait avec un papier et un crayon pour faire ses calculs. À la différence que la tablette va calculer automatiquement les résultats et les afficher, en respectant la calligraphie. On peut écrire a priori n’importe quel type d’expression mathématique, les modifier, les transformer en équations, générer des graphes ou encore assigner une valeur à une variable dans un coin de l’écran.

Cet ajout est complété par un autre, auquel on ne croyait plus : Calculette. Apple a beau avoir présenté cette nouveauté avec un certain humour, on ne comprend toujours pas ce qui a nécessité tant de temps. Les modes basique et scientifique sont présents, de même que l’historique et la conversion des longueurs, poids et devises.

Une préservation de la calligraphie

Cet aspect, évoqué dans Math Notes, est en fait une fonction nommée Smart Script. Elle analyse la calligraphie de l’utilisateur et permet par la suite de la « nettoyer », tout en le redressant. Ce texte est en outre manipulable comme n’importe quel autre qui aurait été tapé au clavier. L’application de la calligraphie fonctionne également pour du texte collé dans la note.

On ne sait pas encore dans quelle mesure ce traitement, tout comme Math Notes, font appel à Apple Intelligence. Si tel est le cas, ce sont deux ajouts majeurs qui disparaissent d’entrée pour toutes les tablettes n’étant pas munies au moins d’une puce M1. Et quand bien même, il faudrait encore résider aux États-Unis, jusqu’à ce qu’Apple élargisse la disponibilité de ses fonctions IA.

Une barre d’onglets personnalisable

Il y a au moins une nouveauté que les utilisateurs sont certains de retrouver : une nouvelle barre d’onglets. Flottante, elle peut être personnalisée avec les éléments présents d’ordinaire dans la zone latérale gauche. Elle est censée apparaître dans toutes les applications proposant des contenus.

Signalons plusieurs autres ajouts, comme les sections rétractables dans Notes ainsi qu’un nouveau mode Lecteur plus épuré dans Safari.

Quelques vieux iPad éliminés

Contrairement à iOS 18, iPadOS 18 ne prend pas en charge tous les modèles de son prédécesseur. L’iPad 6 ainsi que les iPad Pro de 2^e génération passent à la trappe. Il faut donc, au minimum, un iPad 7, mini 5, Air 3 ou un iPad Pro de 3^e génération.

IA pas de consentement

Human Right Watch a analysé une partie de la base de données LAION-5B très utilisée pour entrainer des outils d’IA générateurs d’images et s’est rendu compte qu’elle contiendrait des liens vers des photos d’enfants brésiliens sans leur consentement.

L’ONG Human right watch explique avoir repéré des photos personnelles d’enfants brésiliens dans la base de données LAION-5B. Créée le professeur de lycée allemand Christoph Schuhmann, celle-ci a été notamment utilisée par Stable Diffusion et par Google pour entrainer leurs modèles d’IA génératives de text-to-image.

• Comment LAION a créé un jeu d’images d’entraînement à partir de zéro

Une toute petite partie de la base de données explorée

Ces photos ne figurent pas en tant que telles dans la base de données. LAION-5B liste notamment des liens vers diverses photos qui ont été mises en ligne qu’elle associe à du texte. Elle s’appuie sur l’autre base de données Common Crawl qui parcourt internet et stocke les contenus trouvés.

La chercheuse de l’ONG, Hye Jung Han, a examiné une toute petite partie de LAION-5B (moins de 0,0001 % des 5,85 milliards d’images) mais a trouvé 170 photos d’enfants brésiliens venant d’au moins 10 États du pays.

Des photos de moments intimes

La plupart de ces photos n’ont été vues que par très peu de personne et « semblent avoir bénéficié auparavant d’une certaine intimité » explique Human Rights Watch, qui a vérifié en utilisant des moteurs de recherche.

L’ONG affirme que LAION, l’association allemande fondée par Schuhmann pour gérer la base de données, a confirmé l’existence des liens vers ces photos dans sa base de données et a promis de les supprimer. Mais elle a ajouté qu’il incombait aux enfants et à leurs tuteurs de retirer les photos personnelles des enfants de l’internet, ce qui, selon elle, constitue la protection la plus efficace contre les abus.

• Intelligence artificielle : la CNIL veut (re)concilier « innovation et respect des droits »

Dans une réponse à Wired, LAION a affirmé avoir supprimé les liens vers les contenus signalés par Human Right Watch. Mais un de ses représentants a ajouté que « la suppression des liens d’un ensemble de données LAION ne supprime pas ce contenu du web […] il s’agit d’un problème plus vaste et très préoccupant, et en tant qu’organisation bénévole à but non lucratif, nous ferons notre part pour y remédier ».

Selon l’association, les photos listées par LAION représentaient « des moments intimes comme des bébés naissant entre les mains gantées de médecins, des jeunes enfants soufflant les bougies de leur gâteau d’anniversaire ou dansant en sous-vêtements à la maison, d’élèves faisant un exposé à l’école et d’adolescents posant pour des photos à carnaval du lycée ».

Hye Jung Han explique à Wired que « leur vie privée est violée en premier lieu lorsque leur photo est récupérée et intégrée dans ces ensembles de données. Ensuite, ces outils d’intelligence artificielle sont entrainés à partir de ces données et peuvent donc créer des images réalistes d’enfants ». Elle ajoute que « la technologie est développée de telle sorte que tout enfant qui possède une photo ou une vidéo de lui en ligne est désormais en danger, car n’importe quel acteur malveillant pourrait prendre cette photo, puis utiliser ces outils pour la manipuler à sa guise ».

LAION-5B plus accessible publiquement

Depuis décembre dernier, LAION-5B n’est plus accessible publiquement. L’association a pris cette décision car des chercheurs de Stanford ont identifié 3 226 liens vers des images pédocriminelles potentielles. « La plupart d’entre elles ont été identifiées comme telles par des tierces parties » expliquaient-ils.

Dans un communiqué sur son site, LAION affirmait qu’elle appliquait «  une politique de tolérance zéro à l’égard des contenus illégaux et, dans un souci de prudence, nous retirons temporairement les jeux de données de LAION pour nous assurer qu’ils sont sûrs avant de les republier ».