Trump called off a planned AI executive order just hours before a signing ceremony because he said he was worried the framework could slow America's lead over China. "We're leading China, we're leading everybody, and I don't want to do anything that's going to get in the way of that lead," Trump told reporters. The Associated Press reports: The order would have established a framework for the government to vet the national security risks of the most advanced AI systems before their public release, according to a person familiar with the White House's deliberations with the tech industry but not authorized to speak about it publicly. The directive was being characterized as a voluntary collaboration with participating U.S.-based tech companies, including Anthropic, OpenAI and Google, the person said.
There are competing factions within the administration, said Serena Booth, a computer science professor at Brown University and former AI policy fellow in a Democratic-led Senate committee. "We do see this kind of public fighting," she said. "'We will release an executive order. No, we won't. We're going to sign it this afternoon. Oh, the signing is canceled.' I think this whiplash is because we're seeing these fractures.'"
Some of those divides are balancing what Booth said is a "reasonable idea" to test the most capable AI models before their public release, with a concern that government scrutiny, if it takes too long, could burden AI developers. "It does come at a potential very large cost to innovation and speed of development," she said. "There is, I think, a real risk here and I do see both sides." [...]
"They don't want to do it because it's politically risky in a million different ways," said Dean Ball, now at the Foundation for American Innovation. Ball said he would welcome an executive order that would get those companies working more closely with the government on cybersecurity but "ultimately, I'm fine with them taking time to get this right."
An anonymous reader quotes a report from TechCrunch: OpenAI claims its new reasoning model has produced an original mathematical proof disproving a famous unsolved conjecture in geometry, which was first posed by Paul Erdos in 1946. If this sounds familiar to you, it's because this isn't the first time OpenAI has made such a bold claim. Seven months ago, the AI giant's former VP Kevin Weil posted on X: "GPT-5 found solutions to 10 (!) previously unsolved Erds problems and made progress on 11 others."
It turns out, GPT-5 didn't actually solve those problems; it just found solutions that already existed in the literature. Taunts from rivals like Yann LeCun and Google DeepMind CEO Demis Hassabis followed, and Weil promptly took down his premature post. Today, at least, it seems OpenAI didn't make the same mistake twice. Alongside the announcement, the company published companion remarks (PDF) in support of the disproof from mathematicians like Noga Alon, Melanie Wood, and Thomas Bloom, who maintains the Erdos Problems website, and previously called Weil's post "a dramatic misrepresentation."
[...] The proof, per OpenAI, came from a new general-purpose reasoning model, not a system specifically designed to solve math problems or even this problem in particular. OpenAI says this is significant because it means AI systems are now more capable of holding together long, difficult chains of reasoning and connecting ideas across fields in ways researchers may not have previously explored. That has implications for biology, physics, engineering, and medicine.
OpenAI s’apprêterait à déposer confidentiellement son dossier d’introduction en Bourse aux États-Unis, avec une IPO potentiellement prévue dès septembre 2026, rapporte Reuters le 20 mai 2026. Une opération historique qui pourrait raviver la rivalité entre Sam Altman et Elon Musk, alors que SpaceX prépare elle aussi son arrivée sur les marchés.
Le 20 mai 2026, OpenAI a affirmé qu’un de ses modèles d’IA avait contribué à réfuter une conjecture mathématique formulée par Paul Erdős en 1946. Le problème consiste à déterminer combien de paires de points peuvent être placées exactement à distance 1 dans un plan.
The 2026 Commonwealth Short Story Prize is facing backlash after several winning entries were accused of being AI-generated, with one Caribbean winner's story flagged as fully AI-written by a detector that WIRED says it independently confirmed. From the report: Each year, the Commonwealth Foundation, a nongovernmental organization in London, awards its short story prize to one writer in each of five regions: Africa, Asia, Canada and Europe, the Caribbean, and the Pacific. One overall winner is then selected from that short list. Regional winners take home [about $3,350], while the top winner, to be announced next month, claims [about $6,700]. On May 12, the respected UK literary magazine Granta published the top five 2026 entries -- all previously unpublished, per the rules of the contest -- on its website. (It has hosted the winning submissions for the prize since 2012.) Within days, however, one entry aroused suspicion. "The Serpent in the Grove," a story by Jamir Nazir of Trinidad and Tobago, which had taken honors for the Caribbean region, struck a few people as bearing the stylistic tells of AI-generated text.
"Well, this is a first: a ChatGPT-generated story won a prestigious literary prize," wrote researcher and entrepreneur Nabeel S. Qureshi, a former visiting scholar of AI at the Mercatus Center at George Mason University, in a post on X on Monday. "'Not X, not Y, but Z' sentences everywhere, the 'hums' trope, and plenty of other obvious markers of AI writing. A major milestone for AI, at any rate..." "They say the grove still hums at noon," Nazir's mysterious and atmospheric tale begins. In his screenshot of the opening paragraphs, Quereshi highlighted the second line as what he considered to be a signature example of AI syntax: "Not the bees' neat industry or the clean rasp of cutlass on vine, but a belly sound -- as if the earth swallows a shout and holds it there."
As the literary community undertook a closer read of Nazir's story, many criticized its language and metaphors as nonsensical, wondering how the Commonwealth judges could have seen any merit to them. Others shared screenshots showing that the AI-detection tool Pangram flagged "The Serpent in the Grove" as 100 percent AI-generated, a result that WIRED independently confirmed. (While no AI-detection software is perfect, third-party analysis has consistently determined Pangram to be the most accurate, with a near-zero rate of false positives.)
[...] Besides Nazir, two more winning authors have drawn allegations of using AI in their work. Pangram finds that "The Bastion's Shadow," by Maltese writer John Edward DeMicoli, winner for the Canada and Europe region, is fully AI-generated; it scans "Mehendi Nights," by Indian writer Sharon Aruparayil, winner for the Asia region, as partly AI-generated. Neither DeMicoli nor Aruparayil immediately returned requests for comment when reached through their respective social media accounts. The other two short-listed stories, by Holly Ann Miller of New Zealand and Lisa-Anne Julien of South Africa, deliver "fully human-written" results from Pangram. Wired also reports that one of the judges for the prize has been "accused of using AI to craft her descriptive blurb that accompanied the listing of 'The Serpent in the Grove' as a regional winner.'" Pangram labels the text as "AI-assisted."
OpenAI propose désormais aux entreprises de réserver sa puissance de calcul sur plusieurs années, une offre qui répond à une pénurie de GPU que la société reconnaît elle-même.
En attirant l'un des profils les plus reconnus de l'IA mondiale, Anthropic envoie un signal qui dépasse la compétition pour les talents : elle parie sur l'automatisation de sa propre recherche.
Google is giving its iconic search box its first major redesign since 2001. The new design incorporates, you guessed it, artificial intelligence, "getting bigger and more interactive so that people can ask even longer questions and upload photographs and videos into queries," reports the New York Times. "In addition, people can ask follow-up questions with a chatbot on Google's main search page." From the report: The company will also offer digital assistants, known as agents, to automate searches so that someone who may be apartment hunting can be notified of a new listing without opening a real estate site like Zillow. The search features will be powered by a new artificial intelligence model, Gemini 3.5 Flash. Google said the model had improved on creating software code and performing autonomous tasks, worked faster and was less expensive to run than comparable models.
[...] Google is also bringing one of A.I.'s biggest breakthroughs -- software coding -- to search. When people research complex topics like astrophysics, Gemini can build interactive graphics and simulations behind the scenes to provide a deeper answer than its previous listing of websites. Google said it was introducing an alternative to the agents powered by Anthropic's Claude Code and OpenAI's Codex. Called Gemini Spark, the service is embedded in Gmail, Docs and other Google products, where it can turn meeting notes spread across emails and chats into a single document. It can also read and draft emails. "The open web is on its way out," says Richard Kramer, a financial analyst with Arete Research. "With A.I., Google is reducing everyone to raw data providers."
En pleine messe Google I/O, l’entreprise a finalement révélé son plan pour mettre fin aux doutes que l’on peut avoir sur certaines photos générées par l’IA en utilisant SynthID.
En pleine conférence Google I/O 2026, la firme de Mountain View a dévoilé l'Universal Cart : un panier d’achat universel entièrement dopé à l’IA. En quoi cela consiste réellement ?
Coup de théâtre dans le duel judiciaire de l'année. Le jury a balayé la plainte du milliardaire contre Sam Altman pour une simple question de prescription légale. Elon Musk s'est réveillé beaucoup trop tard, mais il annonce déjà sa riposte en appel.
[Deal du jour] La deuxième génération de lunettes connectées Meta Ray-Ban fait mieux que les premiers modèles, et est une excellente affaire lorsque leur prix baisse.
Amazon is adding AI-generated "podcasts" to Alexa+, letting users request custom audio explainers on any topic featuring two synthetic co-hosts. Variety reports: Seemingly to dispel the notion that these "podcasts" will be AI audio slop, Amazon emphasized that it has deals with major news organizations to ensure "accurate, real-time news and information." Those include the Associated Press, Reuters, the Washington Post, Time magazine, Forbes, Business Insider, Politico and USA Today; publications from Conde Nast, Hearst and Vox Media; and more than 200 local newspapers across the U.S.
In an example clip shared by Amazon of the new Alexa Podcasts feature, the two AI-generated hosts discuss "the latest music releases." A male Alexa+ narrator says more than 50% of music listening now comes from unsigned artists. "The monoculture is just gone," a female-voiced Alexa+ narrator chimes in. The male Alexa+ host says there has been "stoner metal," indie pop and experimental hip-hop music "all dropping on the same Friday," and adds, "That's not chaos -- that's the healthiest the music ecosystem has ever been."
[...] To use Alexa Podcasts, users can simply tell Alexa what topic they're curious about and "it does the rest in minutes." Alexa+ will provide an overview of what it plans to cover, and let you adjust the length and direction before it generates the podcast. When your episode is ready, you'll get a notification on your Echo Show device and the Alexa app.
Coup de théâtre dans le duel judiciaire de l'année. Le jury a balayé la plainte du milliardaire contre Sam Altman pour une simple question de prescription légale. Elon Musk s'est réveillé beaucoup trop tard, mais il annonce déjà sa riposte en appel.
Quand Volkswagen nous a conviés à essayer une partie de sa gamme, nous avons d’abord demandé la liste des voitures. On a vu « Golf GTI Edition 50 », on n’a pas hésité une seconde. Dans un monde où plus de deux autos sur trois testées dans la rubrique sont électrifiées, nous n’allions pas rater l’occasion de prendre le volant d’une compacte sportive 100% thermique. Le monde d’avant nous manquera…
50 ans bien célébrés!
La Volkswagen Golf GTI Edition 50 reprend tous les codes de la compacte sportive comme on les aime, et qui devient une denrée de plus en plus rare pour les raisons que l’on connaît. Et cela commence d’abord par le look, assez proche de celui de la Clubsport. Et pour ne rien gâcher, notre exemplaire dispose du pack « GTI Performance ». En cochant cette option, on profite de jantes spécifiques plus légères, d’un échappement Akrapovič également allégé, ainsi que de pneus semi-slicks Bridgestone Potenza Race. On peut donc considérer que le constructeur ne s’est pas contenté de coller quelques stickers et de souffler des bougies.
Même si l’image d’une Golf sportive est parfois maladroitement associée au profil d’un jeune conducteur intrépide, pour le dire ainsi, nous savons que celle-ci en particulier vaut bien mieux que cela. On a bien conscience que le grand public ne ferait sans doute pas la différence sur la route. Toujours est-il qu’elle a un sacré look. Entre son large bouclier alvéolé, les étriers de frein rouges et le becquet, pour ne citer que ces éléments, tous ces détails contribuent à lui donner une allure franchement badass, surtout dans ce coloris. En tout cas, nous, on adore.
Du sport et des technologies
À bord, Wolfsburg a aussi voulu marquer la différence avec les autres versions de GTI du catalogue. Il y en aura d’ailleurs 50, et pas une de plus, sur notre territoire. On retrouve à l’intérieur les palettes de la R de 333 ch. Pour le reste, la sellerie vintage rappelle ce que vous savez. On note aussi un pédalier rouge. Des touches de la même couleur se retrouvent ici et là dans l’habitacle, ainsi qu’une inscription « GTI 50 » à la base du volant. On connaît déjà le confort de ces sièges, à la fois très agréables et au maintien irréprochable. Ils sont également accompagnés de ceintures rouges. Cette couleur, on l’a bien compris, est indissociable de la griffe GTI maison.
Sincèrement, on se sent plutôt bien à l’intérieur de cette sportive. Elle ne boude pas non plus un certain embourgeoisement, plutôt bienvenu dans l’automobile actuelle. On retrouve ainsi tous les équipements modernes de la compacte routière en bonus : ADAS sophistiqués, dont une conduite semi-autonome de niveau 2, un grand écran avec toute la connectivité attendue sur une auto de ce segment et de ce niveau. En outre, on peut monter assez confortablement à quatre, beaucoup moins à cinq. Avec un coffre à la contenance relativement moyenne, on peut néanmoins envisager cette Golf GTI Edition 50 comme voiture de tous les jours.
325 ch! Juste 8 de moins que la « R »
Pouvoir emmener les enfants en classe verte avec cette voiture ? Vous vous en moquez. Sauf si cela vous laisse l’opportunité de rentrer à vide après les avoir déposés. Et c’est là que l’intérêt de cette voiture prend tout son sens. Il y a d’abord ce moteur, le 2.0 TSI maison, reprogrammé par les sorciers de Wolfsburg pour délivrer 325 ch et 420 Nm de couple, des valeurs supérieures à celles de la Clubsport. Le 0 à 100 km/h est abattu en 5,3 s (-0,3 s) et la vitesse de pointe grimpe à 270 km/h. On ne va pas tourner autour du pot, le bloc est plein comme un œuf ! En revanche, même si Akrapovič s’est impliqué sur l’échappement, on regrette une sonorité peu démonstrative, quel que soit le mode.
Nous n’avons pas essayé le programme « Nürburgring », a priori calibré spécifiquement pour cette piste mythique. En revanche, le mode Sport nous a paru parfaitement approprié pour avaler les lacets de notre parcours entre Séville et le Portugal. Cela nous manquait de faire prendre des tours à un moteur ! Les rapports de la DSG s’enchaînent quasiment sans rupture, tandis que l’aiguille du tachymètre à l’écran s’affole. Était-ce vraiment nécessaire d’accompagner les montées en régime d’un son diffusé dans les haut-parleurs ? On n’en est pas sûrs. Toujours est-il que le moteur ne manque pas de souffle. On peut bien entendu reprendre la main sur la boîte, mais il faut reconnaître qu’elle fait très bien le travail lorsqu’on la laisse faire. On se demande bien ce qu’il reste à la R…
Béni pack GTI Performance
Pour accompagner ce caractère moteur, contrairement à bien des électriques gavées de puissance, on bénéficie ici d’un châssis à la hauteur, qui permet de profiter pleinement de la cavalerie, ou l’inverse. Tout d’abord, la Golf GTI Edition 50 équipée du pack Performance est abaissée de 5 mm par rapport à une GTI « normale ». La suite se joue dans les réglages d’amortissement, dont on vous épargne les détails. L’allègement évoqué plus haut participe également à ce comportement sportif exacerbé. Il faut néanmoins s’assurer d’avoir des gommes bien chaudes avant que la route ne se transforme littéralement en rail. On a le sentiment de maîtriser parfaitement l’auto, compte tenu de la précision du train avant.
Et lorsqu’on enfonce l’accélérateur en sortie de virage, le différentiel électronique piloté tire la voiture vers l’intérieur, ce qui limite l’élargissement de trajectoire, classique sur les tractions puissantes. Sur asphalte sec, grâce à ce dispositif sophistiqué, et avec des gommes très performantes, la motricité est imperturbable. Il faut tout de même bien tenir le volant et rester vigilant. Sur route, nous n’avons pas eu à nous plaindre des freins. La pédale finit par s’enfoncer légèrement si les lignes droites sont courtes, mais rien d’anormal. La question se poserait peut-être sur circuit, à vérifier.
Méchant malus…
Il faut saluer la capacité des concepteurs d’une voiture comme celle-ci à mettre sur la route une auto aussi rapide et accessible pour des conducteurs qui ne sont pas des pilotes aguerris. L’esprit attendu est bien présent, avec une optimisation liée au pack Performance qui satisfera les plus exigeants. Avec 173 g de CO2/km, le malus n’est pas maximal… mais il atteint tout de même près de 30 000 €. Au final, il faut débourser un peu plus de 96 000 € pour se l’offrir chez nous.
John Lennon's last interview — just hours before he was shot on December 8, 1980 — has become a documentary directed by Steven Soderbergh, debuting Saturday at the Cannes Film Festival.
In a new interview with the Associated Press, Soderbergh defends the film's limited use of AI to visualize concepts from that two-hour interview with John Lennon and Yoko Ono:
Soderbergh was resolved to let the audio play. He could finds ways to visualize much of the film, but that still left a large gap where the conversation grows more philosophical. "I worked on everything that could be solved except that for as long as I could," Soderbergh says. "Then there was the inevitable moment of: OK, but really what are we going to do? We just started playing and ran out of time and money. That's where the Meta piece came in." Soderbergh accepted an offer to use Meta's artificial intelligence software to conjure surreal imagery for those sections, which make up about 10% of the film.
When Soderbergh let the news out earlier this year, it prompted an uproar. One of America's leading filmmakers was using AI? In a film about a Beatle, no less? The AI parts (overwhelmingly slammed by critics in Cannes) are fairly banal and don't differ greatly from special effects — there are no deepfakes of Lennon. But they put Soderberg at the forefront of an industrywide debate about the uses of AI in moviemaking. It's a conversation the director, who has made movies on iPhones, is eager to have.
While the film follows John and Yoko's conversation, "I needed a way to follow them in flight visually," Soderbergh says, "or I'm not doing my job." Though when asked about the strong negative reaction, Soderbergh acknowleges that "I knew what was coming. I take it very seriously, and I understand why people have an emotional response to this subject. As I've said before, I feel like I owe people the best version of whatever art I'm trying to make and total transparency about how I'm doing it."
AP: Some fear generative AI will tear apart the film industry. You don't see it as a bogeyman, though.
SODERBERGH: I think most jobs that matter when you're making a movie cannot be performed by this tech and never will be performed by this tech. As it becomes possible for anybody to create something that meets a certain standard of technical perfection, then imperfection becomes more valuable and more interesting. We haven't seen yet someone with a certain amount of creative credibility go full-metal AI on something, and see how people react. I think it's necessary. How do you know where the line is until somebody crosses it?
"I don't think what I'm doing crosses it. Some people may disagree. I don't know where my line is yet. I'm waiting to see...
Today Linus Torvalds announced another Linux release candidate on the kernel mailing list. But he also highlighted "documentation updates" to address a new problem.
"The continued flood of AI reports has basically made the security list almost entirely unmanageable, with enormous duplication due to different people finding the same things with the same tools." (The new documentation says the security team has found "bugs discovered this way systematically surface simultaneously across multiple researchers, often on the same day.")
TORVALDS: People spend all their time just forwarding things to the right people or saying "that was already fixed a week/month ago" and pointing to the public discussion.
Which is all entirely pointless churn, and we're making it clear that AI-detected bugs are pretty much by definition not secret, and treating them on some private list is a waste of time for everybody involved — and only makes that duplication worse because the reporters can't even see each other's reports.
AI tools are great, but only if they actually help, rather than cause
unnecessary pain and pointless make-believe work. Feel free to use
them, but use them in a way that is productive and makes for a better
experience.
The documentation may be a bit less blunt than I am, but that's the
core gist of it.
The new documentation offers this overview. "It turns out that the majority of the bugs reported via the security team are just regular bugs that have been improperly qualified as security bugs due to a lack of awareness of the Linux kernel's threat model."
"So just to make it really clear," Torvalds said at the end of his post. "If you found a bug using AI tools, the chances are somebody else found it too.
"If you actually want to add value, read the documentation, create a patch
too, and add some real value on *top* of what the AI did. Don't be the
drive-by 'send a random report with no real understanding' kind of
person. Ok?"
Today former Google CEO Eric Schmidt "was booed multiple times," reports NBC News, "while discussing AI during a commencement speech at the University of Arizona."
Schmidt had started by remembering how computer platforms "gave everyone a voice" but also "degraded the public square... They rewarded outrage. They amplified our worst instincts. They coarsen the way we speak to each other, and that way, and in the way that we treat each other, is in the essence of a society." But then Schmidt "drew a parallel between artificial intelligence and the transformative impact of the computer — and was immediately met with boos."
"I know what many of you are feeling about that. I can hear you," Schmidt said, addressing the crowd as many continued to boo him. "There is a fear ... there is a fear in your generation that the future has already been written, that the machines are coming, that the jobs are evaporating, that the climate is breaking, that politics is fractured, and that you are inheriting a mess that you did not create, and I understand that fear."
He went on to argue that the future remains unwritten and that the graduating class of 2026 has real power to shape how AI develops — a claim that drew further disapproval from parts of the audience...
He closed by congratulating the class and offering them closing words. "The future is not yet finished. It is now your turn to shape it."
404 Media shared a video on YouTube of the crowd's booing — and what Schmidt said that provoked them:
SCHMIDT: "If you don't care about science that's okay because AI is going to touch everything else as well. [Very loud booing] Whatever path you choose, AI will become part of how work is done..."
"You can now assemble a team of AI agents to help you with the parts that you could never accomplish on your own. [Loud booing] When someone offers you a seat on the rocket ship, you do not ask which seat. You just get on... The rocket ship is here."
xAI has launched Grok Build, "a coding agent of its own to serve as competitor to its rivals' products, such as Anthropic's Claude Code," reports Engadget:
As Bloomberg notes, xAI has been trying to catch up to its rival companies like Anthropic and OpenAI. Elon Musk, the company's founder and CEO, previously admitted that it has fallen behind its competitors when it comes to coding. A couple of months ago, Musk said he was rebuilding xAI "from the foundations up" after several co-founders had left the company. One of the company's executives reportedly told staffers to work on getting Grok to match Claude's performance across various tasks.
More details from PCMag:
Grok Build is currently available in beta to those with a SuperGrok Heavy subscription, which starts at $300 per month. Just download it from the xAI website and log in. It's described as "a powerful new coding agent and CLI for professional software engineering and complex coding work." In its early version, xAI is seeking feedback and looking to fix any bugs... Only a few features have been highlighted, including a plan mode that lets you review, edit, and approve a plan before execution, and support for existing plug-ins and workflows.
"Every link leads to an entry that does not exist yet," explains the GitHub page for a Wikipedia-like site called Halupedia. "Until you click it, at which point an LLM pretends it has always existed and writes it for you, in the deadpan register of a 19th-century scholarly press..."
Every article is invented on demand. The footnotes are also lies... The hardest problem with an infinite, on-demand encyclopedia is internal contradiction... When the LLM writes an article, it is required to add a context="..." attribute on every <a> it inserts, summarising the future article it is linking to (e.g. context="19th-century clerk who formalized footnote drift, Pellbrick's mentor")... When that target article is later requested for the first time, the worker loads the accumulated hints and injects them into the system prompt as "PRIOR REFERENCES — these are CANON". The LLM is instructed that the encyclopedia is hallucinated and absurd, but it must not contradict itself.
Fast Company reports that Halupedia was created by software developer Bartlomiej Strama, who confessed in a Reddit comment that the site came about after a drunk night with a friend. In the week since launch, he says Halupedia has amassed more than 150,000 users."
Beyond indulging in silly alternate histories, what's the point of using Halupedia? Strama hinted at one larger purpose in a reply to a donor on his Buy Me a Coffee page: "Your contribution towards polluting LLM training data will surely benefit society!" he wrote.
The site is licensed as free software under the GPL-3.0 license.
Thanks to long-time Slashdot reader schwit1 for sharing the news.
Connexion
