Fine-tuning is underway and reinforcement learning begins in a few days. 2 to 3 weeks to public release.
This will be a major improvement over the 0.5T v8-small that currently serves all Grok production traffic, especially for difficult coding tasks.
💡 Inside Track & Deep Insight
Elon Musk’s update on Grok foundation model V9-Medium marks a strategic shift from the 0.5T v8-small, more than tripling parameter count to 1.5T. The inclusion of Cursor data—a code editor with AI autocomplete—is a game-changer for programming tasks. This suggests Grok is not just a chatbot but a specialized coding tool, likely targeting developers and enterprises. The reinforcement learning phase, starting in days, will further refine its reasoning, making it competitive with OpenAI’s GPT-4 Turbo or Google’s Gemini. For Tesla, this could accelerate Full Self-Driving (FSD) development by improving code generation and simulation. On the crypto front, Musk’s history with Dogecoin implies Grok might integrate with X payments, potentially boosting DOGE narratives. The 2-3 week release window hints at a rapid deployment, possibly via X Premium, expanding Musk’s AI influence. Stock-wise, while Tesla and SpaceX aren’t directly tied, any AI breakthrough indirectly lifts Tesla’s autonomy story. Crypto markets may see speculative moves on Dogecoin, given Musk’s past tweets. Overall, this is a major technical milestone that could democratize advanced AI coding tools.