Share
Linkedin
Twitter
Facebook
Whatsapp
Copy Link
Users across social media report ChatGPT suddenly got smarter this week, sparking speculation that OpenAI is quietly testing GPT-5.6 inside its platform. The suspected new AI model shows stronger reasoning capabilities and advanced 3D generation, but takes significantly longer to process complex tasks—up to 87 minutes compared to GPT-5.5's 34 minutes in some tests.
Researchers at the University of Pennsylvania used AI to uncover a surprising source of antibiotics: prion proteins known for causing fatal brain diseases. The deep-learning platform identified over 1,000 antimicrobial peptides called prionins, with dozens showing strong activity against drug-resistant bacteria in lab and mouse tests.
NASA tested ERNEST, a compact four-wheeled prototype rover, across 16 miles of California desert terrain with minimal human intervention. The next-generation rover prototype achieved speeds up to 0.6 mph—ten times faster than current Mars rovers—while demonstrating advanced robotic autonomy and the ability to lift its wheels over obstacles that would stop Curiosity and Perseverance.
Hexaware Technologies is investing £25 million to expand its UK operations, creating around 1,200 jobs in AI, digital services, and quantum computing. The Carlyle-backed firm will establish new R&D centers in Manchester and Leeds while expanding its Birmingham delivery center over the next three to five years. The investment was recognized at the G7 Summit as part of the UK government's AI and clean energy agenda.
Researchers at Harvard's Wyss Institute used deep learning to screen 6 million compounds and identified two promising candidates against multi-drug resistant Neisseria gonorrhoeae. The lead compound, called A1, targets alanine racemase, an enzyme critical for bacterial cell wall formation. The study marks the first use of Organ Chip technology for antibiotic preclinical testing, offering a faster validation pipeline.
XDOF emerged from stealth with $70 million to tackle what may be AI robotics' biggest obstacle: the lack of quality training data. Unlike language models trained on internet text, robots need data capturing physical interactions—and that barely exists. The startup is building data pipelines and teleoperation systems that frontier labs are already using, while releasing the largest open-source robot manipulation dataset to date.
Nvidia's GEAR lab has developed ENPIRE, an agent harness framework that enables AI coding agents to autonomously direct robot training without human supervision. The system achieved 99% success rates on tasks like GPU installation and pin insertion, with eight-robot teams completing training in just two hours.
University of Florida researchers trained machine learning models on retinal photographs from over 40,000 patients to identify early Alzheimer's disease risk factors. The AI-powered analysis detected biological and lifestyle indicators like blood pressure, smoking, and insomnia by examining the retina's arteries and optic nerve, offering a low-cost alternative to expensive brain scans.
A UC Davis brain-computer interface has enabled Casey Harrell, a paralyzed ALS patient, to communicate independently for over 3,800 hours across two years with 99% accuracy in controlled tests. Using AI-driven speech synthesis, the BCI implant translates neural activity into speech, allowing him to work full-time as an environmental advocate and have conversations with his daughter who's never heard his voice.
Researchers have developed adaptive deep brain stimulation that adjusts electrical pulses in real time based on patient movement. The system uses machine learning to decode brain signals during walking, standing, and obstacle navigation, delivering personalized neuromodulation that improves walking in Parkinson's patients where conventional therapies fall short.
Meta's three-month-old Applied AI team of 6,500 engineers is in turmoil after forced reassignments to data-labeling tasks sparked widespread employee dissatisfaction. An expletive-laden outburst during a company livestream highlighted the crisis, prompting Mark Zuckerberg and CTO Andrew Bosworth to acknowledge mistakes in the AI reorganization that has left morale at near-record lows.
Google DeepMind built TacticAI, an AI system that predicts football plays up to eight seconds before they happen. Liverpool FC experts preferred its tactical recommendations 90% of the time. Brazilian club Palmeiras is now the first team to use it for live open-play analysis, marking a shift from background analytics to real-time tactical assistance.
A groundbreaking study reveals AI could uncover new physics faster using transfer learning, cutting simulation needs by over tenfold. But researchers discovered a critical flaw: when AI learned how the universe works through standard models, it sometimes became too reliant on prior knowledge, missing genuinely new phenomena that resembled familiar patterns—a problem called negative transfer.
USC football is breaking new ground by naming Conor McQuiston as the first director of Artificial Intelligence in college football. The move comes after the university received a $200 million gift for AI research. McQuiston will transition from his current analytics role as the program expands its technology-driven approach to gain competitive advantages.
Johns Hopkins researchers found that African American patients with diabetes were significantly more likely to receive diabetic eye exam referrals when screened with an FDA-approved AI tool compared to traditional primary care provider referrals. The AI-assisted diagnostic tool analyzed retinal images in real time, providing immediate results that increased patient adherence to essential screenings for diabetic retinopathy.
Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.