LLMs already outperform humans in a huge variety of tasks. ML in general outperform humans in a large variety of tasks. Are all of them AGI? Doubtful.