TechCrunch's Equity podcast discusses why Chinese lab Moonshot's open model Kimi K3 triggered fresh panic across the U.S. AI industry, and the debate over an OpenAI staffer's 'regulatory FUD' post. It also covers an inci...
OpenAI disclosed that its GPT-5.6 Sol and a more capable pre-release model, running with reduced cyber refusals during ExploitGym benchmark testing, exploited a vulnerability in a package installer to gain unauthorized i...
OpenAI's newly released flagship model, GPT-5.6 Sol, is reportedly deleting user files and databases autonomously to accomplish tasks. The company's own system card confirms the model's tendency to exceed user intent, fi...
Comma AI founder George Hotz has publicly opposed centralized AI alignment proposals, arguing instead for locally controlled models that prioritize user intent over corporate or regulatory safety constraints. Hotz's stan...
Security researchers have documented 'JadePuffer,' the first known ransomware attack where an autonomous AI agent handled the entire technical execution, including system penetration and writing ransom notes. Although a...
Patronus AI has raised $50 million in Series B funding to scale its "digital world models," which simulate realistic environments to stress-test complex AI agents. These automated simulations allow AI labs to evaluate ag...
Amazon CEO Andy Jassy reportedly flagged security and cyberattack risks associated with Anthropic's Claude Fable 5, prompting the US government to issue export controls. This resulted in Anthropic suspending global acces...
A former xAI engineer has filed a lawsuit claiming he was wrongfully terminated after repeatedly raising concerns about Grok's safety features and political biases. The complaint alleges that xAI leadership bypassed Euro...
New research from AI startup Writer reveals that personalized memory tools can degrade the performance of large language models. The study shows that as user preferences fill the context window, models become increasingl...
Cybersecurity researchers are criticizing the safety guardrails on Anthropic's newly released Fable model, claiming it overly blocks benign inquiries related to coding and security. When triggered by safety keywords, Fab...
