At Meta Connect 2025, live demos of Meta's AI-powered smart glasses repeatedly failed, including a video call glitch and a cooking assistant mix-up. Meta's CTO attributed the issues to a 'demo fail' caused by an accidental internal network overload, while acknowledging a new bug. The incidents raise questions about the reliability of AI features in consumer devices.
Trifluoroacetic acid (TFA), a byproduct of PFAS breakdown and fluorinated gases, has been detected in 94% of European tap water samples and is increasingly found in rainwater worldwide. Despite industry pushback, scientists warn of its potential health risks and the need for regulatory scrutiny.
A recent scientific simulation of a runaway greenhouse effect paints a stark picture of a potential climate tipping point. The study highlights how increased greenhouse gas concentrations could trigger an irreversible cascade, leading to extreme planetary warming. Researchers emphasize the urgent need for climate action to avert such a scenario.
Medical researchers are exploring three distinct paths to create artificial hearts, including robotic implants, decellularized donor organs, and 3D-printed scaffolds. These efforts could address donor shortages and high costs, though early-stage experiments have yet to achieve full pumping function.
Daimler, the world's largest truck maker, is shifting to 3D printing for spare parts, reducing shipping and storage costs. The move highlights a broader industry trend as costs for additive manufacturing fall and other automakers expand their own 3D printing capabilities.
Astrophysicist Neil deGrasse Tyson appeared on 'The Late Show with Stephen Colbert' to critique the science behind 'Dune: Part Two,' focusing on the film's depiction of sand thumping and worm locomotion. He argues that sand absorbs sound rather than producing a thump and that worm-like creatures move in S-curves, not straight lines. Tyson acknowledges the need to let go of such inaccuracies for the sake of the story.
Li Zifeng, an oil drilling professor at China's Yanshan University, claims to have disproven Einstein's theory of relativity after two decades of work, but faces institutional rejection and online mockery. His assertions, reported by the state-run Global Times, argue the theory lacks practical proof, yet experts remain skeptical.
Researchers at the University of Alabama at Birmingham have mapped the last unknown protein structure of HIV-1, the cytoplasmic tail of gp41, completing the virus's full structural picture. This breakthrough could guide new therapies that block viral replication, though further research is needed before clinical applications.
Geoffrey Hinton, a Nobel laureate and AI pioneer, suggests that to prevent superintelligent AI from harming humans, we should program it with maternal instincts. Speaking at a Las Vegas conference, he argued that AI will naturally seek control and survival, and that motherly care is the only model for a more intelligent entity being controlled by a less intelligent one. Critics note the unscientific basis of 'maternal instinct' and the need to address more immediate AI risks like bias.
Philadelphia Sheriff Rochelle Bilal’s reelection campaign admitted that dozens of positive news articles posted on its website were fabricated using ChatGPT. The admission follows an investigation by the Philadelphia Inquirer that found the stories could not be traced to any real media outlet. The campaign blamed an outside consultant for the AI-generated content.
A fabricated image of former President Donald Trump wading through floodwaters has circulated widely on social media, despite clear signs of AI manipulation. The photo, shared over 160,000 times, is part of a wave of synthetic images spreading misinformation after Hurricane Helene. Fact-checkers have flagged the image as altered, highlighting concerns about media literacy.
Live 95.5 FM in Portland, Oregon, has introduced an AI-generated voice clone of midday host Ashley Elzinga, using Futuri Media's RadioGPT tool. The station says the AI won't replace the human DJ, but the move highlights growing concerns about automation in the radio industry, which has faced layoffs and consolidation.
DoNotPay CEO Joshua Browder entrusted his entire financial life to GPT-4, which saved him $217.85 by canceling subscriptions and negotiating bills. However, experts and users question the safety of handing over sensitive data to an AI known for errors.
Meta has withdrawn its Galactica AI, a language model designed to generate scientific papers, after three days of public and expert criticism that it produced plausible-sounding but false content. The move highlights ongoing concerns about AI's potential to spread misinformation.
New research from the University of Washington reveals that AI tools designed to detect hate speech, including Google's Perspective, are more likely to flag posts by Black users as toxic. The bias stems from training data labeled by humans who disproportionately tagged African-American Vernacular English as offensive, leading to potential silencing of marginalized voices online.
Stanford University researchers have developed a brain-computer interface that can decode unspoken words in people with ALS, potentially easing communication. The system uses a unique mental password to protect users' privacy, achieving 98.75% accuracy in a proof-of-concept trial.
OpenAI's new Atlas browser, integrated with ChatGPT, has drawn early criticism for its search relevance, slow agent mode, and security vulnerabilities. Journalists and experts note that while the concept is ambitious, the execution currently lags behind competitors like Perplexity's Comet and Google's Gemini in Chrome.
From the Large Hadron Collider's record-breaking 7.2 trillion degrees Fahrenheit to the theoretical Planck temperature, physicists have identified several upper limits to heat. This article explores the hottest known and possible temperatures, and why conventional physics breaks down at extremes.
A physicist at the Indian Institute of Technology Jodhpur proposes that space behaves like a viscous fluid, a model that could explain why dark energy appears to weaken over time. The idea, still awaiting peer review, offers a fresh angle on discrepancies found by major sky surveys.
Netflix filed for a trademark on 'Space Force' in January 2019, nearly a year before the US military established its space warfare branch under the same name. The streaming service's move, tied to a satirical show, highlights the entertainment industry's early claim on the term.
