Some sort of "generate descriptions of novel tasks including ways to evaluate performance at those tasks, evaluate quality of the generated tasks+evaluation-metrics, split tasks into subtasks, estimate difficulty of tasks in a way that is is judged on how it compares to a combined estimated difficulty of generated subtasks and to actual success rate and quality" sort of deal?
Physics.
I'm skeptical of the idea that anything is going to derive intelligence from the bottom up, but I'll be super impressed if that's how it goes.
Why not? We started off as single celled organisms and look at where we are now.