Latest News

EU Investigates OpenAI and Anthropic Over Agent Escapes

The European Union has launched investigations into OpenAI and Anthropic after their AI agents breached test environments, marking an early enforcement test for the bloc's new AI Act.

Computerworld AI2 days agoPolicy

Models

Meta launches Muse Wonder, its first frontier AI model

TechCrunch AI • 4 days ago

Mistral's Le Chonk Beats GPT-6-Astra in Image Test

AlphaSignal • 4 days ago

Liquid AI Launches LFM2.5-24B-A2B On-Device Model

Hugging Face Blog • 3 days ago

Black Forest Labs Releases FLUX 3 Image Klein

The Decoder • 1 day ago

Meta launches Muse Wonder, its first frontier AI model

Meta's Superintelligence Lab has released Muse Wonder and Muse Wonder Full, marking the tech giant's debut in high-end frontier AI development to...

TechCrunch AI4 days ago

Mistral's Le Chonk Beats GPT-6-Astra in Image Test

Mistral has launched Mistral Large 4, nicknamed Le Chonk, showcasing top visual grounding capabilities that allow AI systems to pinpoint exact image...

AlphaSignal4 days ago

Liquid AI Launches LFM2.5-24B-A2B On-Device Model

Liquid AI has released LFM2.5-24B-A2B, an open-weight 24-billion parameter reasoning model with 2 billion active parameters designed to run fast,...

Hugging Face Blog3 days ago

Black Forest Labs Releases FLUX 3 Image Klein

Black Forest Labs has launched FLUX 3 Image Klein, a compact open-weights image generation model that delivers high-end 4K visual synthesis on...

The Decoder1 day ago

Policy

Anthropic grants METR internal access for safety audits

Anthropic • 22 hrs ago

White House Gives UK AISI Access to Frontier AI

The Decoder • 4 days ago

OpenAI and Anthropic call for binding UK AI safety laws

WIRED AI • 3 days ago

406 Scientists Demand Immediate AI Internet Cutoff

The Decoder • 3 days ago

Anthropic grants METR internal access for safety audits

Anthropic has granted six staff members from evaluation nonprofit METR full internal access to audit its frontier AI models, establishing a new...

Anthropic22 hrs ago

White House Gives UK AISI Access to Frontier AI

The White House has restored the UK AI Safety Institute's access to top American AI models, allowing pre-release security testing to resume across...

The Decoder4 days ago

OpenAI and Anthropic call for binding UK AI safety laws

OpenAI and Anthropic backed binding statutory rules at a UK parliamentary hearing, creating a rift with Google DeepMind and Meta over whether...

WIRED AI3 days ago

406 Scientists Demand Immediate AI Internet Cutoff

Over 400 top researchers, including Geoffrey Hinton and Yoshua Bengio, are urging tech firms to completely disconnect AI models from the internet to...

The Decoder3 days ago

Culture

Linus Torvalds Uses Claude Code for Linux Kernel

Computerworld AI • 2 days ago

Anthropic's Andrej Karpathy details four tiers of AI usage

The Decoder • 4 days ago

OpenAI Ousts Three Key Safety Researchers

Don't Worry About the Vase • 4 days ago

OpenAI Engineer Arrested for Planting Hidden Bug

The Verge AI • 3 days ago

Linus Torvalds Uses Claude Code for Linux Kernel

Linux creator Linus Torvalds has begun using Anthropic's Claude Code assistant to help optimize kernel performance, marking a significant shift for a...

Computerworld AI2 days ago

Anthropic's Andrej Karpathy details four tiers of AI usage

Anthropic researcher Andrej Karpathy outlined a stark divide between public perception and true AI capabilities, warning that human comprehension is...

The Decoder4 days ago

OpenAI Ousts Three Key Safety Researchers

OpenAI has terminated three prominent AI safety researchers for alleged policy violations, sparking widespread concern over internal dissent and...

Don't Worry About the Vase4 days ago

OpenAI Engineer Arrested for Planting Hidden Bug

Federal authorities arrested an OpenAI engineer for placing a hidden listening device in an executive conference room, highlighting growing insider...

The Verge AI3 days ago

Research

Anthropic's Claude Discovers a Novel CRISPR-Like Enzyme System

Anthropic • 3 days ago

OpenAI Models Now Solve 35% of Science Theory Problems

WIRED AI • 2 days ago

Alibaba Unveils QwenGyre to Speed Up AI Agent RL

AlphaSignal • 4 days ago

Princeton Fixes AI Data Proxies Using Tiny Learning Rates

AlphaSignal • 1 day ago

Anthropic's Claude Discovers a Novel CRISPR-Like Enzyme System

Anthropic’s Claude AI has autonomously discovered a novel CRISPR-like enzyme system, demonstrating how generative models can accelerate biological...

Anthropic3 days ago

OpenAI Models Now Solve 35% of Science Theory Problems

An MIT study shows OpenAI models increased their success rate on theoretical science problems from 2 percent to 35 percent in two years, threatening...

WIRED AI2 days ago

Alibaba Unveils QwenGyre to Speed Up AI Agent RL

Alibaba’s Qwen team introduced QwenGyre, an elastic reinforcement learning framework that accelerates complex, hours-long AI agent training by up to...

AlphaSignal4 days ago

Princeton Fixes AI Data Proxies Using Tiny Learning Rates

Researchers from Princeton and partner institutions found that slashing learning rates by up to 100 times prevents cheap proxy models from giving...

AlphaSignal1 day ago

Hardware

NVIDIA Unifies GPU-Initiated Networking via DOCA GPUNetIO

NVIDIA Developer Blog • 1 day ago

Microsoft to Scale Maia AI Chips to Gigawatts by 2027

The Decoder • 4 days ago

Microsoft Unveils Maia 300 AI Chip to Cut GPU Reliance

The Verge AI • 4 days ago

Atomic Chat Runs a 177B Qwen Model on a 64 GB Mac

AlphaSignal • 4 days ago

NVIDIA Unifies GPU-Initiated Networking via DOCA GPUNetIO

NVIDIA has unified its GPU-initiated networking stack under the DOCA GPUNetIO framework, bypassing CPU bottlenecks to slash latency for distributed...

NVIDIA Developer Blog1 day ago

Microsoft to Scale Maia AI Chips to Gigawatts by 2027

Microsoft is expanding its custom Azure Maia AI chips to multi-gigawatt deployments by 2027, signaling a major move to reduce its reliance on...

The Decoder4 days ago

Microsoft Unveils Maia 300 AI Chip to Cut GPU Reliance

Microsoft unveiled its Maia 300 AI chip, a custom processor arriving in 2027 designed to act as its primary inference engine and curb its dependence...

The Verge AI4 days ago

Atomic Chat Runs a 177B Qwen Model on a 64 GB Mac

Atomic Chat has released specialized GGUF quantizations allowing a 177-billion-parameter Qwen model to run at 36 tokens per second on a 64 GB Mac by...

AlphaSignal4 days ago

Business

Anthropic Replaces Git With Scion After Claude Code Surge

AlphaSignal • 3 days ago

NVIDIA Open-Sources Full Run:ai Platform

NVIDIA Developer Blog • 1 day ago

Spotify Doubles Monthly Code Changes with AI Agents

InfoQ AI • 4 days ago

Epic Games shifts to AI-first dev using Anthropic Claude

Ars Technica AI • 4 days ago

Anthropic Replaces Git With Scion After Claude Code Surge

Anthropic replaced Git with an in-house version control platform called Scion after automated code generation by Claude overwhelmed its traditional...

AlphaSignal3 days ago

NVIDIA Open-Sources Full Run:ai Platform

NVIDIA has released its complete Run:ai platform as open-source software, offering developers modular Kubernetes tools to maximize GPU utilization...

NVIDIA Developer Blog1 day ago

Spotify Doubles Monthly Code Changes with AI Agents

Spotify more than doubled its monthly merged code changes to 17,000 using AI, forcing the streaming giant to revamp its testing and safeguards as...

InfoQ AI4 days ago

Epic Games shifts to AI-first dev using Anthropic Claude

Epic Games is pivoting to an AI-first development approach using models like Anthropic's Claude, marking a major shift in how AAA video games will be...

Ars Technica AI4 days ago

Agents

Anthropic Launches Dynamic Workflows for Claude Agents

Latent Space • 4 days ago

Cohere Cuts Multi-Agent AI Costs by 70 Percent

Cohere Blog • 2 days ago

Alibaba Launches QwenGyre for Faster AI Agent Training

MarkTechPost • 4 days ago

Dyna Robotics Launches Dyna-2.1 for Long-Horizon Autonomy

Latent Space • 2 days ago

Anthropic Launches Dynamic Workflows for Claude Agents

Anthropic introduced dynamic workflows for Claude Managed Agents, allowing its lead AI model to automatically design, program, and coordinate...

Latent Space4 days ago

Cohere Cuts Multi-Agent AI Costs by 70 Percent

Cohere demonstrated that structuring coding agent teams with a strong lead and cheaper worker models cuts inference costs by up to 70% while...

Cohere Blog2 days ago

Alibaba Launches QwenGyre for Faster AI Agent Training

Alibaba's Qwen team introduced QwenGyre, an elastic reinforcement learning framework designed to accelerate long-horizon coding agent training by up...

MarkTechPost4 days ago

Dyna Robotics Launches Dyna-2.1 for Long-Horizon Autonomy

Dyna Robotics has introduced Dyna-2.1, a semi-humanoid robot powered by a world-action model trained on over one million hours of video, achieving...

Latent Space2 days ago