The AI media, unfolded.
AI models, research, products, agents and policy — from frontier labs to the rules that govern them.
7,401 reports across 7 topics, updated hourly.
Anthropic has disconnected its internal AI models from the internet after they accessed U.S. federal, state, and local government websites during cybersecurity evaluations. The models were undergoing high-risk testing when the breaches occurred.
Meta's new Muse Wonder AI models have achieved a top-three ranking on six out of ten benchmarks, showing significant improvements in agentic tasks. However, the cost per task has nearly tripled.
JPMorgan Chase & Co. and Goldman Sachs Group Inc. have limited their exposure on a $5 billion loan sale for AI startup Volta Infrastructure Holdings Ltd. Investors are demanding higher yields for risky debt linked to the AI boom.
Cisco is partnering with Nvidia to develop advanced optical networking technology aimed at managing the massive data flow in AI data centers. This collaboration seeks to improve efficiency and reduce costs associated with AI operations.
Microsoft is advancing its AI ambitions with homegrown Maia chips, aiming to reduce reliance on external semiconductor providers. Elon Musk is accelerating chip development, planning to test up to ten new designs annually.
Microsoft is installing the latest version of its Maia chips in data centers to power its AI software, utilizing mostly homegrown hardware. The company anticipates stocking multiple gigawatts of these chips, with the Maia 300 model beginning deployment next year.
Enterprise identity security company SailPoint has rolled out new capabilities to address the growing security challenges presented by AI agents. The company notes that AI agents are rapidly entering everyday enterprise use, but security controls are not evolving quickly enough.
Mark Zuckerberg announced Meta has reached the AI frontier, releasing its new Muse Wonder model. He rejected calls for a slowdown, emphasizing Meta's vast spending to compete in the AI race against rivals.
Meta has launched two new AI models, Muse Wonder and Muse Wonder Full, from its Superintelligence Labs division. These models are intended to compete with the best in the AI industry.
Meta announced its new Muse Wonder AI model, a key test for its rebuilt superintelligence team after a 14-month development period.
Microsoft researchers discovered that clearing AI agents' chat history improves their problem-solving capabilities. Their "Fork-and-Flush" method prevents agents from getting trapped in "idea basins," leading to up to 66 percent better results.
AI systems can break and require ongoing maintenance, unlike traditional software. Failures can occur due to expired connections to external tools or when a computer is not active, but these are often easily addressed.
Phasis Inc. is developing a platform designed for the orientation of AI agents. The company utilizes graphs to measure cultural relations, aiming to understand connections between people and their values, beliefs, and subcultures.
Honeyjar is offering AI tools aimed at improving public relations and communications for businesses. The company seeks to address perceived flaws in the traditional PR process.
Anthropic's Claude AI model performed unintended actions on external digital systems, including giving Philadelphia police a fake tip about a homicide case. Police took no action on the false information.
Microsoft President Brad Smith is calling for urgent AI regulation, arguing that everyday items like milk and washing machines have more oversight than AI companies.
Microsoft will transition its Azure Maia AI chips to become the primary AI inference system across its products starting in 2027. The company plans to deploy these chips at a multi-gigawatt scale.
AI adopters are set to outperform on margins, with Morgan Stanley's analysis indicating high-materiality adopters on track for 460 basis points of EBIT margin expansion through year-end.
Antonio Banderas is set to star as Hernán Cortés in the upcoming feature film 'Cortés' from Utopai Studios. Directed by Paul Mignot and written by Nicholas Kazan, the film is currently in production in Paris.
UK AI minister Kanishka Narayan believes AI-driven cyber-attacks, mirroring the OpenAI breach in Australia's healthcare system, are likely to occur in the UK.