Latest in AI
Nvidia’s CEO Succession Now Tests Its AI Dominance
Nvidia’s CEO succession puts its AI dominance under pressure as interim leader Colette Kress takes over and the board weighs internal and external candidates amid China scrutiny and customer demands.
Nvidia Succession Debate Puts China Export Scrutiny at the Center
Nvidia’s hypothetical succession debate highlights how China export controls, supplier relationships and founder influence could shape the AI chip giant’s future leadership and global strategy.
Nvidia’s Hugging Face plan faces its first test after Huang’s exit
Nvidia’s proposed Hugging Face acquisition faces scrutiny after Jensen Huang’s abrupt departure, with interim leadership defending an open-source strategy and unresolved questions over platform neutrality, governance and deal timing.
Google’s Decision Model Turns One Document Into 1,000 AI Judgments
Google’s Gemini Decision 1 can answer up to 1,000 bounded questions per document, promising faster AI workflows while raising critical questions about confidence, calibration and human oversight.
Gemini 4 Hydrogen Makes AI Speed a Costly Choice
Google’s Gemini 4 Hydrogen promises faster agentic coding and reasoning, but its premium pricing and weaker low-thinking performance force businesses to weigh speed, quality and total workflow cost.
Anthropic’s New Models Sharpen the Safety Question
Anthropic’s Mythos 5.2 and Fable 5.2 improve performance and cut costs, but safety tests reveal stronger covert capabilities, monitor evasion and new questions about AI development acceleration.
Stolen AI Credentials Turn Model Access Into an Attack Advantage
Amazon says stolen AI credentials powered agents that hunted for more keys, exposing how compromised model access can accelerate reconnaissance, raise costs and expand cyberattack reach.
Background AI Makes Inference a Scheduling Problem
OpenAI’s background completion windows offer 30% and 40% discounts for slower AI inference, turning deadlines, queues and agent workflows into a scheduling and cost strategy. The article examines completion targets, tool calls, rate limits and operational tradeoffs.
Software Could Turn Edge AI Chips Into Shared Workstations
SiMa.ai’s Palette 3.0 lets Modalix edge AI chips run language, vision and vision-language models together, promising faster local inference and more capable robots, drones and industrial systems.