> I use the acronym “AI” to refer to trends in generative AI and large language models.

Please don't propagate the "LLMs are artificial intelligence" PR bullsh*t.

Nah. AI is anything that can often solve inexactly specified problems, often through the use of a combination of probabilities and deductions.

Your problem is thinking that "artificial intelligence" has to be whatever you think "intelligence" is. This is as true as "artificial grass" needing to be a type of grass. It's a constructed imitation that solves some of the problems that we normally use grass for. It can be green and spiky, soft to play sports on and pleasant in appearance, but it still is in no way grass, and has no relationship to grass. It is something that fits its category as a tool in some circumstances. An artificial leg is not a leg, it is a piece of metal or plastic with a hinge in it. An artificial heart is not a heart, it is a pump. They are both limited replacements for another thing in some useful circumstances.

Artificial grass, legs, or hearts don't become actual grass, legs or hearts even if they perform better in all use cases that we have for the actual item. They only become "grass," "legs," and "hearts" once we find the distinction to be largely useless and begin to omit it.