Amazon tops AI ranking | Generative ai Google | Free Generative ai Tools for Images | Traditional ai Examples | Turtles AI

By now, we have all become familiar with Chat GPT, the most renowned of the textual input and output AIs, yet its great success may be close to failure. To talk about this, let us first explain how an AI of this category is evaluated. The most widely used and famous of the methods is to administer multiple-choice questions of various kinds (scientific, historical and question-based or image-based) to the artificial intelligence and evaluate the percentage of correct answers given by the machine. It is also possible to have a few lines of text added to make the reasoning made explicit and thus understand whether it arrived at the answer by chance or merit. Using exactly this method, ScienceQA (which I invite you to visit for more information or to try AI quizzes) administered 21,208 questions to Chat GPT and answered 75.17% correctly. This figure is not incredible but still respectable, were it not for the fact that, in a second test with the same questions, an AI in development by Amazon managed to answer no less than 91.15% of the questions correctly. This was possible thanks to an MMCOT machine approach (we will soon dedicate an article to it) that probably represents the near future of AI. One would now have to wonder what Amazon could do with such a powerful artificial intelligence.