OpenAI CEO Sam Altman took the stand Tuesday in Elon Musk's trial against the company, testifying that Musk repeatedly sought control of OpenAI before leaving in 2018. Altman said he opposed putting AI "under the control of any one person," while Musk's lawyer used a pointed cross-examination to attack Altman's trustworthiness. An anonymous reader shares updates from the testimony via the New York Times: Before Elon Musk left OpenAI in a power struggle in 2018, he wanted to merge the nonprofit artificial intelligence lab with Tesla, his electric car company. Mr. Musk and other OpenAI co-founders met several times to discuss the merger. OpenAI's chief executive, Sam Altman, was even offered a seat on Tesla's board of directors, according to a court document. But folding OpenAI into Tesla would have eliminated the lab's nonprofit status, and that, Mr. Altman said on the witness stand on Tuesday, was something he wanted to avoid. [...] "I believed that A.I. should not be under the control of any one person," Mr. Altman said. [...] Mr. Altman testified about his feud with Mr. Musk. He said he had become worried that Mr. Musk, who provided the early investment money for OpenAI, wanted to take control of the lab. He described what he called a "particularly harrowing moment" when his OpenAI co-founders asked Mr. Musk what would happen to his control of a potential for-profit when he died. Mr. Altman said Mr. Musk had replied that the control would pass to his children. "I was not comfortable with that," Mr. Altman said. When Mr. Musk lost a power struggle for control of the lab, he left, forcing Mr. Altman to find another big financial backer in Microsoft. But Mr. Altman ran into trouble in 2023 when OpenAI's board fired him because, as several of its members have testified in the trial, it didn't trust him. Steven Molo, Mr. Musk's lead lawyer, homed in on Mr. Altman's trustworthiness during an aggressive cross-examination. "Are you completely trustworthy?" Mr. Molo asked. "I believe so," Mr. Altman answered. After questioning Mr. Altman's trustworthiness for nearly 20 minutes, Mr. Molo turned to Mr. Altman's relationship with Mr. Musk. Mr. Altman said that after he met Mr. Musk in the mid-2010s, Mr. Musk had occasionally expressed concern about the dangers of A.I. But Mr. Musk spent far more time saying he was worried that companies like Google would get ahead in A.I. development, Mr. Altman said. (Mr. Musk testified in the trial that he had wanted to create OpenAI to prevent Google from controlling the technology.) Mr. Altman, the lawyer intimated, took advantage of Mr. Musk's concerns and was never sincere about his own A.I. fears. "Are you a person who just tells people things they want to hear whether those things are true or not?" Mr. Molo asked. The lawyer also questioned whether Mr. Atman, who became a billionaire through years of tech investments, was self-dealing through OpenAI. Mr. Molo showed a list of Mr. Altman's personal investments across a number of companies that stand to benefit from their association with OpenAI. They included Helion Energy, a start-up that has deals with Microsoft and OpenAI, and Cerebras, a chip maker in business with OpenAI. Mr. Molo asked if Mr. Altman, who is on OpenAI's board as well as its chief executive, would ever fire himself. "I have no plans to do that," Mr. Altman said. OpenAI's odd journey from nonprofit lab to what it is today -- a well-funded, for-profit company that is still connected to a nonprofit called the OpenAI Foundation with an endowment that could be worth more than $130 billion -- provided grist for Mr. Molo's questions about Mr. Altman's motivations. He implied that Mr. Altman could have continued to build OpenAI as a pure nonprofit. But the only way to build such a valuable charity was to raise billions through a for-profit venture, Mr. Altman responded. Still, the giant sums being raised appeared to upset Mr. Musk. In late 2022, according to court documents, Mr. Musk sent a text to Mr. Altman complaining that Microsoft was preparing to invest $10 billion in OpenAI. "This is a bait and switch," Mr. Musk said at the time. But Mr. Altman, under questioning from his own lawyers, said: "Every step of the way, I have done my best to maximize the value of the nonprofit. I would point out that there are not a lot of historical examples of a nonprofit at this scale." Before Altman took the stand, OpenAI board chair Bret Taylor continued his testimony that began on Monday. He said Elon Musk's 2024 bid to buy the company's assets appeared to conflict with his lawsuit and was rejected because the board did not believe OpenAI's mission should be controlled by one person. "We did not feel like it was appropriate for one person to control our mission," he said. Recap: Microsoft CEO Satya Nadella Testifies In OpenAI Trial (Day Nine) Sam Altman Had a Bad Day In Court (Day Eight) Sam Altman's Management Style Comes Under the Microscope At OpenAI Trial (Day Seven) Brockman Rebuts Musk's Take On Startup's History, Recounts Secret Work For Tesla (Day Six) OpenAI President Discloses His Stake In the Company Is Worth $30 Billion (Day Five) Musk Concludes Testimony At OpenAI Trial (Day Four) Elon Musk Says OpenAI Betrayed Him, Clashes With Company's Attorney (Day Three) Musk Testifies OpenAI Was Created As Nonprofit To Counter Google (Day Two) Elon Musk and OpenAI CEO Sam Altman Head To Court (Day One)

Read more of this story at Slashdot.

Wine's Wayland native driver has taken another step forward with now supporting the pointer warp "wp_pointer_warp_v1" protocol...

South Korea's presidential policy chief is calling for a "citizen dividend" that would return some AI-driven profits and tax revenue to the public. The Straits Times. From the report: Presidential policy chief Kim Yong-beom said in a Facebook post that a portion of the profits and tax revenue derived from the artificial intelligence boom "should be structurally returned to all citizens." That is because, Mr Kim argued, the economic gains from AI are based at least partly on industrial infrastructure built by the country over five decades. Mr Kim's comments come after tens of thousands of people gathered outside Samsung's main chip hub in April to demand employees get a greater share of AI profits. The company's labour union wants 15 per cent of operating profit handed to chip-division employees. The union has threatened an 18-day strike starting May 21. Workers have pointed to rising payouts at SK Hynix, which in 2025 agreed to allocate 10 per cent of its annual operating profit to a performance bonus pool, as evidence they deserve more pay. "Excess profits in the AI era are, by nature, concentrated," Mr Kim wrote. Memory companies, core engineers and asset holders are highly likely to receive substantial benefits, while much of the middle class may experience only indirect effects.

Read more of this story at Slashdot.

For those making use of OpenZFS on Linux or FreeBSD, OpenZFS 2.4.2 is out today as the newest stable release of this ZFS file-system implementation...

Instructure, the company behind the widely used Canvas learning platform, says it reached an agreement with the hackers who stole 3.5 terabytes of student and university data. The company says it received "digital confirmation" that the information was destroyed and that affected schools and students would not be extorted. The BBC reports: Paying cyber criminals goes against the advice of law enforcement agencies around the world, as it can fuel further attacks and offers no guarantee the data has been deleted. In previous cases, criminals have accepted ransom payments but lied about destroying stolen data, instead keeping it for resale. For example, when the notorious LockBit ransomware group was hacked by the National Crime Agency, police found stolen data had not been deleted even after payments had been made. Instructure said in a statement on its website that protecting students' and education staff data was its primary motivation. "While there is never complete certainty when dealing with cyber criminals, we believe it was important to take every step within our control to give customers additional peace of mind, to the extent possible," the company said. Instructure did not set out the terms of the agreement but said that it meant that: - the data was returned to the company - it received "digital confirmation of data destruction" - it had been informed that no Instructure customers would be extorted as a result of the incident - the agreement covers all affected customers, with no need for individuals to engage with the hackers

Read more of this story at Slashdot.

FreeBSD 15.0 had aimed to provide a KDE desktop install option from its text-based OS installer to make for a more compelling FreeBSD out-of-the-box desktop experience. That was then delayed to FreeBSD 15.1 but that didn't end up materializing. Now the KDE desktop install option is diverted to FreeBSD 15.2...

An anonymous reader quotes a report from the Financial Times (via Ars Technica): Amazon employees are using an internal AI tool to automate non-essential tasks in a bid to show managers they are using the technology more frequently. The Seattle-based group has started to widely deploy its in-house "MeshClaw" product in recent weeks, allowing employees to create AI agents that can connect to workplace software and carry out tasks on a user's behalf, according to three people familiar with the matter. Some employees said colleagues were using the software to automate additional, unnecessary AI activity to increase their consumption of tokens -- units of data processed by models. They said the move reflected pressure to adopt the technology after Amazon introduced targets for more than 80 percent of developers to use AI each week, and earlier this year began tracking AI token consumption on internal leader boards. "There is just so much pressure to use these tools," one Amazon employee told the FT. "Some people are just using MeshClaw to maximize their token usage." Amazon has told employees that the AI token statistics would not be used in performance evaluations. But several staff members said they believed managers were monitoring the data. "Managers are looking at it," said another current employee. "When they track usage it creates perverse incentives and some people are very competitive about it."

Read more of this story at Slashdot.

Google is teasing a new line of "Googlebook" laptops for this fall, powered by a new Android-and-ChromeOS-derived operating system that will run Chrome, Android apps, phone-connected apps and files, and deeply integrated Gemini features. The company says Chromebooks will continue "after the launch of Googlebook" and "...all Chromebooks will continue to receive support through their device's existing date commitment." The Verge reports: "We'll have more to share on the exact OS branding later this year," Peter Du of Google's global communications team tells The Verge. [...] Googlebooks will have a Magic Pointer feature that offers contextual suggestions whenever you shake your cursor and point it at something on the screen. Google's examples include setting up a meeting by pointing at a date in an email or selecting images of furniture and a living space to visualize them together. Beyond your mouse pointer, Googlebooks will also feature the custom AI-created widgets that Google is also debuting today for Android phones and Wear OS smartwatches. I don't know what kind of horrors people will be able to make into widgets, but Google gives the example of making one to organize your flights, hotel information, restaurant reservations, and another for creating a countdown timer for an upcoming family reunion. (It's always flights, hotels, and restaurants, isn't it?) While there are many outstanding questions to be answered about Googlebooks, the biggest and most obvious ones are what will these laptops look like, what chips will be in them, and what will they cost? We've got none of that so far. Google only has some initial renders of a mysterious Googlebook and the promise that it's working with Acer, Asus, Dell, HP, and Lenovo to make the first models. There are no model names. No specs. Nada. Google isn't even saying if the laptop in its renders is made by a partner or a tease of some first-party Pixel-like Googlebook to come or is just a cool mockup. The one distinct hardware feature shown, the bar of glowing Google-colored light, will be a signature of all Googlebooks. (Sure, bring on the RGB. Why not?)

Read more of this story at Slashdot.

Microsoft and G42's planned $1 billion AI data center in Kenya has stalled amid disagreements over power commitments, with President William Ruto saying the country would need to "switch off half the country" to support the project at full scale. Tom's Hardware reports: The project, announced in May 2024 during Ruto's visit to Washington, was supposed to bring a geothermal-powered data center to the Olkaria region in Kenya's Rift Valley. G42 was to lead construction, with the facility running Microsoft Azure in a new East Africa cloud region. The first phase targeted 100 megawatts of capacity and was expected to be operational by this year, with a long-term goal of scaling to 1 gigawatt. President Ruto isn't exaggerating about shutting off half the country's power. Kenya's total installed electricity capacity sits between 3,000 and 3,200 megawatts, and peak demand reached a record 2,444 megawatts in January, according to data from KenGen, the country's government-owned electricity producer. The full 1 gigawatt build would therefore have consumed roughly a third of the country's total capacity, and even the first 100 megawatts would have required a significant share of the Olkaria geothermal complex's output, which currently generates around 950MW across all its plants. John Tanui, principal secretary at Kenya's Ministry of Information, told Bloomberg that the project hasn't been withdrawn and that talks are continuing, adding that the "scale of the data center they [Microsoft] wanted to do still requires some structuring." A separate 60-megawatt project with local developer EcoCloud is also still under discussion. [...] Microsoft is spending $190 billion on capex in 2026, and the company adds approximately 1 gigawatt of data center capacity every three months globally. But power constraints are proving to be a universal bottleneck: nearly half of planned U.S. data center builds this year have been delayed or canceled due to shortages of electrical infrastructure.

Read more of this story at Slashdot.

The EU plans to target "addictive design" features on TikTok, Instagram, and other platforms, including endless scrolling, autoplay, push notifications, and recommendation loops that can steer children toward harmful content. European Commission President Ursula von der Leyen said new regulation could arrive later this year, alongside an EU age-verification app meant to make child-safety rules easier to enforce. CNBC reports: "We are taking action against TikTok and its addictive design -- endless scrolling, autoplay, and push notifications. The same applies to Meta, because we believe Instagram and Facebook are failing to enforce their own minimum age of 13," Von der Leyen said. "We are investigating platforms that allow children to go down 'rabbit holes' of harmful content -- such as videos that promote eating disorders or self-harm," she added. The EU's executive arm has also developed its own age verification app, which has the "highest privacy standards in the world," according to Von der Leyen. Member states will soon be able to integrate it into their digital wallets, and it can easily be enforced by online platforms. "No more excuses -- the technology for age-verification is available," the EU chief said. The EU Commission could have a legal proposal prepared as soon as the summer, as it awaits the advice and findings of its 'Special Panel of experts on Child Safety Online.'

Read more of this story at Slashdot.

Today's Patch Tuesday is a busier one than normal for the quarter. Both AMD and Intel have rolled out new updates for Linux customers among other security disclosures today. Thankfully though the vulnerabilities don't appear to be too widespread or impactful...

DXVK-NVAPI 0.9.2 is now available for this implementation of NVIDIA's NVAPI/NVOFAPI interfaces atop DXVK and VKD3D-Proton that is used in turn by Valve's Steam Play (Proton) for enhanced NVIDIA Linux gaming support...

An anonymous reader quotes a report from Reuters: EBay on Tuesday rejected a $56 billion takeover bid from the much smaller GameStop over financing doubts, calling the proposal "neither credible nor attractive." EBay, which has roughly four times GameStop's market value, also underscored that its turnaround efforts under CEO Jamie Iannone have boosted growth, with its stock returning 201% since Iannone took the position six years ago. "We have concluded that your proposal is neither credible nor attractive," eBay Chairman Paul Pressler said in a statement. "eBay's Board is confident the company, under its current management team, is well-positioned to continue to drive sustainable growth." He also pointed to concerns with GameStop's bid, including its financing, its impact on eBay's long-term growth and the leadership structure of a potentially combined company. Last week, GameStop's CEO Ryan Cohen delivered one of the most memorable CNBC interviews in recent memory... initially disinterested, then increasingly hostile, with little eye contact, few real answers to basic questions, and repeated robotic deflections to "check the website." It's worth a watch if you have a few extra minutes.

Read more of this story at Slashdot.

An IBM engineer posted the first set of patches enabling the Rust programming language support for the Linux kernel to be built on the s390 architecture...

D’ici quelques minutes, dans le cadre du The Android Show I/O Edition 2026, Google devrait apparemment abandonner l’appellation Chromebook au profit d’une nouvelle marque : Googlebook. XDA Developers a brièvement mis en ligne un article, désormais supprimé, qui apporte quelques indications... [Tout lire]

Les Zimaboard 2 sont des MiniPC montés en format carte. Equipées de processeurs Intel Twin Lake N150 et proposées avec de la mémoire LPDDR5x-4800 et un stockage eMMC, elles sont touchées comme les autres par une augmentation du prix des composants.

Équipées de 8 à 16 Go de mémoire vive en LPDDR5x-4800 et 32 à 64 Go de stockage de base en eMMC, les Zimaboard 2 subissent de plein fouet les hausses de tarif. J’imagine que même sur le segment processeur, la situation doit être compliquée. À moins qu’Intel ne fasse une exception pour la marque, la majorité des constructeurs de MiniPC m’expliquent que l’approvisionnement en puces Intel Twin Lake est devenu extrêmement compliqué. C’est peut-être ce que la marque indique être un « composant clé » de son produit.

Une hausse massive des ZimaBoard 2 

Toujours est-il que ces cartes particulières vont grimper de tarif. Dès le 25 mai, le modèle ZimaBoard 2 832 en 8 Go / 32 Go va passer de 279 $ HT à 339 $ HT… une hausse de 60$ assez douloureuse. Le modèle ZimaBoard 2 1664 en 16 Go / 64 Go passe de son côté de 349 à 399 $ HT. Une hausse paradoxalement moins sensible sur ce second modèle.

Si vous aviez l’ambition de vous construire un petit serveur ou autre avec ce type de carte, il ne vous reste donc que quelques jours. Toutes les commandes passées en amont de la hausse seront facturées au prix actuel. À partir du 25 mai à minuit pile (EDT et donc 6H00 du matin en France), les augmentations seront sévères.

Zimaboard 2 : un serveur maison de la taille d’un smartphone

Les Zimaboard 2 vont également augmenter de prix © MiniMachines.net. 2026

Marvel sombre, gastronomie italienne et divorce explosif au cinéma : Disney+ mai 2026 ratisse large, avec un calendrier dense le 12 mai et un duo Cumberbatch/Colman attendu dès le 13. Le tour d’horizon des sorties qui comptent. Le mois en chiffres Punisher (12 mai) Stanley Tucci S2 La Guerre des Roses Rivals S2 Calendrier complet […]

L'article Disney+ mai 2026 : Punisher, Stanley Tucci, Guerre des Roses a été publié en premier sur Bbox-Mag

The FCC has softened its ban on foreign-made consumer routers, allowing vendors to keep issuing broader software and firmware updates for devices already in use in the U.S. through at least January 2029. Dark Reading reports: Under the original FCC ruling, foreign manufacturers were permitted to provide only limited maintenance and security patches to US customers through March 2027. In a public note (PDF) on May 8, the FCC extended that deadline to at least January 2029 and also expanded the scope of permissible updates. The FCC will now allow foreign manufacturers to provide not just minor security fixes and changes, but also more major software and firmware updates that could affect router functionality, which previously required additional FCC review. The agency described the revisions as intended to ensure the continued safety of already deployed foreign-made consumer routers in the US. "The FCC likely issued this revision in response to the operational realities of network security and the slow pace of equipment replacement," says Jason Soroko, senior fellow at Sectigo. "Replacing millions of embedded devices across national infrastructure requires immense time and capital, and abandoning existing systems to a completely unpatched state would create an immediate vulnerability." "This waiver significantly alleviates the most pressing fears tied to the initial ban by preventing a sudden and dangerous security vacuum," added Soroko.

Read more of this story at Slashdot.