To understand AI’s true capabilities, experimenting with the latest tools is essential. Hands-on testing reveals strengths and weaknesses more clearly.Take large language models (LLMs) as an example. They exhibit emergence—unexpected behaviors that even developers struggle to explain. Some abilities surpass current theories, leaving researchers puzzled.If developers find it hard to fully grasp LLMs, it’s even harder for others to assess them accurately. That’s why AI experiments are crucial—they provide a glimpse into the latest LLM capabilities.