OpenAI released ChatGPT, an optimized language model for dialogue, for testing purposes. The model is trained using Reinforcement Learning from Human Feedback (RLHF).

First tests demonstrate an amazing performance.

Example: