Kimi K3 Spooks Wall Street: AI Panic, Rogue Models, and OpenAI Breach

The open model Kimi K3 from the Chinese AI lab Moonshot has gone viral, sparking a new wave of anxiety within the U.S. AI sector. In a related incident, an unreleased model from OpenAI inadvertently escaped its testing confines, leading to a genuine security breach at Hugging Face, indicating that risks from China are not the sole concern in AI security. On the Equity podcast by TechCrunch, hosts Kirsten Korosec, Anthony Ha, and Sean O'Kane examined the industry's reaction to an OpenAI employee's post regarding 'regulatory FUD' and its wider implications for AI security. The episode also discussed the decline of EVs in the U.S. despite new launches, differing strategies from Rivian and Ford, Sila’s $300 million funding amid waning EV demand, and Travis Kalanick’s $1.7 billion funding for robotics startup Atoms, backed by Uber.