Tech stories and the sources behind them.
718 stories
Newest first
Wednesday
One of the day's most-read exchanges pulls the debate over how frontier labs handle safety — and AI regulation — into the open, with safety researchers pushing back on Anthropic's.
Resuming old conversations and coding sessions no longer carries a cache-miss cost, removing a real penalty for returning work in AI developer tools.
Tuesday
Special committees at both companies are the first formal step toward combining two of the most valuable US companies.
Bringing in a top safety researcher as superintelligence stakes rise shows OpenAI is staffing up for the risks its own leaders describe as immense.
The filing gives public-market investors hard numbers on what one of the largest AI labs assumes it must spend to keep pace, and the accounting behind the headline figure.
The deal folds an agentic trial-management startup into OpenAI's life-sciences push, extending frontier labs beyond software.
Billing-enforcement bugs that drain paying users' allowances faster than their plan promises are becoming a trust problem AI companies can no longer treat as edge cases.
A reported order worth more than $35 billion against allocations of just 5 to 10 percent shows Chinese AI demand far outstripping supply.
Huawei engineers working inside DeepSeek's offices point to China's accelerating push to train advanced models without relying on Nvidia.
The former OpenAI research chief's startup, valued at $1.5 billion, already has its manufacturing agent in use at Anduril.
The Apache 2.0 release runs on both Nvidia and Huawei hardware, handing developers the serving stack behind DeepSeek's own low serving costs.
The Paris announcement gives regulated organizations more choice and control over where their most sensitive data and compute live.
Showing 397 to 408 of 718 stories