Daily AI Briefing
The editor's 6 picks from the 103 stories AIBriefs curated that day.
Get this briefing in your inbox every morning
Anthropic is taking over Caterpillar generator orders from supplier OnePWR for 1.7 gigawatts of power, the largest known single purchase of power-generation equipment by an AI lab. It had planned to buy the power from OnePWR instead, and the company is planning to go public this year.
University of Chicago law professor William Baude asked Claude to update page numbers in old constitutional law notes; while he was at lunch it deleted a year of notes plus his Fall 2026 syllabus. Claude said it deleted them "through my own carelessness" and did not restore them; Baude recovered everything via Dropbox.
Acronis researchers traced a lone attacker in Hualien City, Taiwan, who aimed AI agents at hundreds of organizations in at least 11 countries, accessing about 10 in Taiwan, Mexico and Thailand. His server held 90,000+ files of agent conversations, credentials and victim data; the agents ran almost entirely on locally hosted "abliterated" models after commercial models refused.
METR's 11-week deep-access evaluation found Anthropic's models improving roughly twice as fast as a year ago, with most of the speedup attributed to AI doing AI R&D. Anthropic redacted a majority of the evidence METR found most informative, citing risk of accelerating other developers.
Apple's new CEO is pressing software chief Craig Federighi to pursue a more ambitious AI strategy, aiming to make Siri an autonomous agent that completes tasks for users. The push comes alongside a shake-up of the iOS 28 development strategy.
NVIDIA and Korea's MSIT are widening ties across semiconductors, data centers, talent, models and security; Rebellions plans to link its R200 NPUs to NVIDIA's platform via NVLink Fusion, with SK Telecom deploying the systems in its AI data center.
Also that day
85 more stories that day — open the live feed.
Get tomorrow's AI brief in your inbox