call_end

    • Ar chevron_right

      OpenAI and Nvidia’s $100B AI plan will require power equal to 10 nuclear reactors

      news.movim.eu / ArsTechnica • 22 September

    On Monday, OpenAI and Nvidia jointly announced a letter of intent for a strategic partnership to deploy at least 10 gigawatts of Nvidia systems for OpenAI's AI infrastructure, with Nvidia planning to invest up to $100 billion as the systems roll out. The companies said the first gigawatt of Nvidia systems will come online in the second half of 2026 using Nvidia's Vera Rubin platform.

    "Everything starts with compute," said Sam Altman, CEO of OpenAI, in the announcement. "Compute infrastructure will be the basis for the economy of the future, and we will utilize what we're building with NVIDIA to both create new AI breakthroughs and empower people and businesses with them at scale."

    The 10-gigawatt project represents an astoundingly ambitious and as-yet-unproven scale for AI infrastructure. Nvidia CEO Jensen Huang told CNBC that the planned 10 gigawatts equals the power consumption of between 4 million and 5 million graphics processing units, which matches the company's total GPU shipments for this year and doubles last year's volume. "This is a giant project," Huang said in an interview alongside OpenAI CEO Sam Altman and President Greg Brockman.

    Read full article

    Comments

    • Ar chevron_right

      Anti-vaccine groups melt down over reports RFK Jr. to link autism to Tylenol

      news.movim.eu / ArsTechnica • 22 September

    Health Secretary and anti-vaccine activist Robert F. Kennedy Jr. is reportedly poised to announce that use of Tylenol (aka acetaminophen, paracetamol) during pregnancy is linked to autism—an unproven assertion that has sent Kennedy's anti-vaccine allies into a rage.

    "We didn’t wait 20 years for Bobby to finally speak and then get served Tylenol as an answer," anti-vaccine group Georgia Coalition for Vaccine Choice wrote in an unhinged Facebook post on Monday morning. "If that's all we hear - is that the end? Not thimerosal. Not aluminum. Not MMR. Not Hep B. Not the insane schedule pushed after pharma got liability protection. Are we supposed to just forget?"

    Children's Health Defense (CHD)—the anti-vaccine group founded by Kennedy—even got in on the backlash, retweeting a post on Monday about parents who falsely blame vaccines for their children's neurological condition, with the statement: "THIS WAS NOT CAUSED BY TYLENOL."

    Read full article

    Comments

    • Ar chevron_right

      DeepMind AI safety report explores the perils of “misaligned” AI

      news.movim.eu / ArsTechnica • 22 September • 1 minute

    Generative AI models are far from perfect, but that hasn't stopped businesses and even governments from giving these robots important tasks. But what happens when AI goes bad? Researchers at Google DeepMind spend a lot of time thinking about how generative AI systems can become threats, detailing it all in the company's Frontier Safety Framework. DeepMind recently released version 3.0 of the framework to explore more ways AI could go off the rails, including the possibility that models could ignore user attempts to shut them down.

    DeepMind's safety framework is based on so-called "critical capability levels" (CCLs). These are essentially risk assessment rubrics that aim to measure an AI model's capabilities and define the point at which its behavior becomes dangerous in areas like cybersecurity or biosciences. The document also details the ways developers can address the CCLs DeepMind identifies in their own models.

    Google and other firms that have delved deeply into generative AI employ a number of techniques to prevent AI from acting maliciously. Although calling an AI "malicious" lends it intentionality that fancy estimation architectures don't have. What we're talking about here is the possibility of misuse or malfunction that is baked into the nature of generative AI systems.

    Read full article

    Comments

    • Ar chevron_right

      Here’s how potent Atomic credential stealer is finding its way onto Macs

      news.movim.eu / ArsTechnica • 22 September

    Ads prominently displayed on search engines are impersonating a wide range of online services in a bid to infect Macs with a potent credential stealer, security companies have warned. The latest reported target is users of the LastPass password manager.

    Late last week , LastPass said it detected a widespread campaign that used search engine optimization to display ads for LastPass macOS apps at the top of search results returned by search engines, including Google and Bing. The ads led to one of two fraudulent GitHub sites targeting LastPass, both of which have been taken down. The pages provided links promising to install LastPass on MacBooks. In fact, they installed a macOS credential stealer known as Atomic Stealer, or alternatively, Amos Stealer.

    Dozens targeted

    “We are writing this blog post to raise awareness of the campaign and protect our customers while we continue to actively pursue takedown and disruption efforts, and to also share indicators of compromise (IoCs) to help other security teams detect cyber threats,” LastPass said in the post.

    Read full article

    Comments

    • Ar chevron_right

      Three crashes in the first day? Tesla’s robotaxi test in Austin.

      news.movim.eu / ArsTechnica • 22 September

    These days, Austin, Texas feels like ground zero for autonomous cars. Although California was the early test bed for autonomous driving tech, the much more permissive regulatory environment in the Lone Star State, plus lots of wide, straight roads and mostly good weather, ticked enough boxes to see companies like Waymo and Zoox set up shop there. And earlier this summer, Tesla added itself to the list . Except things haven't exactly gone well.

    According to Tesla's crash reports, spotted by Brad Templeton over at Forbes , the automaker experienced not one but three crashes, all apparently on its first day of testing on July 1. And as we learned from Tesla CEO Elon Musk later in July during the ( not-great ) quarterly earnings call, by that time, Tesla had logged a mere 7,000 miles in testing.

    By contrast, Waymo's crash rate is more than two orders of magnitude lower, with 60 crashes logged over 50 million miles of driving . (Waymo has now logged more than 96 million miles.)

    Read full article

    Comments

    • Ar chevron_right

      Our fave Star Wars duo is back in Mandalorian and Grogu teaser

      news.movim.eu / ArsTechnica • 22 September

    Disney CEO Bob Iger has been under fire for several days now for pulling Jimmy Kimmel Live off the air "indefinitely," with Disney+'s cancellation page actually crashing a couple of times from all the traffic as people rushed to make their displeasure known. So what better time for the studio to release the first teaser trailer for The Mandalorian and Grogu , a feature film spinoff from its megahit Star Wars series The Mandalorian ? Grogu and Mando, together again on an exciting space adventure, will certainly be a crowd-pleaser.

    Grogu (aka Baby Yoda) won viewers' hearts from the moment he first appeared onscreen in the first season of The Mandalorian , and the relationship between the little green creature and his father-figure bounty hunter has only gotten stronger. With the 2023 Hollywood strikes delaying production on S4 of the series, director Jon Favreau got the green light to make this spinoff film.

    Per the official logline:

    Read full article

    Comments

    • Ar chevron_right

      F1 in Azerbaijan: This sport is my red flag

      news.movim.eu / ArsTechnica • 22 September • 1 minute

    Almost a decade old now, the Azerbaijan Grand Prix is part of the modern breed of Formula 1 street circuits. Set on the shores of the Caspian Sea, the anti-clockwise layout manages to combine some of F1's highest top speeds with a rather fiddly section through the old city, all of it lined with walls and barriers to punish mistakes. It's a low-downforce track, with mostly slow corners, similar to Montreal and Monza.

    Despite the questionable record of the hosting country—something all too many F1 races can also offer, including now the three held here in the US—I have a soft spot for watching Baku, with its interesting mix of old and new architecture, and it usually puts on a good race. I particularly love the helicopter and drone shots that follow the action with a God's eye view, giving you a glimpse behind the building facades and into this city by the sea.

    Friday's practice sessions gave the Ferrari-supporting Tifosi something to look forward to, with Lewis Hamilton and Charles Leclerc topping the times for FP2. Hamilton, who has looked at sea this year after his switch to Ferrari, was made the bubbliest we've seen him all season when being interviewed on Friday, and his teammate Leclerc has four pole positions to his name at Baku, further stoking the hope.

    Read full article

    Comments

    • Ar chevron_right

      What climate targets? Top fossil fuel producing nations keep boosting output

      news.movim.eu / ArsTechnica • 22 September

    The last two years have witnessed the hottest one in history, some of the worst wildfire seasons across Canada, Europe and South America and deadly flooding and heat waves throughout the globe. Over that same period, the world’s largest fossil fuel producers have expanded their planned output for the future, setting humanity on an even more dangerous path into a warmer climate.

    Governments now expect to produce more than twice as much coal, oil and gas in 2030 as would be consistent with the goals of the Paris Agreement, according to a report released Monday . That level is slightly higher than what it was in 2023, the last time the biennial Production Gap report was published.

    The increase is driven by a slower projected phaseout of coal and higher outlook for gas production by some of the top producers, including China and the United States.

    Read full article

    Comments

    • Ar chevron_right

      A history of the Internet, part 3: The rise of the user

      news.movim.eu / ArsTechnica • 22 September • 1 minute

    Larry Page and Sergey Brin met each other at a graduate student orientation at Stanford in 1996. Both were studying for their PhDs in computer science, and both were interested in analyzing large sets of data. Because the web was growing so rapidly, they decided to start a project to improve the way people found information on the Internet.

    They weren’t the first to try this. Hand-curated sites like Yahoo had already given way to more algorithmic search engines like AltaVista and Excite, which both started in 1995. These sites attempted to find relevant webpages by analyzing the words on every page.

    Page and Brin’s technique was different. Their “BackRub” software created a map of all the links that pages had to each other. Pages on a given subject that had many incoming links from other sites were given a higher ranking for that keyword. Higher-ranked pages could then contribute a larger score to any pages they linked to. In a sense, this was a like a crowdsourcing of search: When people put “This is a good place to read about alligators” on a popular site and added a link to a page about alligators, it did a better job of determining that page’s relevance than simply counting the number of times the word appeared on a page.

    Read full article

    Comments