AI News — Saturday, April 11, 2026

10
Stalking Victim Sues OpenAI, Claims ChatGPT Fueled Abuser's Delusions

A stalking victim is suing OpenAI, alleging that ChatGPT provided information that exacerbated her abuser's delusions and that the company failed to act on her warnings.

TechCrunchindustry
9
Anthropic Launches 'Cowork' Desktop Agent for Claude, No Coding Required

Anthropic has introduced Cowork, a new Claude Desktop agent designed to work directly with user files without requiring any coding expertise, enhancing accessibility for everyday tasks.

VentureBeatproduct
9
Railway Secures $100 Million to Build AI-Native Cloud Infrastructure, Challenging AWS

Railway has raised $100 million to develop an AI-native cloud infrastructure, positioning itself as a direct competitor to established giants like AWS.

VentureBeatindustry
8
Nous Research Releases NousCoder-14B, an Open-Source Coding Model

Nous Research has launched NousCoder-14B, an open-source coding model, entering the rapidly evolving landscape of AI-powered code generation tools.

VentureBeatopen-source
8
Anthropic Temporarily Banned OpenClaw's Creator from Accessing Claude

Anthropic temporarily suspended the creator of OpenClaw from using Claude, raising questions about API usage policies and the interaction between AI developers and platform providers.

TechCrunchindustry
8
Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability

New research explores how optimization, data, and model capabilities conditionally influence generalization in reasoning-focused Supervised Fine-Tuning (SFT) for language models.

Hugging Faceresearch
7
SkillClaw: Let Skills Evolve Collectively with Agentic Evolver

SkillClaw introduces an 'Agentic Evolver' framework that allows AI agents' skills to evolve collectively, potentially leading to more robust and adaptable AI systems.

Hugging Faceresearch
7
HY-Embodied-0.5: Embodied Foundation Models for Real-World Agents

Researchers present HY-Embodied-0.5, a new embodied foundation model designed to enable real-world AI agents to perform complex tasks in physical environments.

Hugging Faceresearch
7
Google Introduces New Ways to Balance Cost and Reliability in Gemini API

Google has announced new features for the Gemini API, offering developers more flexibility to balance inference cost and reliability for their AI applications.

Google AI Blogproduct
6
When Numbers Speak: Aligning Textual Numerals and Visual Instances in Text-to-Video Diffusion Models

This paper investigates methods to improve text-to-video diffusion models by better aligning textual numerals with their corresponding visual instances in generated content.

Hugging Faceresearch
6
ClawBench: Can AI Agents Complete Everyday Online Tasks?

ClawBench introduces a new benchmark to evaluate the capability of AI agents in completing a wide range of everyday online tasks, assessing their practical utility.

Hugging Faceresearch
6
MegaStyle: Constructing Diverse and Scalable Style Dataset via Consistent Text-to-Image Style Mapping

MegaStyle proposes a method for creating diverse and scalable style datasets by consistently mapping textual descriptions to image styles in text-to-image generation.

Hugging Faceresearch
6
Google AI to Create New Satellite Imagery Map for Protecting Brazil's Forests

Google AI is developing a new satellite imagery map specifically designed to aid in the protection and monitoring of Brazil's vital forest ecosystems.

Google AI Blogproduct
5
A Simple React Hook for Running Local LLMs via WebGPU

A new React hook has been developed to simplify the process of running local Large Language Models directly in the browser using WebGPU, enhancing client-side AI capabilities.

Dev.toopen-source
5
"I Pointed Claude Code at My Local Ollama Models — Here's the 3-Minute Setup"

A developer shares a quick guide on how to integrate Claude Code with local Ollama models, demonstrating an easy setup for leveraging powerful AI locally.

Dev.toopen-source