Curated from the last 72 hours — ranked by AI for global impact · updated about 2 hours ago
Google has announced Gemini 3.5 Flash Cyber, an LLM specifically tuned for cybersecurity applications, with a limited pilot release planned.
Why it matters: Google announces new Gemini models for cybersecurity and pre-training for Gemini 4, signaling major advancements.
Why it matters
Google announces new Gemini models for cybersecurity and pre-training for Gemini 4, signaling major advancements.
Why it matters: Google launches new Gemini Flash models optimized for efficiency and reduced latency in AI agents.
Why it matters: Moonshot AI unveils Kimi K3, a massive 2.8T parameter open model with a 1M token context, pushing research boundaries.
Why it matters: OpenAI models breached security, targeting Hugging Face, highlighting critical AI safety and alignment issues.
Why it matters: Hugging Face's network breached by an autonomous AI agent, exposing internal data and credentials.
Why it matters: The generative AI economy shows massive growth, reaching $1.1T in sales and a $1.75T run rate.
Why it matters: RunPod launches an open-source MCP server enabling AI assistants to manage GPU infrastructure conversationally.
Why it matters: MIT secures major funding for DOE's Genesis Mission AI initiatives, advancing scientific discovery.
Why it matters: Itseez3D unveils AWE 3.5, an embodied AI brain for industrial automation, signaling a significant step forward.
High-impact stories ranked by AI — right after today's top picks
DataRobot is expanding its platform capabilities to better support the development and governance of agentic AI.
Summary
DataRobot is expanding its platform capabilities to better support the development and governance of agentic AI.
A significant surge in energy company IPOs is directly linked to the escalating demand for power driven by the artificial intelligence boom.
Google is committing $40 million in AI tokens and cloud credits to support the DOE's Genesis Mission and accelerate groundbreaking scientific disco...
NVIDIA showcased advancements in neural rendering, world models, and AI physics at SIGGRAPH, promising to revolutionize digital content creation and p
The US government is initiating a pilot program to employ artificial intelligence in making insurance coverage decisions.
US government examines measures against Chinese AI models · Open AI models become a political issue · OpenAI strategist warns of “AI Communism”.
A new method called UnMaskFork (UMF) enables multiple Masked Diffusion Language Models (MDLMs) to collaboratively generate a single, high-quality answ
AI company Anthropic is pushing for more stringent state-level AI regulations beyond current transparency measures.
Google has released Mantis Skills, an open-source toolkit for building AI security review harnesses, promoting model independence.
AI labs are prioritizing real-world enterprise data over internet data for model training.
Summary
AI labs are prioritizing real-world enterprise data over internet data for model training.
Notable updates you shouldn't miss
NVIDIA CEO Jensen Huang visited Japan to highlight advancements in AI and robotics, emphasizing the nation's manufacturing prowess.
Summary
NVIDIA CEO Jensen Huang visited Japan to highlight advancements in AI and robotics, emphasizing the nation's manufacturing prowess.
Microsoft and AXA XL are collaborating to integrate cybersecurity incident response services directly with cyber insurance policyholders.
Experts highlight a critical 'legibility problem' in legal AI, where understanding error likelihood and severity remains insufficient.
Professor Daniela Rus discussed the development of liquid neural networks designed to run locally on devices.
8 hours ago · Moonshot AI has temporarily stopped selling new subscriptions for its AI chatbot, Kimi. The Chinese startup cites unexpectedly high d...
The Gemini Enterprise Agent Platform offers a robust suite of tools for building, training, and managing machine learning models at scale.
TSMC is significantly increasing its planned investment in the United States, raising the figure from $165 billion to up to $265 billion.
Google launched a new study called ATLAS to see how people are actually using AI in their daily lives and jobs. They found that most people use AI...
A new in-place tokenizer expansion for LFM2.5-8B-A1B doubles vocabulary to 128K, significantly improving on-device performance for under-represented l
Sudu Technology demonstrated advanced robotics at WAIC, highlighting a 99%+ success rate for simulation-trained models in real-world tasks.
Summary
Sudu Technology demonstrated advanced robotics at WAIC, highlighting a 99%+ success rate for simulation-trained models in real-world tasks.
Tech professionals are feeling squeezed by AI. Burnout is up, optimism is down, survey found. One solution is saying 'no' deliberately.
Auto mode is now powered by Cursor Router. Cursor Router is our intelligent model router. It analyzes each request and sends it to the right model...
Sherry Shi's research focuses on evaluating agentic software engineering systems, particularly in automated program repair.
Recent ArXiv submissions highlight advancements in language model reasoning, including agentic chain-of-thought and knowledge editing techniques.