Comment by Nevermark

> Everyone needs to figure out how AI works for them.

That is the real takeaway.

I use LLM’s for dozens of things. It took time to find how model strengths best connect with mine.

People who only evaluate models for one purpose over short timespans are going to have a hard time.

It’s like models come from a different cognitive & communication culture. Eliminating that barrier takes more experimentation and experience than logic.

I think their creativity is also often a function of the prompters creativity.

The harder you (creatively) push them into a creative corner, the more genuinely creative they get.

Not unlike how a good teacher can get unexpected positive results from students when they are channeled and pushed.