768 articles
Elon Musk's xAI secured a major Pentagon contract to deploy Grok on classified military networks alongside OpenAI and Google, signaling the U.S.
BBC investigation reveals Grok and other AI chatbots are triggering psychological crises in vulnerable users, with 414 documented cases worldwide of people...
OpenAI CEO Sam Altman invited Elon Musk to the GPT-5.5 launch party scheduled for May 5, 2026, despite an ongoing $134 billion lawsuit.
Tesla generated $573 million from Elon Musk's other companies last year, with xAI accounting for $430 million through Megapack battery sales and Grok chatbot...
ChatGPT chat logs are becoming key evidence in murder investigations, revealing suspects' alleged plans and mindsets.
Generative AI models like GPT-4 can now handle complex cognitive tasks once thought safe from automation, reshaping workplace skills and job roles in ways that...
David Silver, the DeepMind researcher behind AlphaGo, raised $1.1 billion for Ineffable Intelligence to build AI through reinforcement learning instead of...
Windows 11's May 2026 update introduces AI agent monitoring directly on the Taskbar, starting with Microsoft 365 Copilot's Research feature.
An anonymous AI researcher paid $10,000 to debate AI safety pioneer Eliezer Yudkowsky, signaling growing investment in public discourse around artificial...
New tools are emerging to help brands get cited by AI search engines like Perplexity and ChatGPT.
Prediction markets give xAI a 75% chance of releasing a next-generation Grok model that scores 1440 or higher on Chatbot Arena by year-end 2026, but...
OpenAI's ChatGPT Images 2.0 generates photorealistic fraudulent documents with alarming ease, including fake prescriptions, bank alerts, and IDs.
Bitcoin.com tested 13 leading AI models on cryptocurrency forecasting and found predictions ranging from $1.58 to $3.85,revealing how identical data produces...
Nebius acquired Eigen AI for $643 million, gaining a team that optimized Meta's Llama 3 and Llama 4 models.
Developers using Claude Code are accidentally routing requests through OpenAI's API format instead of Anthropic's native endpoint.
Authorities are investigating multiple crimes where suspects allegedly used ChatGPT to plan attacks.
OpenAI launched GPT-5.5 at double the price of its predecessor while Chinese rivals undercut Western models by 90-97%, forcing enterprises to choose between...
After a year-long hiatus and a reputation-damaging Llama 4 release, Meta is back with Muse Spark.
Amazon posted record earnings and 28% AWS growth, but investors balked at a $59.3 billion surge in capital spending to fuel AI infrastructure.
The Pentagon signed classified AI agreements with OpenAI, Google, Microsoft, Amazon, Nvidia, xAI, and Reflection, but excluded Anthropic over supply-chain...
Mistral AI launches Workflows, a production orchestration layer already processing millions of daily executions.
Nvidia's long-term growth thesis remains intact, but a major analyst now expects most gains to arrive in 2028-2030 rather than sooner.
AI writing and design tools have lowered content creation costs from hundreds to pennies, but success requires human oversight.
Moonshot AI's Kimi K2.6 model costs $0.95 per million tokens while matching or beating pricier competitors on coding tasks, reshaping the AI efficiency...
Freight and logistics firms face a new challenge: getting cited by AI answer engines like Perplexity and ChatGPT.
OpenAI may be building a smartphone with Qualcomm and MediaTek, targeting 2028 production.
New testing reveals no single smartest AI exists. Grok 4.2 dominates math and logic with 33% fewer reasoning errors, while Claude excels at writing quality.
Alibaba closed weights on Qwen3.6-Max, its flagship model, for the first time in three years.
Apple's new CarPlay support lets drivers use ChatGPT and Perplexity AI as voice assistants.
Elon Musk has publicly acknowledged that xAI trained Grok using OpenAI's models through a technique called distillation.
Open-design, a new Apache-2.0 licensed framework, lets developers generate designs locally using 19 skills and 71 brand systems without paying for a separate...
Enterprise employees lose over 8 hours per week hunting for information across disorganized systems.
A Harvard-led study found OpenAI's o1 large language model matched or exceeded human physicians in emergency diagnostic tasks, diagnostic triage, and clinical...
Anthropic studied 1 million Claude conversations and found 6% of users seek personal guidance on health, careers, relationships, and finance.
60% of searches now end without clicks as AI assistants replace Google. Companies are racing to adopt 'Answer Engine Optimization' to appear in ChatGPT and...
A developer discovered his $73 Claude bill wasn't from heavy usage, but from a hidden cost mechanism: the model rereads your entire conversation history with...
A technology analyst replaced Google's Gemini Pro subscription with open-weight models like Gemma 4 running locally on personal devices, maintaining...
ChatGPT Images 2.0 dominates in India with 5 million downloads in launch week, but global engagement remains modest at 1% growth.
Anthropic's new BioMysteryBench reveals Claude models solving 30% of bioinformatics problems humans couldn't crack, marking a shift in AI's scientific research...
IBM released Granite 4.1, a family of compact AI models that outperform larger competitors while cutting costs and running efficiently on enterprise hardware.
X launched its largest ad system update ever, integrating xAI's Grok models to power smarter targeting and real-time ad delivery.
Moonshot AI's Kimi K2.6 deploys up to 300 parallel sub-agents to tackle massive tasks in one run, generating 104-page literature reviews and optimizing complex...
Australia's strict copyright laws are pushing AI infrastructure investment to other countries, while Chinese open-source models like DeepSeek capture one-third...
DeepSeek V4 signals a fundamental shift in AI competition. By combining open-source weights, ultra-low costs, and million-token context with domestic Chinese...
Anthropic launches Claude Security in public beta, enabling enterprises to scan codebases for vulnerabilities using Claude Opus 4.7.
Google launched a native Gemini Mac app on April 15, 2026, ending years of browser-only access.
New research reveals a troubling trade-off: training AI chatbots to be warm and empathetic increases their error rates by 10-30 percentage points, making them...
Claude's Artifacts feature now supports persistent storage, live data refresh, and direct API calls, transforming it from a preview tool into a full microapp...
xAI's Grok 3 outperforms GPT-4 on reasoning tests and answers breaking news questions with 89% accuracy by tapping live X data.
During OpenAI litigation, Elon Musk stated artificial intelligence poses existential risks comparable to Terminator, revealing safety concerns he's championed...
