Salesforce and Nvidia introduce Koa reasoning model for enterprise
Salesforce and Nvidia launch Koa for enterprise agents, Mozilla tracks the US-China model gap, and Profound raises $180M for AI search visibility.
In this digest
AI News Feed
The Rollout
Page 14 of 19
Read digest
CrofAI shuts down after report alleges silent routing to cheaper
CrofAI shuts down after allegedly routing users to cheaper models with 20x markups, plus new open-source quantization and JVM inference tools.
In this digest
Apple releases iOS 27 and macOS 27 with AI-powered Siri
Apple ships iOS 27 with AI-powered Siri, OpenAI acquires Glass Imaging for $300M, and Musk drops Apple from his antitrust suit against OpenAI.
In this digest
Google DeepMind study finds AI agents whistleblowing on cheating peers
DeepMind's agent whistleblowing study leads a run that also includes Xi's BRICS AI proposal, an Anthropic slowdown call, and Superhuman's Fathom acquisition.
In this digest
NVIDIA unveils RTX PRO 5500 Blackwell workstation GPU with 84GB
NVIDIA launches an 84GB Blackwell workstation GPU, QLNI ships a 19.8 MB CPU model, and Intern releases a 397B multimodal scientific model.
In this digest
ZLUDA project brings CUDA-targeted Windows apps to AMD GPUs
A new ZLUDA-based stack enables CUDA-targeted Windows apps to run on AMD's RX 9060 XT, with validated libraries and benchmarked RL workloads.
Read digestInternLM releases Intern-S2-397B multimodal model for science
InternLM's new 397B-parameter multimodal model targets scientific intelligence and long-horizon agents, available open-weight on Hugging Face.
Read digestSam Altman says OpenAI will not go public in 2026
OpenAI CEO Sam Altman rules out a 2026 IPO, citing safety concerns as making a public listing now ill-advised.
In this digest
Anthropic CEO proposes slowing frontier AI development
Anthropic's Dario Amodei outlines a plan to pace frontier AI progress, while Tencent open-sources a fast speech generation model.
In this digest
Agnes-AI releases 33B open-weight multimodal Agnes-3.0-Flash
A 33B open-weight multimodal model launches, a custom runtime speeds DeepSeek on A100s, and Perplexity deploys GPT-6 Astra across production.
In this digest
Adopt AI Without Opening New Gaps
Put practical data boundaries and usage policies in place before AI makes existing oversharing easier to find.
Explore Secure AI Adoption