[Kednicma]

We aren't completely starved yet of the bottom-up approach, but I agree that it's somewhat limited. We can explain choice and free will in bottom-up terms, which complements the article's explanations of memory and signaling.

In the classic video game "Link's Awakening" (chosen to fit the article's theme), there is a maze of signs. Each sign points towards another sign. Reading the signs in the order that they point to each other, following the chain of arrows, leads to a prize. The player is local and does not know where the prize is, but the signs encode global information about the maze. The player's memory is limited and can only remember one sign at a time, but that is sufficient. It seems to me that cells communicate and act using similar local/global distinctions.

> We reject a simplistic essentialism where humans have ‘real’ goals, and everything else has only metaphorical ‘as if’ goals.

This is the philosophical meat of the article, and the tough takeaway for the reader. The reader must admit that their own goals, since they are human, are not quite "real" in a way which somehow transcends the accidental success of blindly-evolving low-level components. Rather, humans do the same sort of predictive modeling, blind guessing, and lucky incidence that we see in "simpler" life forms.