Part 9/15:
Musk critiques the state of current AI technology, highlighting the massive computational resources required and how the models today fall far short of human capabilities. For example, he notes that large language models (like GPT) still cannot outperform humans in writing a novel despite enormous compute power.
He points out that current AI models are primarily imitative, heavily reliant on "training data"—but they lack genuine understanding or reasoning ability. Musk argues that brute-force scaling—adding more data and compute—may not be enough; instead, a better understanding of the underlying science of neural networks is crucial.