Is Google's Gemini... legit?

**Google's Gemini** AI model is generating significant discussion and skepticism, especially regarding its **32-shot chain of thought** MMLU claim and **32k context window**. The community is comparing Gemini's performance and capabilities with **OpenAI's GPT-4** and **GPT-3.5**, highlighting the upcoming **Gemini Pro** and **Gemini Ultra** models on the Bard platform. Users report various **OpenAI service issues** including chatbot errors and subscription problems. Discussions also cover **prompt engineering techniques**, AI model evaluation comparing **GPT-4**, **Claude 2.1**, and **PaLM2**, and improvements in speech and multimodal capabilities. The bot now supports reading and summarizing links from platforms like arXiv, Twitter, and YouTube, enhancing user interaction.