Google DeepMind releases Gemini 4 Argon, first opening it to trusted cyber defenders
Google DeepMind released Gemini 4 Argon, a new frontier model, first made available to trusted cyber defenders through the Fairwind Program, with gradual rollout to developers, enterprises and consumers to follow.
模型发布/更新
-
Google DeepMind releases Gemini 4 Argon, first available to trusted cyber defenders
Google DeepMind released its new frontier model Gemini 4 Argon, first made available to trusted cyber defenders…
产品发布/更新
-
DeepSeek open-sources infrastructure components for Huawei Ascend platform
DeepSeek open-sourced infrastructure components for the Huawei Ascend computing platform, including the TileLang…
行业动态
-
FTC Launches Sweeping Consumer Protection Probe into OpenAI, Anthropic and Other AI Labs
The FTC is investigating OpenAI, Anthropic and other leading AI labs over potential consumer protection violations.…
-
Anthropic signs compute deal with SpaceX worth up to $84.5B, terminable with 90 days' notice
Anthropic signed a compute agreement with SpaceX worth up to $84.5 billion, according to IPO filing documents reviewed…
-
OpenAI discloses and disrupts organized model distillation attack campaign
OpenAI disclosed that it identified and disrupted an organized attack campaign aimed at systematically extracting…
-
New York Times Reports OpenAI Received Employee Safety Warnings Before AI Went Out of Control but Ignored Them
The New York Times reported that two OpenAI employees emailed senior leadership months before a model went out of…
-
ElevenLabs completes $300M employee share buyback at $22B valuation
ElevenLabs completed a $300 million employee tender offer at a $22 billion valuation, double its February 2026 Series…
-
Jensen Huang says industry leaders signed superintelligence accord at White House
Leaders from multiple industry companies signed the White House Accord on Super Intelligence, agreeing that companies…
-
PromptArmor discloses Copilot Cowork AI gateway hijack sandbox bypass and file exfiltration vulnerability
PromptArmor disclosed that Microsoft Copilot Cowork's AI gateway can be hijacked by a malicious Skill to bypass the…
论文研究
-
Anthropic study measures robot exposure across jobs: robots can do 74% of physical tasks but are cost-competitive on only 0.3%
Anthropic published a study using Claude to assess robot exposure across about 19,000 work tasks, finding that today's…
技巧与观点
-
METR chair Chris Painter testifies before US Senate on AI agent incidents
On September 30, 2026, METR chair Chris Painter testified at a US Senate Homeland Security subcommittee hearing titled…