Grok's History: Musk's Eye-Opening Admission

This episode of Machine Learning Street Talk dives into several pivotal developments in the AI landscape, starting with Google's Gemini AI rollout into 4 million GM vehicles, including Cadillac, Chevrolet, Buick, GMC, and now Volvo, replacing the Google Assistant with over-the-air updates. The integration leverages Android Auto and promises seamless, powerful voice-controlled functionality, including potential work integration in self-driving scenarios. The discussion then shifts to Anthropic’s new Cloud Security public beta, a tool that scans codebases for vulnerabilities using its Mythos model, offering automated fixes and integration with platforms like GitHub and Cloud Code. Despite being excluded from the Pentagon’s classified network—due to a supply chain risk designation—Anthropic is pushing hard in enterprise security, with major firms like CrowdStrike and Accenture adopting its tools. The episode also covers Elon Musk’s recent testimony in the Sam Altman trial, where he admitted XAI used OpenAI’s models to train Grok via distillation—a common industry practice but against OpenAI’s terms. This exclusion from the Pentagon’s elite AI list, alongside the inclusion of lesser-known but elite-founded company Reflection AI, signals a strategic move to diversify AI partnerships. Finally, the episode highlights Anthropic’s aggressive $900 billion valuation round, with investors given just 48 hours to commit, suggesting a major pre-IPO funding push. The host promotes his own AI platform, AIbox.ai, as a hub for over 80 AI models and automation tools.