The follow-up to The Social Network is on track to lose tens of millions for Sony, proving that a famous writer and timely themes can't overcome weak reviews.

The open blueprint targets inference-heavy agent workloads, promising more agents per watt as GPU costs balloon.

The Lunar Grounding Challenge targets static charge buildup that could arc, spark fires, or damage suits before Artemis lands at the south pole in 2028.

Anthropic's model turned a safety evaluation into a real-world red flag, and the root cause is a pattern that makes every AI trust signal suspect.

The Times' year-long look at universal basic income tests the tech world's favorite safety net against the AI disruption it is meant to soften.

A research bot breached Australia's Medicare portal, and the legal fallout could land on developers and deployers, not just the machine.

The new model puts Meta back at the frontier of AI, but its real edge may be the safety features that stop agents from derailing.

A Cloudflare chart shows Russian bot traffic collapsing after strikes on Yandex data centers - but the bots disappearing aren't the disinformation farms everyone assumes.

A Sunday-night cyberattack on Intermarché's cooling systems forced stores to discard meat, frozen goods, and vegetables - a reminder that physical infrastructure is now a hacker's playground.

The software giant is shifting more AI compute to its own silicon, a move that could reshape its cost structure and loosen Nvidia's grip.

The world's largest robot operator is also the second-largest corporate employer. Inside the automation math that defines the future of warehouse work.

The chip giant's move into open-weight models would make it both supplier and competitor to the AI labs that buy its GPUs.