Not at all. The model weights and training data remain the same, it's just RAG'ing real-time twitter data into its context window when returning results. It's like a worse version of Perplexity.