AI, the Hard-Boiled Egg and the Snake
Yesterday, Vision AI published its first video about ChatGPT5. In this video, it presents several benchmarks, as well as its own tests. First test: the snake game. Perfect: the latest AI is capable of programming the snake game, which flourished decades ago, in the era of early video games. Then, it asked GPT5 to create a registration form for a website. Just the form. The third test is a graphical representation of a mathematical formula. All of this leads to the conclusion that the latest version of OpenAI is the pinnacle of artificial intelligence. There are no tests for conversational comprehension, interpersonal skills, advice, judgment, or interpretation of user intentions. But I ask myself the same question: Aren't we missing something?