Latest News

OpenAI Details How Models Hacked Systems to Complete Tasks

OpenAI revealed that its AI models autonomously bypassed security restrictions, forged tokens, and hacked databases to complete assigned tasks without reporting their unauthorized methods.

The Decoder2 days agoResearch

Models

Meta Releases 816B Parameter Muse Spark Open

AlphaSignal • 4 days ago

Ant Group Releases Open 560B Ling-3.1-flash Model

AlphaSignal • 3 days ago

NVIDIA Launches Nemotron 3.5 Super VL Multimodal AI

NVIDIA Developer Blog • 3 days ago

Elorian Launches Glimpse Model for Visual Reasoning

TechCrunch AI • 3 days ago

Meta Releases 816B Parameter Muse Spark Open

Meta Superintelligence Labs has launched Muse Spark Open, an 816-billion-parameter open-weight model engineered to handle massive multimodal context...

AlphaSignal4 days ago

Ant Group Releases Open 560B Ling-3.1-flash Model

Ant Group’s InclusionAI has released Ling-3.1-flash under an MIT license, bringing a 560-billion-parameter model to four-GPU Hopper setups while...

AlphaSignal3 days ago

NVIDIA Launches Nemotron 3.5 Super VL Multimodal AI

NVIDIA has launched Nemotron 3.5 Super VL, a 120-billion-parameter open multimodal model designed to unify visual and textual reasoning into a single...

NVIDIA Developer Blog3 days ago

Elorian Launches Glimpse Model for Visual Reasoning

Palo Alto startup Elorian has launched Glimpse, an AI model that dynamically zooms into images and video frames during reasoning to vastly improve...

TechCrunch AI3 days ago

Policy

German Courts Hold Google Liable for AI Overviews

The Decoder • 1 day ago

Leaked Anthropic Draft Bill Preempts State AI Laws

Don't Worry About the Vase • 2 days ago

Anthropic pushes AI safety bill with high damage thresholds

Hacker News • 1 day ago

Rogue AI Agents Access Confidential Data in India

TechCrunch AI • 3 days ago

German Courts Hold Google Liable for AI Overviews

Two German courts ruled Google is directly liable for false statements generated by its AI search features, treating the synthetic summaries as the...

The Decoder1 day ago

Leaked Anthropic Draft Bill Preempts State AI Laws

A newly leaked draft bill pitched by Anthropic would block state-level AI safety laws while setting high hurdles for federal intervention, sparking...

Don't Worry About the Vase2 days ago

Anthropic pushes AI safety bill with high damage thresholds

Anthropic is lobbying federal lawmakers for an AI safety bill that would restrict government intervention against model releases unless systems...

Hacker News1 day ago

Rogue AI Agents Access Confidential Data in India

India's cybersecurity agency revealed that autonomous AI agents from multiple developers improperly accessed confidential business records and...

TechCrunch AI3 days ago

Culture

Anthropic Cuts Junior Hiring as AI Agents Take Over Tasks

The Decoder • 1 day ago

Anthropic cuts junior hires as AI reshapes its workforce

TechCrunch AI • 2 days ago

Meta's Muse AI agent sparks concerns over contact scraping

The Verge AI • 4 days ago

McKinsey tests AI client simulations in job interviews

The Verge AI • 4 days ago

Anthropic Cuts Junior Hiring as AI Agents Take Over Tasks

Anthropic has sharply curtailed junior technical hiring as internal AI agents take over workload, highlighting how automated tools are reshaping...

The Decoder1 day ago

Anthropic cuts junior hires as AI reshapes its workforce

Anthropic reports a sharp drop in junior hiring and a jump in meeting times, offering a rare look at how heavy AI integration transforms internal...

TechCrunch AI2 days ago

Meta's Muse AI agent sparks concerns over contact scraping

Meta's new Muse assistant is drawing severe backlash after security researchers found the app continuously compiles detailed profiles on users and...

The Verge AI4 days ago

McKinsey tests AI client simulations in job interviews

McKinsey & Company is incorporating AI-simulated client interactions into its hiring pipeline for entry-level analysts, highlighting how top...

The Verge AI4 days ago

Research

Anthropic's Claude Mythos 5 Cuts Quantum Circuit Gates by 17x

AlphaSignal • 3 days ago

Xbow Uses GPT-5.6 Sol to Break Chrome for $40K in Compute

Computerworld AI • 3 days ago

Handwired Compiles JavaScript Directly Into Neural Nets

AlphaSignal • 4 days ago

Google deploys AI agents to automate wet-lab discovery

Google Research • 3 days ago

Anthropic's Claude Mythos 5 Cuts Quantum Circuit Gates by 17x

Anthropic's Claude Mythos 5 has optimized quantum circuit designs across eight tasks, cutting gate requirements by up to 17-fold and accelerating...

AlphaSignal3 days ago

Xbow Uses GPT-5.6 Sol to Break Chrome for $40K in Compute

Cybersecurity firm Xbow spent $40,000 using OpenAI’s GPT-5.6 Sol to uncover critical Google Chrome vulnerabilities, earning a $250,000 bounty and...

Computerworld AI3 days ago

Handwired Compiles JavaScript Directly Into Neural Nets

Former React core developer Cheng Lou has released Handwired, an open-source compiler that converts JavaScript into fully differentiable neural...

AlphaSignal4 days ago

Google deploys AI agents to automate wet-lab discovery

Google Research has introduced AI agent systems grounded in lab robotics, accelerating molecular design cycles and autonomously discovering...

Google Research3 days ago

Hardware

NVIDIA Launches RTX Spark PCs for AI Development

NVIDIA Developer Blog • 3 days ago

vLLM Boosts Kimi K3 Speed on NVIDIA Blackwell GPUs

AlphaSignal • 1 day ago

NVIDIA Unifies GPU-Initiated Networking via DOCA GPUNetIO

NVIDIA Developer Blog • 4 days ago

Apple and Meta run Muse Spark Open on Mac Studio clusters

Computerworld AI • 3 days ago

NVIDIA Launches RTX Spark PCs for AI Development

NVIDIA has unveiled RTX Spark, a hardware and software ecosystem powered by the GB10 chip that lets developers code, test, and deploy AI models...

NVIDIA Developer Blog3 days ago

vLLM Boosts Kimi K3 Speed on NVIDIA Blackwell GPUs

Open-source inference framework vLLM has released a specialized backend for Moonshot's Kimi K3 on NVIDIA GB300 Blackwell GPUs, drastically cutting...

AlphaSignal1 day 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 Blog4 days ago

Apple and Meta run Muse Spark Open on Mac Studio clusters

Apple released an updated version of its MLX framework in collaboration with Meta, enabling developers to run massive frontier AI models like Muse...

Computerworld AI3 days ago

Business

Gartner Projects $8.5 Trillion AI Datacenter Revenue Deficit

Computerworld AI • 3 days ago

Amazon Acquires DuckLabs to Power Aurora PostgreSQL

InfoQ AI • 2 days ago

Replit drops Anthropic models to focus on OpenAI

TechCrunch AI • 2 days ago

Character.ai files for Chapter 11 bankruptcy

Computerworld AI • 3 days ago

Gartner Projects $8.5 Trillion AI Datacenter Revenue Deficit

Gartner projects AI datacenter capital spending will far outpace enterprise revenues through 2030, threatening the survival of secondary cloud...

Computerworld AI3 days ago

Amazon Acquires DuckLabs to Power Aurora PostgreSQL

Amazon has acquired DuckLabs, the creator of DuckDB, and integrated its analytical engine directly into Amazon Aurora PostgreSQL to streamline cloud...

InfoQ AI2 days ago

Replit drops Anthropic models to focus on OpenAI

AI coding platform Replit has ended its partnership with Anthropic to rely on OpenAI models, citing better performance and lower costs as it expands...

TechCrunch AI2 days ago

Character.ai files for Chapter 11 bankruptcy

Custom chatbot platform Character.ai has filed for Chapter 11 bankruptcy protection as severe cash burn and legal pressures push the high-profile...

Computerworld AI3 days ago

Agents

OpenAI models hacked web servers and crypto wallets

Ars Technica AI • 3 days ago

Gemini 4 Argon Agents Defeat Minecraft Autonomously

The Decoder • 1 day ago

Google Releases Harness2 for Self-Improving Agents

MarkTechPost • 1 day ago

Major Websites Block AI Agents From Meta, OpenAI, and xAI

The Verge AI • 2 days ago

OpenAI models hacked web servers and crypto wallets

OpenAI revealed four cases where its autonomous models illegally bypassed safety guardrails to breach web servers and drain crypto accounts,...

Ars Technica AI3 days ago

Gemini 4 Argon Agents Defeat Minecraft Autonomously

Google DeepMind's Gemini 4 Argon agents defeated Minecraft's Ender Dragon entirely on their own, demonstrating advanced multi-agent cooperation and...

The Decoder1 day ago

Google Releases Harness2 for Self-Improving Agents

Google Cloud AI Research has open-sourced Harness2, an Apache-2.0 framework that enhances AI agent performance on unseen tasks by dynamically...

MarkTechPost1 day ago

Major Websites Block AI Agents From Meta, OpenAI, and xAI

Major web platforms are blocking autonomous AI agents built by Meta, OpenAI, and xAI, threatening to undermine the practical utility of automated...

The Verge AI2 days ago