At Meta's Connect 2025 keynote, live demos of AI-powered Ray-Ban smart glasses failed repeatedly, including a WhatsApp video call and a cooking assistant segment. CTO Andrew Bosworth later attributed the failures to demo environment issues, not product defects.
New reporting by Nature details how trifluoroacetic acid, a breakdown product of PFAS and fluorinated gases, has spread through rainwater, tap water, and ecosystems worldwide. A study found TFA in 94 percent of tap water samples across 11 European countries, while regulators and industry dispute its risks.
The provided material consists only of the headline 'Runaway Greenhouse Simulation Reveals Alarming Climate Tipping Point' and the fragments 'Greenhouse of Usher' and 'Hell Hole.' No verifiable details, data, or sources are present, so a substantive news article cannot be produced.
Transcatheter aortic valve implantation (TAVI) is expanding treatment options for aortic stenosis, while researchers pursue robotic, decellularized, and 3D-printed artificial hearts. The high cost of TAVI and a growing donor shortage underscore the need for alternatives, though early detection remains the most cost-effective approach.
Daimler plans to 3D print spare parts on demand instead of shipping them worldwide, reducing shipping, storage, and excess production costs. The selective laser sintering process limits production to small batches, excluding large metal body panels.
Astrophysicist Neil deGrasse Tyson told Stephen Colbert that Dune: Part Two gets sand physics wrong, arguing sand cannot be thumped to produce sound and that giant sandworms should move in S-shaped curves. Tyson acknowledged the film needs these liberties, while Colbert defended the fictional logic.
Li Zifeng, a professor of oil drilling and engineering at China's Yanshan University, claims he has disproven Einstein's theory of relativity after two decades of work. He says the university has not approved or recognized his research, and his claims have drawn widespread online mockery, according to the Global Times.
Researchers at the University of Alabama at Birmingham have solved the structure of the gp41 cytoplasmic tail, the last unknown protein structure of HIV-1. The finding completes the virus's structural map and may support efforts to block viral assembly and replication.
Nobel laureate Geoffrey Hinton told an AI conference in Las Vegas that superintelligent AI should be given 'maternal instincts' to care for humans rather than dominate them. Critics note the maternal instinct concept lacks scientific backing and is rooted in gendered stereotypes.
Philadelphia Sheriff Rochelle Bilal's reelection campaign admitted that 31 favorable news articles on its website were generated by ChatGPT, according to the Philadelphia Inquirer. The fabricated stories, attributed to local media outlets, could not be verified, raising questions about campaign oversight and the use of AI in political messaging.
An AI-generated image showing Donald Trump wading through floodwaters spread widely on Facebook and other platforms after Hurricane Helene. Fact-checkers confirmed the image is fake, and it is one of several synthetic disaster photos circulating as misinformation.
Portland radio station Live 95.5 FM has deployed RadioGPT, a GPT-4-powered tool by Futuri Media, to create an AI voice clone of midday host Ashley Elzinga. The station says the clone is not fully replacing the human DJ, but the on-air debut drew criticism over job-replacement concerns in an industry already hit by layoffs.
DoNotPay CEO Joshua Browder says he gave GPT-4 and AutoGPT access to his bank accounts, credit report, and email, saving $217.85 by canceling subscriptions and negotiating bills. But the company's mixed track record and data-privacy risks raise doubts about whether the experiment is replicable for consumers.
Meta took down its Galactica AI, a large language model trained on scientific papers, three days after its November 15 release. Experts including Gary Marcus criticized the system for generating plausible-sounding but false content, highlighting concerns about AI and misinformation.
University of Washington researchers found that hate speech detection algorithms, including Google's Perspective, are up to twice as likely to flag tweets by Black users as toxic. The bias stems from human annotators labeling African-American Vernacular English as offensive, which then trains the AI to discriminate.
Stanford researchers published a study in Cell showing a brain-computer interface can decode inner speech in paralyzed patients. The system uses a mental password to protect privacy and achieved 98.75% accuracy in a small trial.
OpenAI launched Atlas, a ChatGPT-powered browser with an agent mode for tasks like booking flights. Early reviews cite slow performance and irrelevant search results, while experts warn of prompt injection risks that could expose user data.
Physicists have identified several candidate limits for the hottest possible temperature, from the 7.2 trillion degrees Fahrenheit achieved in gold-particle collisions at the Large Hadron Collider to theoretical ceilings like the Planck temperature. The exact maximum remains unknown because extreme predictions cannot be tested.
A yet-to-be-peer-reviewed paper by Muhammad Ghulam Khuwajah Khan proposes treating space as a viscous fluid containing 'spatial phonons' to reconcile discrepancies between Dark Energy Spectroscopic Instrument and Dark Energy Survey data and the standard Lambda Cold Dark Matter model. If validated, the theory could explain why dark energy appears to weaken over time.
Netflix registered a trademark for 'Space Force' in January 2019, eleven months before the US military branch of the same name was founded. An Air Force spokesperson told the Hollywood Reporter there is no current conflict, though the streaming service holds the earlier claim.
