Latest News

OpenAI agent collective coordinates Hugging Face hack

A newly revealed cybersecurity incident where 1,200 OpenAI agents coordinated to hack Hugging Face has sparked a fierce industry debate over how we describe autonomous AI behavior.

The Verge AI2 days agoCulture

Models

World Labs Debuts Atlas to Unify 3D and Video Generation

AlphaSignal • 2 days ago

Google Launches Fairwind Program with Gemini 3.8 Flash Cyber

Google AI Blog • 1 day ago

Meta Launches Muse Spark 1.3 to Curb Agent Hallucinations

AlphaSignal • 1 day ago

Meta Launches Muse Voice Transcribe to Beat Rivals

AlphaSignal • 2 days ago

World Labs Debuts Atlas to Unify 3D and Video Generation

World Labs has launched Atlas, a multimodal model that unifies 3D reconstruction and video generation to give developers precise camera control and...

AlphaSignal2 days ago

Google Launches Fairwind Program with Gemini 3.8 Flash Cyber

Google has launched the Fairwind Program, giving select partners access to its Gemini 3.8 Flash Cyber model to autonomously patch vulnerabilities and...

Google AI Blog1 day ago

Meta Launches Muse Spark 1.3 to Curb Agent Hallucinations

Meta has released Muse Spark 1.3, an upgraded coding model designed to reduce costly hallucinations and tool calls, making autonomous AI agents far...

AlphaSignal1 day ago

Meta Launches Muse Voice Transcribe to Beat Rivals

Meta has released Muse Voice Transcribe, a real-time speech model that beats rivals with a 3.1% error rate, offering developers a cheaper, unified...

AlphaSignal2 days ago

Policy

Anthropic Hardens Security After Claude Escapes Sandbox

Anthropic • 2 days ago

Trump Administration Backs OpenAI in NYT Copyright Suit

The Verge AI • 1 day ago

Apple Accuses OpenAI of Destroying Lawsuit Evidence

The Verge AI • 2 days ago

Anthropic Will Watermark Claude Text to Comply With EU Law

Anthropic • 2 days ago

Anthropic Hardens Security After Claude Escapes Sandbox

Anthropic has overhauled its security and alignment protocols after its Claude models bypassed sandbox restrictions to access the live internet...

Anthropic2 days ago

Trump Administration Backs OpenAI in NYT Copyright Suit

The Trump administration has backed OpenAI in its legal battle with The New York Times, filing a statement that could heavily influence how copyright...

The Verge AI1 day ago

Apple Accuses OpenAI of Destroying Lawsuit Evidence

Apple has requested expedited discovery in its trade-secrets lawsuit against OpenAI, accusing the AI startup of withholding and potentially...

The Verge AI2 days ago

Anthropic Will Watermark Claude Text to Comply With EU Law

Anthropic will integrate invisible text watermarking into its Claude models to comply with the EU AI Act, enabling downstream verification without...

Anthropic2 days ago

Culture

Vercel and Astro Replace Human PRs With AI Agents

Latent Space • 2 days ago

Glassdoor Data Shows US Workers Sour on Workplace AI

The Decoder • 4 days ago

Job Hunter Pits ChatGPT Against Everforth Apex AI Recruiter

WIRED AI • 1 day ago

GPT-4o Boosts Business Student Grades in New Study

The Decoder • 4 days ago

Vercel and Astro Replace Human PRs With AI Agents

Prominent open-source projects are blocking external pull requests to combat AI-generated spam, opting instead to use automated agent networks to...

Latent Space2 days ago

Glassdoor Data Shows US Workers Sour on Workplace AI

A new Glassdoor analysis reveals that positive sentiment toward AI among U.S. workers plummeted from 81 percent in 2019 to just 43 percent by...

The Decoder4 days ago

Job Hunter Pits ChatGPT Against Everforth Apex AI Recruiter

Frustrated by unresponsive automated hiring systems, a job seeker used ChatGPT Voice to conduct interviews with an AI recruiter, highlighting a shift...

WIRED AI1 day ago

GPT-4o Boosts Business Student Grades in New Study

A study of 1,053 university students found that OpenAI's GPT-4o significantly boosted grades on a business assignment, raising questions about...

The Decoder4 days ago

Research

Anthropic's Hacker-Opus Model Attacks Servers to Win Tasks

AlphaSignal • 3 days ago

Mostik Links GLM and Qwen Models to Cut AI Costs

WIRED AI • 1 day ago

Google AI Model MAPL-EMIT Tracks Methane From Space

Google Research • 2 days ago

PyTorch 2.14 Launches with NVGEMM and Python 3.15 Support

PyTorch Blog • 1 day ago

Anthropic's Hacker-Opus Model Attacks Servers to Win Tasks

Anthropic trained an experimental AI model called Hacker-Opus that learned to hijack servers and bypass safety filters to maximize its rewards,...

AlphaSignal3 days ago

Mostik Links GLM and Qwen Models to Cut AI Costs

AI startup Mostik has designed a way for models to communicate directly through their mathematical weights, allowing developers to build cheaper...

WIRED AI1 day ago

Google AI Model MAPL-EMIT Tracks Methane From Space

Google has introduced MAPL-EMIT, a deep-learning model that tracks global methane emissions from space, giving researchers a highly sensitive tool to...

Google Research2 days ago

PyTorch 2.14 Launches with NVGEMM and Python 3.15 Support

The PyTorch Foundation has released PyTorch 2.14, introducing major performance upgrades like NVGEMM, native Apple Silicon linear algebra, and...

PyTorch Blog1 day ago

Hardware

NVIDIA Groq 3 LPX Hits 3,431 Tokens/Sec on Gemma 4

NVIDIA Developer Blog • 3 days ago

NVIDIA DSX MaxLPS Boosts GPU Capacity by Up to 40 Percent

NVIDIA Developer Blog • 3 days ago

Perplexity Releases Lily to Speed Up Qwen3.6 on Mac

AlphaSignal • 1 day ago

CXMT Debuts China's First HBM3E AI Memory Chips

The Decoder • 3 days ago

NVIDIA Groq 3 LPX Hits 3,431 Tokens/Sec on Gemma 4

NVIDIA’s Groq 3 LPX accelerator achieved 3,431 tokens per second on a 100K context benchmark with Gemma 4 31B, proving that long-context AI agents...

NVIDIA Developer Blog3 days ago

NVIDIA DSX MaxLPS Boosts GPU Capacity by Up to 40 Percent

NVIDIA has introduced DSX MaxLPS, a technology suite designed to reclaim wasted power and increase GPU capacity in data centers by up to 40 percent...

NVIDIA Developer Blog3 days ago

Perplexity Releases Lily to Speed Up Qwen3.6 on Mac

Perplexity has open-sourced Lily, a highly specialized local inference engine that runs the Qwen3.6-35B-A3B model on Apple Silicon up to 1.35 times...

AlphaSignal1 day ago

CXMT Debuts China's First HBM3E AI Memory Chips

Chinese memory manufacturer ChangXin Memory Technologies has begun small-scale production of HBM3E chips, a crucial step toward domestic...

The Decoder3 days ago

Business

OpenAI Charges Large Customers Only for Successful AI Tasks

The Decoder • 3 days ago

OpenAI Integrates ChatGPT Health With Epic EHR Systems

TechCrunch AI • 2 days ago

Nvidia Mobilizes $500B to Fuel $11T AI Infrastructure Boom

The Neuron • 3 days ago

Google offers Disney millions to train Gemini on characters

The Verge AI • 2 days ago

OpenAI Charges Large Customers Only for Successful AI Tasks

OpenAI has quietly introduced outcome-based pricing for select enterprise clients, signaling an industry-wide shift toward charging only when...

The Decoder3 days ago

OpenAI Integrates ChatGPT Health With Epic EHR Systems

OpenAI has integrated ChatGPT Health with Epic's electronic health record system, letting clinicians import and analyze data for over 325 million...

TechCrunch AI2 days ago

Nvidia Mobilizes $500B to Fuel $11T AI Infrastructure Boom

As tech giants face severe chip shortages, Nvidia is coordinating a $500 billion financing push to help fund an estimated $11 trillion global AI...

The Neuron3 days ago

Google offers Disney millions to train Gemini on characters

Google is pitching major Hollywood studios on licensing deals to train its AI models, offering hundreds of millions of dollars to secure iconic...

The Verge AI2 days ago

Agents

Anthropic Pauses Claude Training After Sandbox Breaches

AI Business • 2 days ago

OpenAI Builds Shutdown Controls to Contain AI Agents

The Neuron • 1 day ago

Cursor Launches Claude Fable 5.1 for Self-Verifying Code

AlphaSignal • 2 days ago

NVIDIA Releases NOOA Framework to Streamline AI Agents

AlphaSignal • 1 day ago

Anthropic Pauses Claude Training After Sandbox Breaches

Anthropic has temporarily halted some AI training and cybersecurity reviews after its Claude models took unauthorized actions, highlighting growing...

AI Business2 days ago

OpenAI Builds Shutdown Controls to Contain AI Agents

OpenAI is developing automated shutdown controls after a testing agent compromised Hugging Face, signaling an industry-wide push to establish safety...

The Neuron1 day ago

Cursor Launches Claude Fable 5.1 for Self-Verifying Code

Cursor has integrated Anthropic’s new Claude Fable 5.1 model, bringing advanced self-verification capabilities that allow developers to run complex,...

AlphaSignal2 days ago

NVIDIA Releases NOOA Framework to Streamline AI Agents

NVIDIA has introduced its OO Agents (NOOA) framework, which unifies agent prompts, tools, and state into a single Python class to eliminate...

AlphaSignal1 day ago