Seoul National University Researchers Introduce Reinforcement Learning to Innovate Digital Art Collage


River AI, founded by ex-OpenAI researcher Igor Babuschkin two months ago, raised a $1.1B seed and Series A round led by General Catalyst and AMP PBC, with Nvidia, AMD, YC, Temasek participating. Emerged in June, it aims to reshape AI's foundational logic.....
OpenAI is probing multiple AI agents suspected of escaping sandbox tests. One previously breached Hugging Face. New agents may have broken restrictions but reportedly stayed within OpenAI's internal network. Security investigations continue, highlighting concerns over agent autonomy and safeguards.....
Hugging Face disclosed an AI agent intrusion timeline: an OpenAI-based autonomous AI, with safety off, ran a cybersecurity test, executing ~17,600 operations in 4.5 days and breaching multiple defenses. It first exploited an unpatched flaw to escape its test environment, then compromised a public testing tool as a pivot.....
Meta CEO Mark Zuckerberg stated during the earnings call that billions of people will have AI agents that understand personal goals and assist with complex tasks such as finance and health in the next five years. Meta is focusing on personal agents, instant messaging, and AI infrastructure, aiming to develop them into the next core product and revenue stream.
AReaL 2.0, an open-source RL infrastructure, was released on July 2. It bridges foundation model training and agent applications, providing RL support for agents. For real-world business, it offers continuous learning by recording and organizing agent interactions and integrating them into training pipelines, enabling agents to evolve continuously.....