The release of Google's Gemini model has caused a lot of excitement, followed by a lot of debate, some controversy, mistrust, and yet another way to look at big tech and the future of AI. What exactly happened? Is Gemini better than the current version of ChatGPT? If so, in what way? If not, in what way? Did Google lie about their new product? _x000D_ _x000D_ The promotional news said Gemini is ahead in 30 out of 32 tasks. What does that mean? What are those 'tasks'? When engineers are comparing two AI models, what are they comparing? How do engineers evaluate a model's intelligence level?_x000D_ _x000D_ In this club, we'll take a look at Gemini, compare it with GPT, and clarify the fancy technical jargon related to the current large language models. We'll attempt to break through the controversy by understanding the products, and comparing them objectively, and we'll speculate about the future._x000D_ _x000D_ 谷歌双子座模型的发布引起了很多兴奋，随之而来的是很多争论，一些争议，不信任，这是看待大型科技和人工智能未来的另一种方式。到底发生了什么?双子座比当前版本的ChatGPT更好吗?如果是，以什么方式?如果不是，以什么方式?谷歌对他们的新产品撒谎了吗?_x000D_ _x000D_ 宣传新闻说，双子座在32项任务中有30项领先。这是什么意思?这些“任务”是什么?当工程师比较两个人工智能模型时，他们在比较什么?工程师如何评估模型的智能水平?_x000D_ _x000D_ 在本俱乐部中，我们将研究一下谷歌的双子座模型Gemini，将其与GPT进行比较，并澄清与当前大型语言模型相关的花哨技术术语。我们将尝试通过了解产品来突破争议，并客观地比较它们，我们将推测未来。