AI News — Thursday, November 12, 2026
Nvidia's stock has fallen sharply amid reports that its board is preparing to remove CEO Jensen Huang following a major scandal.
A detailed timeline outlines the events surrounding the alleged affair between Nvidia CEO Jensen Huang and a suspected Chinese spy, leading to calls for his ouster.
Despite widespread claims of AI boosting productivity, a QA team's internal discussions reveal a more complex reality, highlighting challenges and unexpected issues.
Google introduces the Gemini Decisions API, designed to provide fast, calibrated multimodal decision-making capabilities for developers.
OpenAI launches 'Businesses in ChatGPT,' a new offering tailored to help companies integrate and leverage ChatGPT for various business operations.
Researchers present a novel method for generating streaming video in just a few steps by effectively shifting conditions, promising more efficient video synthesis.
As speculation mounts about Jensen Huang's potential departure, TechCrunch offers a speculative list of individuals who might step in as Nvidia's next CEO.
Google Labs unveils Tabletop, a new platform offering customizable and collaborative virtual spaces for groups to plan, play, and work together using AI.
JioStar's chief architect shares insights into the challenges and successes of implementing AI discovery systems for a massive user base of 500 million.
A new research paper introduces Tree-Ising, a method that uses bit-level Ising edits to effectively repair and improve the performance of large language models.
New research explores how agentic robots can achieve advanced manipulation tasks through 'world in-context learning,' improving their adaptability and skill acquisition.
A study proposes using evidence-linked memory to mitigate performance degradation in LLM agents, ensuring more reliable and accurate responses by grounding information.
OpenAI highlights initiatives and tools designed to empower caregivers by integrating AI solutions into their daily routines for support and efficiency.
This article provocatively argues that while AI-generated code often faces criticism for being unmaintainable, human-written code frequently suffers from similar issues.
Researchers introduce dNaViT, demonstrating that discrete codes can serve as effective general-purpose visual representations, simplifying and unifying visual processing.