Thursday, June 18, 2026

28 stories
AI Enhances Drug Synthesis Process
OpenAIResearch

AI Enhances Drug Synthesis Process

OpenAI and Molecule.one developed a nearly autonomous AI chemist utilizing GPT-5.4. The system successfully optimized a difficult chemical reaction used in medicinal research.

Read the source →
Google DeepMind to Streamline UK Housing Approvals
Google DeepMindProduct

Google DeepMind to Streamline UK Housing Approvals

The UK government is collaborating with Google DeepMind to develop an AI prototype. This tool is designed to accelerate the decision-making process for residential construction planning.

Read the source →
Medical AI Matches Doctors in Disease Management
Google AIResearch

Medical AI Matches Doctors in Disease Management

New study results published in Nature indicate that Google's AMIE system performs as well as primary care physicians. The AI demonstrates a strong ability to help manage complex health issues through conversation.

Read the source →
MolmoMotion Predicts 3D Movement
Hugging FaceResearch

MolmoMotion Predicts 3D Movement

Researchers have introduced MolmoMotion, a system designed to forecast 3D motion. The model uses natural language instructions to guide its movement predictions.

Read the source →
France Implements National AI Infrastructure
NVIDIAPolicy

France Implements National AI Infrastructure

France is activating the AI factories and computing capacities planned a year ago with NVIDIA's help. This infrastructure now supports the deployment of AI agents and startup applications across the country.

Read the source →
SageMaker AI Updates Async Inference
AWS Machine LearningProduct

SageMaker AI Updates Async Inference

Amazon has introduced inline payload support for SageMaker AI Async Inference. This update allows users to include data directly in API requests instead of uploading files to Amazon S3 first.

Read the source →
Disabling AI Features in Google Docs
TechCrunchProduct

Disabling AI Features in Google Docs

Google Docs users can now remove the automated Gemini writing prompts from their workspace. A new guide explains the steps required to turn off these AI pop-ups.

Read the source →
Midjourney Enters Health Hardware
The VergeProduct

Midjourney Enters Health Hardware

CEO David Holz introduced the Midjourney Scanner, a device designed for full-body ultrasound imaging. The company also intends to establish a wellness spa in San Francisco.

Read the source →
White House Orders Access Cut for SK Telecom
WiredPolicy

White House Orders Access Cut for SK Telecom

The US government directed Anthropic to block SK Telecom from using the Claude Mythos model. This action followed concerns regarding the Korean company's reported links to China.

Read the source →
OpenAI Recruits Former Google Executive
The DecoderOther

OpenAI Recruits Former Google Executive

Noam Shazeer, a key contributor to the Gemini project and the original Attention paper, has moved to OpenAI. This transition follows a brief return to Google through a multi-billion dollar agreement with Character.AI.

Read the source →
AI Agents Assist Robots in Hardware Tasks
Ars TechnicaResearch

AI Agents Assist Robots in Hardware Tasks

Nvidia has developed a system where AI coding agents train robots to perform technical manual labor. These robots successfully learned to remove zip ties and install GPUs.

Read the source →
OpenAI Debuts LifeSciBench
OpenAIResearch

OpenAI Debuts LifeSciBench

OpenAI has released a new evaluation tool called LifeSciBench. Developed and vetted by specialists, it tests the ability of AI models to manage complex decisions and tasks within the life sciences.

Read the source →
Hugging Face Bridges Software and Robotics
Hugging FaceProduct

Hugging Face Bridges Software and Robotics

The Hugging Face Hub is now connecting with physical robot hardware. This integration is powered by LeRobot and Strands Agents.

Read the source →
NVIDIA Launches Beta for XR AI Framework
NVIDIAProduct

NVIDIA Launches Beta for XR AI Framework

NVIDIA has released a public beta designed to help developers create multimodal AI agents. This framework is specifically tailored for use with XR devices and augmented reality glasses.

Read the source →
Amazon Quick Adds Autonomous Agents
AWS Machine LearningProduct

Amazon Quick Adds Autonomous Agents

Amazon Quick has introduced autonomous agents that perform tasks independently to save users time. The update also includes an activity feed for prioritization and a tool to extract insights across multiple data sources via a single query.

Read the source →
Enterprises Struggle with AI Costs
TechCrunchOther

Enterprises Struggle with AI Costs

Many companies are finding it difficult to determine the actual return on investment for their AI spending. After initial periods of unrestricted usage, some firms are now scaling back licenses or facing depleted budgets.

Read the source →
AI Aids in Identifying Rare Pediatric Genetic Disorders
OpenAIResearch

AI Aids in Identifying Rare Pediatric Genetic Disorders

Scientists employed an OpenAI reasoning model to assist doctors with difficult medical cases. This approach successfully identified 18 new diagnoses for children with previously unexplained conditions.

Read the source →
Challenge to LoRA Dominance
Hugging FaceResearch

Challenge to LoRA Dominance

Hugging Face is exploring whether new methods can outperform LoRA, the current leading technique for fine-tuning AI models. The initiative aims to find more efficient ways to adapt large models.

Read the source →
NVIDIA Supports AI Evolution in Advertising
NVIDIAProduct

NVIDIA Supports AI Evolution in Advertising

NVIDIA is collaborating with partners at Cannes Lions to integrate autonomous AI operations into marketing. The focus is on providing the infrastructure necessary to handle the scale and speed of modern AI tools.

Read the source →
Amazon Launches Bedrock AgentCore Harness
AWS Machine LearningProduct

Amazon Launches Bedrock AgentCore Harness

Amazon has released a tool that allows developers to deploy production-ready AI agents using just two API calls. These agents operate in isolated environments with the ability to execute code, access the web, and maintain memory across different sessions.

Read the source →
New US Regulations Accelerate Data Center Grid Access
TechCrunchPolicy

New US Regulations Accelerate Data Center Grid Access

FERC has instructed grid operators to expedite the connection process for AI data centers. However, the mandate does not resolve existing deficits in electricity availability.

Read the source →
Debating AI Risk Governance
The VergePolicy

Debating AI Risk Governance

Recent discussions have centered on who holds the authority to determine when artificial intelligence becomes too hazardous. These conversations follow a complex situation involving Anthropic and the Trump administration regarding the Fable 5 model.

Read the source →
AI Matches Physician Performance in Medical Trials
The DecoderResearch

AI Matches Physician Performance in Medical Trials

New research in Nature indicates that specialized AI can diagnose illnesses and suggest treatments as effectively as human doctors. However, the studies noted that the underlying models used in these systems are already becoming obsolete.

Read the source →
Evaluating Agentic Capabilities of Open Models
Hugging FaceResearch

Evaluating Agentic Capabilities of Open Models

Hugging Face has introduced a method to benchmark open-source models using custom tools. This approach tests how effectively these models can act as autonomous agents within specific environments.

Read the source →
Karamo Brown Debuts AI Wellness App
TechCrunchProduct

Karamo Brown Debuts AI Wellness App

Life coach Karamo Brown has released a new health and wellness platform called Kē. The application features an AI-powered digital version of Brown to guide users through personal growth and nutrition.

Read the source →
Adobe Adds AI Chatbots to Creative Cloud
The VergeProduct

Adobe Adds AI Chatbots to Creative Cloud

Adobe is introducing dedicated AI assistants across its primary design and editing software. These tools are currently available in a public beta for applications including Photoshop and Premiere.

Read the source →
VR-Controlled Humanoids in Shenzhen
WiredProduct

VR-Controlled Humanoids in Shenzhen

IO-AI Tech is utilizing virtual reality equipment to operate humanoid robots. This system allows workers in China's hardware hub to control machines using their own physical movements.

Read the source →
LeCun Predicts Market Crash for Major AI Labs
The DecoderOther

LeCun Predicts Market Crash for Major AI Labs

Yann LeCun suggests that companies like OpenAI and Anthropic may collapse due to unsustainable investor subsidies and high operational costs. The researcher recently secured $1 billion for his own venture, AMI Labs, which utilizes a different AI methodology.

Read the source →