One of my favorite articles. I see people using this as an argument for why LLMs are fundamentally flawed, but that fails to realize that most humans don't get into the details. In fact, many don't have the vocabulary to be in the details.

There is an argument to be made for using humans, but I think we need humans to be more capable, more curious, more adaptive. LLMs are far better than the average human, but they are fundamentally inferior to the motivated human working with an LLM to augment themselves at the frontier.