[kirth_gersen]

The jagged edge of LLMs' capabilities makes it really hard to explain what it's good at, and therefore what it's good for. It can be genius level at one thing and then extremely foolish at a seemingly related task in ways that are hard to intuit. It is still early days and AI is improving at a startling pace. So, eventually even the floor of competency will rise above the threshold of "average human". That's my guess anyways. Asking what AI is good for is kind of like asking what a toddler is good for. The question itself is premature.