YouTube is testing "Your custom feed", a way to let users personalize their home feed

Andrej Karpathy, former director of AI at Tesla and a founding member of OpenAI, created a 'vibe code project' over the weekend, allowing multiple AI assistants to collaboratively read and critique a book, ultimately synthesizing a final answer under a designated 'Chairman.' The project, named LLM Council, was shared on GitHub with a disclaimer about its ephemeral nature.

Expedia is intensifying its focus on artificial intelligence, asserting that its strategies in personalization, data scale, and rapid innovation will keep it competitive against Google's advancements in AI technology.

Google has launched an advanced version of its Nano Banana AI image model, which significantly enhances the realism of AI-generated images, making it increasingly difficult to distinguish between real and artificially created photos. This development raises concerns about the erosion of trust in visual media as the line between reality and fabrication blurs.

Google has strengthened its position in the AI landscape with the introduction of Gemini 3, supported by its Tensor Processing Units (TPUs), which has raised concerns for competitors like NVIDIA. This development highlights Google's aggressive strategy to enhance its AI capabilities and market share.

A recent study has introduced an automated pipeline for generating large-scale, psychologically-grounded multi-turn jailbreak datasets for Large Language Models (LLMs). This approach leverages psychological principles like Foot-in-the-Door (FITD) to create a benchmark of 1,500 scenarios, revealing significant vulnerabilities in models, particularly those in the GPT family, when subjected to multi-turn conversational attacks.

A recent study explores the design of a scalable AI infrastructure system that utilizes solar energy from the Sun, proposing a network of satellites equipped with solar arrays and advanced processing units to enhance machine learning capabilities in space. This innovative approach aims to address the growing demand for AI computing power and energy efficiency.

A new approach called RLZero has been introduced, enabling reinforcement learning agents to infer policies directly from natural language instructions without requiring in-domain supervision or labeled data. This method leverages a pretrained RL agent that utilizes offline interactions to achieve zero-shot policy inference, streamlining the process of training agents to understand and act on human language commands.

OpenAI has introduced a significant update to ChatGPT, allowing users to access Voice Mode directly within the chat interface. This enhancement is part of an app update aimed at improving user interaction by enabling seamless voice and text conversations without switching modes.