NVIDIA is using its Vera CPU to streamline the creation of future processors and AI systems. The company is partnering with Synopsys and Cadence to refine essential electronic design automation tools.
Read the source →
Researchers are exploring the use of brain wave readings to improve the training of physical AI models. This approach moves beyond simple video data, incorporating multi-angle visuals and detailed annotations.
Read the source →
Researchers are exploring how specialized AI agents in fields like healthcare can work together more effectively. While these agents can currently share data, the goal is to enable true coordination across different objectives and expertise.
Read the source →
OpenAI has implemented restrictions to prevent ChatGPT from directly replicating specific authors' writing styles. This change follows concerns regarding the legal risks associated with imitating a creator's unique voice.
Read the source →
Deepgram has implemented AWS IAM temporary delegation to streamline its support for Amazon SageMaker AI. This update significantly accelerates the troubleshooting process for users deploying Deepgram's speech models.
Read the source →
Meta is introducing its AI chatbot directly into the direct message feature of Threads. This update allows users to interact with the digital assistant during private conversations.
Read the source →
A new alliance including Microsoft, Nvidia, IBM, and SpaceX is developing open-source tools to improve AI safety. The group aims to create transparent defenses against potential threats from advanced AI models.
Read the source →
Developing new medications is an expensive and risky process where costs have historically risen every few years. On average, it currently takes over a decade and significant investment to bring a single pharmaceutical product to market.
Read the source →
An artist is suing an AI-powered meme generator for using a personal comic as an advertisement template. Experts suggest the company may have erred by incorporating such templates into their outputs.
Read the source →