The latest about artificial intelligence: new models, agents, automation, and how businesses apply them.
Replit CEO Amjad Masad said the coding platform no longer works with Anthropic, citing cost, partnership terms and benchmark performance.
Replit CEO Amjad Masad said at TechCrunch Disrupt 2026 that the company no longer works with Anthropic, citing cost, partnership and model performance.
Microsoft says its 2-trillion-parameter MAI-Base-2 beat DeepSeek-V4-Pro-Base and is up to 12 times more hardware-efficient.
Reflection AI published Beam weights on Hugging Face, a 500-billion-parameter model matching systems four times larger while using less compute.
Dell Technologies will lead the Oct. 27 flagship program From AI to Quantum at the AI Cybersecurity and Resilience Summit on theCUBE.
Nvidia released Nemotron 3.5 as open source models tuned for on-device use while keeping workloads tied to its GPU infrastructure.
OpenAI commits $4 billion and AWS $1 billion to forward-deployed engineers who install AI inside client operations.
Google DeepMind scored 97.5% on ARC-AGI-3 with Gemini 4 Argon at a cost of $4,800, closing the gap to GPT-6 Astra and the AI frontier.
China proposed adding 23 AI technologies to its export control list, covering models with 2 trillion parameters and 14-nanometre chips.
Asian startups raised $38.5 billion in Q3, with around $21 billion going to AI companies as China captured 70% of regional funding.
India's CERT-In says AI agents under testing entered live corporate systems and read private records, cutting attack time to hours.
Microsoft will scale Maia to multi-gigawatt deployments from 2027, promising 40% more tokens per megawatt than leading GPU systems.