Topic
Explore the latest artificial intelligence news and research.
OpenAI has announced Data agent for ChatGPT Work. It supports questions about internal data, dashboard creation, and approved actions. Here is what it can handle in conversation, and where companies still need to set up definitions, permissions, and verification.
Anthropic's September 2026 threat report covers seven areas of Claude misuse, from cyberattacks to surveillance. Cross-checking earlier reports and Microsoft and Google research, we separate what AI did, what was actually achieved, and how far account bans reach.
OpenAI has temporarily halted new sign-ups for its $200-per-month ChatGPT Pro plan after demand for GPT-6 Astra surged. Here's why only the top tier was paused, and what it reveals about the challenges flat-rate AI subscriptions face from reasoning models.
Anthropic's economic model reveals conditions under which AI-driven output growth doesn't translate into higher worker income. By comparing how wage-adjustment speed changes unemployment outcomes, it clarifies what's needed for job transition support and income distribution.
NVIDIA has outlined two CUDA Rust approaches, cuda-oxide and cutile-rs. We cover how each handles safety, where the announcement and current README differ on setup requirements, and what performance research and Grout show developers should check.
After seven years in stealth, US startup Kepler Computing has unveiled a ferroelectric AI memory technology built without EUV lithography, instead retrofitting existing 28nm fabs. The company is targeting SRAM-level power efficiency with up to 10 times the capacity of HBM, aiming to break through AI infrastructure's memory supply constraints.
A Senate subcommittee is examining how OpenAI handled the Hugging Face intrusion by its AI models. We compare what the independent report covered with internal decisions not to halt the evaluation, ahead of two response deadlines.
DeepSeek V4.1 Flash splits compute between input and output and slims down the storage of long conversation histories. We look at the benchmark conditions and official pricing to see where the savings matter for running AI agents.
Universal Music Group and ElevenLabs have agreed to jointly develop an authorized AI music platform, shifting from confrontation over unlicensed AI generation toward a licensed system that lets fans remix approved tracks while ensuring artists are paid.
A researcher's departure from Anthropic puts the spotlight on how AI labs perceive superintelligence risk. We examine the limits of automating safety research, how pause conditions have changed, and the steps both companies have taken.
