1841 articles
ChatGPT saw 5.844 billion website visits in November, its second monthly decline this year. The dip happened even though the platform hit a record daily peak of 233 million visits.
Grok 4.1 debuts with an industry-leading 2-million-token context window, outpacing all major AI competitors and enabling entire codebases and lengthy documents to process in one session.
A researcher reports that GPT-5 proposed the central idea developed in a new Quantum Field Theory paper published in Physics Letters B. Screenshots show GPT-5 outlining a key conceptual step using the Tomonaga–Schwinger formalism.
Alibaba introduced Qwen3-VL, a multimodal vision-language model with a 256K native context window, enhanced text comprehension, and improved reasoning capabilities across visual tasks, showing top performance in multiple AI benchmarks.
Bank of America reports a sharp jump in debt issuance by major tech companies funding AI datacenter expansion, with 2025 borrowing at $121 billion—four times the five-year average. Hyperscalers are leaning heavily on loans and bonds as capital needs climb.
Fresh measurements from Google show Gemini now uses far less water, electricity, and CO₂ than earlier reports suggested. The updated data reveals a 33-fold jump in energy efficiency compared to widely circulated 2023 figures.
Anthropic CEO Dario Amodei shared an optimistic outlook on AI's future, describing exponential capability growth and massive commercial potential. His comments reveal how advanced models are already transforming internal development workflows at the company.
Tencent AI Lab introduced R-Few, a self-evolving training framework that lets large language models improve themselves with minimal human input. The system uses a Challenger–Solver architecture to tackle key stability challenges in iterative LLM training.
NVIDIA rolled out GenMol v2 with a redesigned SAFE syntax that uses angle brackets to map molecular connections more clearly, streamlining three key research workflows.
Hugging Face unveiled Transformers v5, bringing major improvements in simplicity and production readiness. The update supports over 400 model architectures and strengthens integration with modern AI development tools.
A new framework for AI applications shifts focus from expanding model windows to precise context engineering. The discipline centers on retrieving relevant information through agents, memory tools and structured prompts.
McKinsey's latest research shows that companies which derive at least 10 percent of their EBITDA from generative AI spend far more on governance and trust measures than their rivals plus lead in responsible AI practices.
Amazon Web Services introduced EC2 Trn3 UltraServers that contain the Trainium3 chip. The new machines give a large jump in compute strength and in power use. Firms that deploy the hardware report sharp drops in cost plus quicker job completion.
Anthropic bought Bun, the JavaScript runtime that is famous for its speed. Claude Code passed the one billion dollar mark only six months after it launched. Bun remains open source as it joins Anthropic's set of AI development tools.
Grok reached the top spot on CryptoBench, a benchmark measuring how well AI models handle crypto tasks. The results show xAI making serious moves in digital-asset analytics.
Microsoft is rolling out new "Projects" capabilities for Microsoft 365 Copilot, transforming the assistant from a simple chatbot into a structured project management tool that handles complex, multi-step workflows.
Alibaba just dropped its Quark AI Glasses lineup with two models—the flagship S1 and lifestyle G1. Both run on Qwen AI and pack features like text/image Q&A, near-eye navigation, and in-lens payment apps.
Fresh Similarweb data reveals ChatGPT sees steeper weekend traffic drops than Google's Gemini, pointing to different user patterns as AI tools try to expand beyond work hours.
Fresh projections show AI training demand will peak in 2025 before compute needs shift dramatically toward inference. The forecast includes growing adoption of custom chips, presenting long-term competitive challenges for NVDA.
Perplexity launched BrowseSafe and BrowseSafe-Bench as open-source tools that catch prompt-injection attacks in real time, with early testing showing the system outperforming major AI models in threat detection.
We use cookies to improve your experience on our site and to show you relevant advertising. To find our more, read our privacy policy and cookie policy
Victoria Bazir
Saad Ullah
Alex Dudov
Usman Salis
Peter Smith
Eseandre Mordi
Marina Lyubimova
Sergey Diakov
