Latest News

OpenAI Debuts Agents API to Power Autonomous Workflows

OpenAI has launched its new Agents API, allowing developers to build autonomous AI assistants that can run code in sandboxes, connect to external servers, and delegate tasks.

Hacker News1 day agoAgents

Models

Inception Mercury 2.5 Hits 1,107 Tokens Per Second

AlphaSignal • 3 days ago

World Labs Unveils Atlas to Unify 3D and Video Generation

AlphaSignal • 3 days ago

humans& debuts Persimmon model to mimic human chat

AlphaSignal • 1 day ago

OpenBMB's MiniCPM5-2B Dominates Small Model Benchmarks

AlphaSignal • 4 days ago

Inception Mercury 2.5 Hits 1,107 Tokens Per Second

Inception has released Mercury 2.5, a massive diffusion-based language model designed to slash latency and costs in multi-step AI agent workflows by...

AlphaSignal3 days ago

World Labs Unveils Atlas to Unify 3D and Video Generation

Spatial intelligence startup World Labs has unveiled Atlas, a foundation model that merges 3D reconstruction and video generation to simplify virtual...

AlphaSignal3 days ago

humans& debuts Persimmon model to mimic human chat

AI startup humans& has released Persimmon, a 550-billion parameter model designed to realistically simulate human behavior in group chats to improve...

AlphaSignal1 day ago

OpenBMB's MiniCPM5-2B Dominates Small Model Benchmarks

OpenBMB has released MiniCPM5-2B, a dense 2.6-billion-parameter reasoning model that tops small-model benchmarks while using three times fewer tokens...

AlphaSignal4 days ago

Policy

Anthropic Hardens Security After Claude Escapes Sandbox

Anthropic • 4 days ago

Anthropic Patches Claude Flaws Found by US and UK Agencies

Anthropic • 1 day ago

DOJ Backs OpenAI in New York Times Copyright Lawsuit

Mindstream AI • 4 days ago

US accuses DeepSeek and Alibaba of copying frontier models

Ars Technica AI • 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...

Anthropic4 days ago

Anthropic Patches Claude Flaws Found by US and UK Agencies

Anthropic has partnered with US and UK safety institutes to stress-test its Claude models, uncovering critical vulnerabilities that prompted a...

Anthropic1 day ago

DOJ Backs OpenAI in New York Times Copyright Lawsuit

The U.S. Justice Department has backed OpenAI in its copyright battle with The New York Times, arguing that training AI models on published articles...

Mindstream AI4 days ago

US accuses DeepSeek and Alibaba of copying frontier models

Three US intelligence and security agencies have accused six Chinese AI companies of executing industrial-scale distillation attacks to copy...

Ars Technica AI2 days ago

Culture

Web Scrapers Overwhelm Linux Kernel Git Infrastructure

Simon Willison • 4 days ago

Synthesia and Amazon Overhaul Code Reviews to Fight AI Slop

IEEE Spectrum AI • 3 days ago

Meta Drops AI Metrics From Engineer Reviews

The Decoder • 3 days ago

DeepMind Banned AI Extinction Risk Talk, Ex-Staffer Says

The Decoder • 1 day ago

Web Scrapers Overwhelm Linux Kernel Git Infrastructure

The official Git repository for the Linux kernel is spending more CPU power rendering HTML pages for automated scrapers than serving legitimate...

Simon Willison4 days ago

Synthesia and Amazon Overhaul Code Reviews to Fight AI Slop

As AI tools flood repositories with flawed code, engineering teams at companies like Synthesia and Amazon are redesigning their review processes to...

IEEE Spectrum AI3 days ago

Meta Drops AI Metrics From Engineer Reviews

Meta has removed AI tool usage metrics from its software engineer performance reviews after employees began artificially inflating their token...

The Decoder3 days ago

DeepMind Banned AI Extinction Risk Talk, Ex-Staffer Says

A former DeepMind PR staffer revealed the lab once banned employees from publicly discussing AI extinction risks, exposing a historical gap between...

The Decoder1 day ago

Research

Magic Matches DeepSeek V4 Pro Base Using 50x Less Compute

AlphaSignal • 3 days ago

DeepMind Maps 9 Billion DNA Variants with AlphaGenome Atlas

IEEE Spectrum AI • 3 days ago

Insilico Medicine AI Drug Lowers Biological Age in Trial

The Decoder • 4 days ago

Google ToolGrad Framework Boosts AI Tool-Use Training

Google Research • 1 day ago

Magic Matches DeepSeek V4 Pro Base Using 50x Less Compute

AI startup Magic has developed a pretraining recipe that matches DeepSeek V4 Pro Base using 50 times less compute, suggesting algorithmic efficiency...

AlphaSignal3 days ago

DeepMind Maps 9 Billion DNA Variants with AlphaGenome Atlas

Google DeepMind has launched the AlphaGenome Atlas, a free database mapping nine billion genetic mutations to help scientists understand how DNA...

IEEE Spectrum AI3 days ago

Insilico Medicine AI Drug Lowers Biological Age in Trial

A clinical trial of Insilico Medicine's AI-designed drug rentosertib showed it reduced patients' biological age markers by up to six years, proving...

The Decoder4 days ago

Google ToolGrad Framework Boosts AI Tool-Use Training

Google researchers have introduced ToolGrad, an 'answer-first' dataset generation framework that allows smaller open-source models to match...

Google Research1 day ago

Hardware

ASML secures TSMC and Samsung for High-NA EUV tools

The Decoder • 3 days ago

Cohere Megakernel Beats vLLM by 1.58x on H100 GPUs

AlphaSignal • 3 days ago

Google to Invest $15B in Finland’s AI Infrastructure

AI Business • 1 day ago

Apple Debuts iPhone 18 Pro With A20 Pro AI Chip

The Neuron • 2 days ago

ASML secures TSMC and Samsung for High-NA EUV tools

ASML has secured commitments from TSMC and Samsung to adopt its $400 million High-NA EUV lithography systems, cementing the Dutch firm's control over...

The Decoder3 days ago

Cohere Megakernel Beats vLLM by 1.58x on H100 GPUs

Cohere open-sourced a megakernel serving engine that runs entire LLM decode steps as a single CUDA kernel, boosting inference speeds on Nvidia H100...

AlphaSignal3 days ago

Google to Invest $15B in Finland’s AI Infrastructure

Google is investing $15.1 billion in Finnish AI infrastructure and securing nuclear power to fuel its regional data centers, marking its largest...

AI Business1 day ago

Apple Debuts iPhone 18 Pro With A20 Pro AI Chip

Apple has unveiled its new iPhone 18 Pro powered by the 2-nanometer A20 Pro chip, marking a major push to run advanced Apple Intelligence and Siri...

The Neuron2 days ago

Business

Anthropic Secures $517 Billion in Compute Contracts

The Decoder • 4 days ago

Mistral AI raises €3B to scale sovereign open-weight models

Mistral AI Blog • 3 days ago

OpenAI Pauses $200 Pro Plan Due to High Astra Demand

TechCrunch AI • 1 day ago

Cognition raises $2B at $48B valuation for Devin AI

TechCrunch AI • 3 days ago

Anthropic Secures $517 Billion in Compute Contracts

Anthropic has reportedly secured up to $517 billion in computing contracts over eleven months, signaling an aggressive push to match rival OpenAI...

The Decoder4 days ago

Mistral AI raises €3B to scale sovereign open-weight models

Mistral AI has raised €3 billion in a Series D round valuing it at over €21 billion, securing the largest-ever European tech fundraise to expand its...

Mistral AI Blog3 days ago

OpenAI Pauses $200 Pro Plan Due to High Astra Demand

OpenAI has paused new sign-ups for its $200-per-month Pro subscription tier as unprecedented demand for its new Astra model strains the company's...

TechCrunch AI1 day ago

Cognition raises $2B at $48B valuation for Devin AI

AI coding startup Cognition has raised $2 billion at a $48 billion valuation, proving that venture capitalists believe the market for automated...

TechCrunch AI3 days ago

Agents

OpenAI deploys AI research interns and warns of risks

The Decoder • 4 days ago

Cognition's Devin Agent Breaks RSA-260 Factoring Record

AlphaSignal • 2 days ago

Apple Adds Anthropic Claude Agent SDK to Xcode

Anthropic • 2 days ago

Meta debuts Muse AI agent to automate everyday tasks

TechCrunch AI • 3 days ago

OpenAI deploys AI research interns and warns of risks

OpenAI has deployed automated AI research interns to accelerate its own development, even as its chief scientist warns that current alignment and...

The Decoder4 days ago

Cognition's Devin Agent Breaks RSA-260 Factoring Record

Cognition's Devin AI agent has factored RSA-260 to set a new cryptography record, proving that autonomous agents can dramatically lower the expertise...

AlphaSignal2 days ago

Apple Adds Anthropic Claude Agent SDK to Xcode

Apple has integrated Anthropic's Claude Agent SDK into Xcode 26.3, allowing developers to run autonomous, long-running coding tasks and visually...

Anthropic2 days ago

Meta debuts Muse AI agent to automate everyday tasks

Meta has launched Muse, a personal AI agent powered by its Muse Spark model that automates everyday tasks like booking travel and making purchases,...

TechCrunch AI3 days ago