news aggregator

An anonymous reader quotes a report from CalMatters: A California attorney must pay a $10,000 fine for filing a state court appeal full of fake quotations generated by the artificial intelligence tool ChatGPT. The fine appears to be the largest issued over AI fabrications by a California court and came with a blistering opinion (PDF) stating that 21 of 23 quotes from cases cited in the attorney's opening brief were made up. It also noted that numerous out-of-state and federal courts have confronted attorneys for citing fake legal authority. "We therefore publish this opinion as a warning," it continued. "Simply stated, no brief, pleading, motion, or any other paper filed in any court should contain any citations -- whether provided by generative AI or any other source -- that the attorney responsible for submitting the pleading has not personally read and verified." The opinion, issued 10 days ago in California's 2nd District Court of Appeal, is a clear example of why the state's legal authorities are scrambling to regulate the use of AI in the judiciary. The state's Judicial Council two weeks ago issued guidelines requiring judges and court staff to either ban generative AI or adopt a generative AI use policy by Dec. 15. Meanwhile, the California Bar Association is considering whether to strengthen its code of conduct to account for various forms of AI following a request by the California Supreme Court last month. The Los Angeles-area attorney fined last week, Amir Mostafavi, told the court that he did not read text generated by the AI model before submitting the appeal in July 2023, months after OpenAI marketed ChatGPT as capable of passing the bar exam. A three-judge panel fined him for filing a frivolous appeal, violating court rules, citing fake cases, and wasting the court's time and the taxpayers money, according to the opinion. Mostafavi told CalMatters he wrote the appeal and then used ChatGPT to try and improve it. He said that he didn't know it would add case citations or make things up.

Read more of this story at Slashdot.

However, the changes could lead to more offshoring

In a surprise announcement on Friday, President Trump issued a proclamation on the H-1B visas many tech companies use to import qualified foreign workers. The headlines mentioning a $100,000 fee caused panic among many visa holders, leading the White House to issue a clarification: Only new applicants will cost their companies this exorbitant price.…

An anonymous reader shares a report: The iPhone 17 Pro and 17 Pro Max appear to provide little resistance to scratches and scuffs around the sharp edges of the camera bump. Tech blogger Zack Nelson demonstrates this weakness in a durability test on his JerryRigEverything YouTube channel, explaining that the anodized aluminium layer on the iPhone 17 Pro and 17 Pro Max "does not stick to corners very well" -- creating a weak point in the coating. This is a known issue with the electrochemical anodizing process, so it was a design decision Apple knowingly made. "For some reason, Apple didn't add a chamfer, fillet, or radius around the camera plateau, and I think it was intentional, so it looks cooler," Nelson says in the video. "But that decision to look cool out of the box is going to plague everyone who owns this phone down the road." The video shows that everyday objects, like a coin or house key carried in the same pocket as the iPhone 17 Pro, can chip away at the anodized coating around the sharp corners of the camera bump. However, that same mildly aggressive scratching on the flat surface of the camera plateau only produced dust that could be easily wiped away.

Read more of this story at Slashdot.

Not old enough to drink, old enough to be accused of causing millions in damage

A teen surrendered to Las Vegas police and was booked on suspicion of breaking into multiple Las Vegas casino networks in 2023, as part of a series of hacks attributed to Scattered Spider.…

The rise of self-driving cars could eventually cost many ride-hailing drivers their jobs -- and that's a big problem, Uber CEO Dara Khosrowshahi said. From a report: Khosrowshahi spoke about the issue onstage this month at a summit hosted by the "All-In" podcast, which posted a video of the conversation on Wednesday. At the summit, Khosrowshahi was asked about concerns that gig workers, who have played a key role in Uber's development, will eventually lose their jobs as self-driving cars become more prevalent. The Uber CEO said he expects human drivers to continue working alongside self-driving cars in Uber's network in the coming years. "For the next five to seven years, we're going to have more human drivers and delivery people, just because we're going so quickly," Khosrowshahi said. "But, I think, 10 to 15 years from now, this is going to be a real issue," he said about drivers losing their jobs.

Read more of this story at Slashdot.

Spooky season is nearly here. Want to be scared? There are fresh betas to try

Two of the biggest names in fixed-release distros are nearly finished and ready to drop. You can taste them now, but they're not fully baked yet.…

Microsoft is working on bringing support for setting a video as your desktop wallpaper on Windows 11. From a report: Hidden in the latest Windows 11 preview builds, the feature lets you set an MP4, MOV, AVI, WMV, M4V, or MKV file as your wallpaper, which will play the video whenever you view the desktop. For many years, users have wanted the ability to set a video as a desktop background. It's a feature that many Linux distributions support, and macOS also supports the ability to set a moving background as your lock screen. Windows Vista did support setting videos as your wallpaper, but only as part of the Ultimate SKU via a feature called DreamScene.

Read more of this story at Slashdot.

Cloud and apps chiefs step up as Safra Catz moves upstairs. Larry remains Larry

Oracle on Monday named Clay Magouyrk and Mike Sicilia as its new co-chief executives, replacing Safra Catz, who will shift into the role of executive vice chair of the board after more than a decade as top dog.…

Safety watchdog doubts SpaceX can ready HLS in time for 2027 Artemis mission

NASA's Aerospace Safety Advisory Panel (ASAP) has cast doubt on SpaceX's Starship making the 2027 Artemis III lunar landing deadline.…

Nvidia will invest up to $100 billion in OpenAI as the AI lab builds data centers requiring 10 gigawatts of power capacity. The 10-gigawatt deployment equals 4 to 5 million GPUs -- the same number Nvidia will ship globally this year. Building one gigawatt of data center capacity costs $50 to $60 billion, including approximately $35 billion for Nvidia chips and systems. The first phase begins in the second half of 2026 using Nvidia's next-generation Vera Rubin systems. The investment adds Nvidia to OpenAI's investor roster alongside Microsoft, SoftBank, and Thrive Capital at a $500 billion valuation. Nvidia CEO Jensen Huang described the investment as "additive to everything that's been announced and contracted."

Read more of this story at Slashdot.

An anonymous reader shares a report: Venture capitalists in clean tech are starting to say out loud what they've suspected for a while: China's dominance has left key sectors in the West uninvestable. A group of eight VCs from Western firms agreed to share with Bloomberg the details of a July road trip across China during which they visited factories, spoke with startup investors, and interviewed founders of companies. They knew China had raced ahead in sectors like batteries and "everything around energy," but seeing how big the gap was firsthand left them wondering how European and North American competitors can even survive, says Talia Rafaeli, a former investment banker at both Goldman Sachs and Barclays who's now a partner at Kompas VC. "Everyone needs to take this kind of trip," she said.

Read more of this story at Slashdot.

Subscribing to an online service is often as easy as a click of a button. Is it illegal if it takes a maze of clicks to cancel? That issue is at the heart of a civil trial beginning this week that will scrutinize the tactics Amazon uses to entice consumers to sign up for its signature Prime service -- and to steer them away from leaving. WSJ: The Federal Trade Commission alleges the online giant has duped nearly 40 million customers, in violation of consumer-protection laws. It is seeking civil penalties, refunds to consumers and a court order prohibiting Amazon from using subscription practices that could confuse or deceive customers. The case, which will unfold in a Seattle courtroom, is a top test of the agency's enforcement campaign against allegedly deceptive digital subscription practices. Amazon's Prime membership, the largest paid subscription program in the world with at least 200 million users, has helped the company become an integral part of consumers' shopping habits. The FTC, which sued Amazon in 2023, alleges the company tricked people into signing up for the service without their knowledge or consent, including by obscuring details about billing and the terms of free trials. It says Amazon created a labyrinth to make it hard to cancel, which the company dubbed "Iliad," a reference to Homer's epic about the long, arduous Trojan War. The FTC says Amazon required customers to navigate four webpages and chose from 15 options to cancel a Prime membership. The company streamlined the process in April 2023, ahead of the filing of the criminal complaint. The FTC won an initial pretrial victory last week when a federal judge ruled that Amazon did violate consumer-protection laws by taking Prime members' billing information before disclosing the terms of the membership. But he said jurors still would have to consider whether the customers gave their consent to enroll and whether Amazon provided a simple cancellation mechanism.

Read more of this story at Slashdot.

Long-time contributor Ellen Davis steps down after GitHub access shake-up and governance dispute

A decade-long RubyGems maintainer, Ellen Davis (also known as duckinator), has resigned from Ruby Central following what she described as a "hostile takeover" of the open source project.…

Major U.S. corporations are mandating more office time but seeing minimal compliance changes. Companies now require 12% more in-office days than in early 2024, according to Work Forward data tracking 9,000 employers. Yet Americans continue working from home approximately 25% of the time, unchanged from 2023, Stanford economist Nicholas Bloom's monthly survey of 10,000 Americans shows. The New York Times ordered opinion and newsroom staff to four days weekly starting November. Microsoft mandates three days beginning February for Pacific Northwest employees. Paramount and NBCUniversal gave staff ultimatums: commit to five and four days respectively or take buyouts. Amazon faced desk and parking shortages after its full-time mandate, temporarily backpedaling in Houston and New York. Nearly half of senior managers would accept pay cuts to work remotely, a BambooHR survey of 1,500 salaried employees found.

Read more of this story at Slashdot.

Ratings agency points out there's a risk of relying on a small number of buyers

Ratings agency Moody's has pointed to the dangers inherent in Oracle's $300 billion agreement with OpenAI - one of the deals contributing to a staggering $455 billion pipeline of obligations for Big Red's cloud infrastructure.…

The US will charge companies $100,000 for each new H-1B visa starting February 2026 under Project Firewall. According to a new analysis, the fee exceeds average H-1B salaries at firms like TCS where engineers earn $105,000 annually. Previous visa costs ranged from $2,000 to $33,000. Indians hold an estimated 70% of H-1B visas. The fee eliminates five to six years of profit per engineer. Typical engineers deployed to American client sites generate $150,000 to $200,000 in annual billings at 10% operating margins, producing $15,000 to $20,000 in yearly profit. J.P. Morgan states the move "prices out the utility of H-1B as a source of labor supply." But it might not be bad for the IT giants. Major Indian IT firms derive only 0.2% to 2.2% of their workforce from H-1B approvals after years of reducing visa dependence, according to India Dispatch. New approvals alone account for under 0.4% of headcount. Morgan Stanley estimates companies could offset 60% of the financial impact through increased offshoring and selective price increases. The net damage to operating profit would stay contained at around 50 basis points or a 3% to 4% hit to earnings spread across the renewal cycle. Companies plan to accelerate geographic arbitrage by routing more work to India, Canada, and Latin America. Firms can maintain their existing visa holder base while letting normal turnover occur over three to six years.

Read more of this story at Slashdot.

Thinnest yet still fixable, though not without effort

iFixit has given Apple's slimline new smartphone, the iPhone Air, a thumbs-up for repairability, praising its easy access to key components, despite being the thinnest handset Cupertino has built so far.…

Airport staff revert to manual ops as travellers urged to use self-service check-in where possible

The EU's cybersecurity agency today confirmed that ransonmware is the cause of continued disruption blighting major airports across Europe.…

Protected content in some Blu-ray and DVD applications broken

Microsoft has added another entry to its growing list of problematic updates in the Windows Hall of Shame, this time causing Digital TV and Blu-ray applications to stutter and freeze when playing protected content.…

It's not just glitches at the launch of the Meta Ray-Ban Display smart glasses... The New York Times remains skeptical of its market share: [Meta's] smart glasses remain a niche. As of February, Meta had sold about two million of its $300 Ray-Ban Meta camera glasses since their 2023 debut, and it hopes to sell 10 million annually by the end of 2026, which is a tiny amount for a company this size. In the last decade, Meta has spent over $100 billion on its virtual and augmented reality division, which includes its smart glasses and is not profitable. Last quarter, the division reported a $4.5 billion loss, nearly the same as a year ago. "Meta's Smart Glasses Might Make You Smarter. They'll Certainly Make You More Awkward," joked a recent Wired headline. But the Wall Street Journal does report there's "a growing group of blind users... finding the devices to be more of a life-enhancing tool than a cool accessory." Jonathan Mosen, executive director at the nonprofit National Federation of the Blind said he'd like to see Meta continue to invest in the glasses. "It's giving significant accessibility benefits at a price point people can afford." He has used them a few times to record video of ride-share drivers refusing to give him and his wife a ride because she travels with a guide dog. Denying rides to people with service animals is illegal in many countries, including the U.S. Another concern for blind users is that AI assistants in general are prone to making errors, or so-called hallucinations, which may not be apparent. Aaron Preece, who is blind and editor in chief of American Foundation for the Blind's AccessWorld magazine, said Meta's glasses recently failed to correctly read the number on the door to his home. "I just can't trust it," he said. "It's more of a novelty than something I'd use on a day-to-day basis." When it comes to innovative technology, CNET seems more excited about Meta's display-controlling "neural wristband" accessory. Instead of camera-based hand tracking, these muscle-sensing bands "can register gestural moves like pinches, taps, thumb swipes, and maybe even typing over time..."

Read more of this story at Slashdot.