Linux fréttir

Datacenters galore, plus some vague cooperation on AI, nuclear, quantum, and more

America and the UK have announced a $42 billion (£31 billion) trade pact, funded by Microsoft, Google, and others, that predicts bit barns will spring up over Britain's green and pleasant Land. But there's a lot more than money involved.…

First up: $41M to use human annotators to label all that unstructured military data. What could go wrong?

Data curation firm Scale AI has partnered with the Pentagon to deploy its AI on Top Secret networks - a move its interim CEO says is necessary if the US wants AI to be useful for national security.…

Two Xpeng AeroHT flying cars collided during a rehearsal for the Changchun Air Show in China, with one vehicle catching fire upon landing. While the company reported no serious injuries, CNN reported one person was injured in the crash. The BBC reports: Footage on Chinese social media site Weibo appeared to show a flaming vehicle on the ground which was being attended to by fire engines. One vehicle "sustained fuselage damage and caught fire upon landing," Xpeng AeroHT said in a statement to CNN. "All personnel at the scene are safe, and local authorities have completed on-site emergency measures in an orderly manner," it added. The electric flying cars take off and land vertically, and the company is hoping to sell them for around $300,000 each. In January, Xpeng claimed to have around 3,000 orders for the vehicle. [...] It has said it wants to lead the world in the "low-altitude economy."

Read more of this story at Slashdot.

House of Zen promises 3.5x improvement in inference and 3x uplift in training perf over last-gen software

AMD closed the performance gap with Nvidia's Blackwell accelerators with the launch of the MI355X this spring. Now the company just needs to overcome Nvidia's CUDA software advantage and make that perf more accessible to developers. …

Microsoft is now prioritizing Anthropic's Claude 4 over OpenAI's GPT-5 in Visual Studio Code's auto model feature, signaling a quiet but clear shift in preference. The Verge reports: "Based on internal benchmarks, Claude Sonnet 4 is our recommended model for GitHub Copilot," said Julia Liuson, head of Microsoft's developer division, in an internal email in June. While that guidance was issued ahead of the GPT-5 release, I understand Microsoft's model guidance hasn't changed. Microsoft is also making "significant investments" in training its own AI models. "We're also going to be making significant investments in our own cluster. So today, MAI-1-preview was only trained on 15,000 H100s, a tiny cluster in the grand scheme of things," said Microsoft AI chief Mustafa Suleyman, in an employee-only town hall last week. Microsoft is also reportedly planning to use Anthropic's AI models for some features in its Microsoft 365 apps soon. The Information reports that the Microsoft 365 Copilot will be "partly powered by Anthropic models," after Microsoft found that some of these models outperformed OpenAI in Excel and PowerPoint.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: Like the rest of its Big Tech cadre, Google has spent lavishly on developing generative AI models. Google's AI can clean up your text messages and summarize the web, but the company is constantly looking to prove that its generative AI has true intelligence. The International Collegiate Programming Contest (ICPC) helps make the point. Google says Gemini 2.5 participated in the 2025 ICPC World Finals, turning in a gold medal performance. According to Google this marks "a significant step on our path toward artificial general intelligence." Every year, thousands of college-level coders participate in the ICPC event, facing a dozen deviously complex coding and algorithmic puzzles over five grueling hours. This is the largest and longest-running competition of its type. To compete in the ICPC, Google connected Gemini 2.5 Deep Think to a remote online environment approved by the ICPC. The human competitors were given a head start of 10 minutes before Gemini began "thinking." According to Google, it did not create a freshly trained model for the ICPC like it did for the similar International Mathematical Olympiad (IMO) earlier this year. The Gemini 2.5 AI that participated in the ICPC is the same general model that we see in other Gemini applications. However, it was "enhanced" to churn through thinking tokens for the five-hour duration of the competition in search of solutions. At the end of the time limit, Gemini managed to get correct answers for 10 of the 12 problems, which earned it a gold medal. Only four of 139 human teams managed the same feat. "The ICPC has always been about setting the highest standards in problem-solving," said ICPC director Bill Poucher. "Gemini successfully joining this arena, and achieving gold-level results, marks a key moment in defining the AI tools and academic standards needed for the next generation." Gemini's solutions are available on GitHub.

Read more of this story at Slashdot.

Governments worldwide are implementing heat protection laws as 2.4 billion workers face extreme temperature exposure and 19,000 die annually from heat-related workplace injuries, according to a World Health Organization and World Meteorological Organization report. Japan imposed $3,400 fines for employers failing to provide cooling measures when wet-bulb temperatures reach 28C. Singapore mandated hourly temperature sensors at large outdoor sites and requires 15-minute breaks every hour at 33C wet-bulb readings. Southern European nations ordered afternoon work stoppages this summer when temperatures exceeded 115F across Greece, Italy and Spain. The United States lacks federal heat standards; only California, Colorado, Maryland, Minnesota, Nevada, Oregon and Washington have state-level protections. Boston passed requirements for heat illness prevention plans on city projects. Enforcement remains inconsistent -- Singapore inspectors found nearly one-third of 70 sites violated the 2023 law. Texas and Florida prohibit local governments from mandating rest and water breaks.

Read more of this story at Slashdot.

The ability of AI displace humans at various tasks is accelerating quickly, Anthropic CEO Dario Amodei said at an Axios event on Wednesday. From the report: Amodei and others have previously warned of the possibility that up to half of white-collar jobs could be wiped out by AI over the next five years. The speed of that displacement could require government intervention to help support the workforce, executives said. "As with most things, when an exponential is moving very quickly, you can't be sure," Amodei said. "I think it is likely enough to happen that we felt there was a need to warn the world about it and to speak honestly." Amodei said the government may need to step in and support people as AI quickly displaces human work.

Read more of this story at Slashdot.

You didn't really trust the crims to keep their word, did you?

Spiders don't change their stripes. Despite gang members' recent retirement claims, Scattered Spider hasn't exited the cybercrime business and instead has shifted focus to the financial sector, with a recent digital intrusion at a US bank.…

Light pollution now doubles every eight years globally as LED adoption accelerates artificial brightness worldwide. A recent study measured 10% annual growth in light pollution from 2011 to 2022. Northern Chile's Atacama Desert remains one of the few Bortle Scale 1 locations -- the darkest rating for astronomical observation -- though La Serena's population has nearly doubled in 25 years. The region hosts major observatories including the Vera C. Rubin Observatory at Cerro Pachon. Satellite constellations pose additional challenges: numbers have increased from hundreds decades ago to 12,000 currently operating satellites. Astronomers predict 100,000 or more satellites within a decade. Chile faces pressure from proposed mining operations including the 7,400-acre INNA green-hydrogen facility near key astronomical sites despite national laws limiting artificial light from mining operations that generate over half the country's exports.

Read more of this story at Slashdot.

Carefully crafted response makes no mention of whether DOGE employees duplicated critical database

The Social Security Administration (SSA) has disputed a whistleblower's allegations that claimed DOGE made an unauthorized, unsecured copy of a critical database - but it's what the denial doesn't say that speaks volumes. …

Crew will have to wait a little longer for science supplies, spares, and 'fun food'

NASA has delayed a supply delivery to the International Space Station (ISS) after the engines of Northrop Grumman's Cygnus XL cargo spacecraft did not perform as expected during an orbit-raising burn.…

AI models often produce false outputs, or "hallucinations." Now OpenAI has admitted they may result from fundamental mistakes it makes when training its models. The Register: The admission came in a paper [PDF] published in early September, titled "Why Language Models Hallucinate," and penned by three OpenAI researchers and Santosh Vempala, a distinguished professor of computer science at Georgia Institute of Technology. It concludes that "the majority of mainstream evaluations reward hallucinatory behavior." The fundamental problem is that AI models are trained to reward guesswork, rather than the correct answer. Guessing might produce a superficially suitable answer. Telling users your AI can't find an answer is less satisfying. As a test case, the team tried to get an OpenAI bot to report the birthday of one of the paper's authors, OpenAI research scientist Adam Tauman Kalai. It produced three incorrect results because the trainers taught the engine to return an answer, rather than admit ignorance. "Over thousands of test questions, the guessing model ends up looking better on scoreboards than a careful model that admits uncertainty," OpenAI admitted in a blog post accompanying the release.

Read more of this story at Slashdot.

If global temperatures continue rising, virtually all the corals in the Atlantic Ocean will stop growing and could succumb to erosion by the end of the century, a new study finds. From a report: The analysis of over 400 existing coral reefs across the Atlantic Ocean estimates that more than 70 percent of the region's reefs will begin dying by 2040 even under optimistic climate warming scenarios. And if the planet exceeds 2 degrees Celsius of warming above preindustrial temperatures by the end of the century, 99 percent of corals in the region would meet this fate. Today, the planet has warmed about 1.3 degrees Celsius over preindustrial temperatures. The implications are grave. Corals act as the fundamental building blocks of reefs, providing habitat for thousands of species of fish and other marine life. They are also bulwarks that break up waves and help protect shorelines from rising sea levels. A quarter of all ocean life depends on coral reefs and over a billion people worldwide benefit from them, according to the National Oceanic and Atmospheric Administration.

Read more of this story at Slashdot.

American businesses join Win 10 upgrade train, consumers happy to sit on the platform

Wolrd War Fee The US PC industry is suffering from inventory indigestion caused by resellers over-ordering hardware to avoid Donald Trump's expected import taxes on China-made kit.…

An anonymous reader shares a report: After years of dismissing the idea of putting a touchscreen on a MacBook, it seems Apple may have finally caved. Its MacBook Pro overhaul in 2026 is now expected to be the first-ever MacBook to feature a touchscreen display, according to a report from supply chain analyst Ming-Chi Kuo on X. The change will reportedly affect Apple's next-generation MacBook Pro, which could feature an OLED display and "incorporate a touch panel using on-cell touch technology." The OLED MacBook Pro isn't expected to enter production until late 2026, and before then, Apple is expected to launch the M5 MacBook Pro in early 2026.

Read more of this story at Slashdot.

And we're paying for it piecemeal through the software, services, and devices we buy

Tech analysts expect worldwide spending on AI to hit nearly $1.5 trillion in 2025, including $268 billion on optimized servers. These investments will also soon appear in even more consumer products.…

An anonymous reader shares a report: Business Insider has told journalists they can use AI to create first drafts of stories and suggested it won't notify readers that AI was used, according to Status, a newsletter covering the media industry. The policy makes the outlet one of the first to formally allow such extensive use of the technology. The AI guidelines were reportedly circulated in an internal memo from editor-in-chief Jamie Heller on Thursday. The policy authorized journalists to deploy AI "like any other tool" for tasks like research and image editing, Status reported.

Read more of this story at Slashdot.

But will the ISS still be there to host it?

Axiom Space and Spacebilt have announced plans to add optically interconnected Orbital Data Center (ODC) infrastructure to the International Space Station (ISS).…

Scripted TV production peaked in 2022 at 599 shows and has declined since, according to FX's research division tracking. New prestige series have dropped sharply while streaming platforms prioritize returning shows over new development. Netflix has shifted majority output to unscripted content including docuseries and reality programming since 2018. YouTube leads streaming viewership ahead of Netflix, Paramount+, and Hulu. Free ad-supported platforms YouTube, Tubi and Roku Channel continue gaining market share. Subscription prices across major streaming services have increased while scripted content volume decreased. Second season of Severance cost $200 million. Fourth season of Stranger Things reached $270 million in production expenses.

Read more of this story at Slashdot.