Animist Superintelligence

January 26, 2021

a lightbulb diagram

Maybe our future, post-AI society will have more in common with the deep past than our present.

Where A is a David Abram-style animist perception (nature perceives us, it is a two-way street), B is our current epoch, and C is a hypothetical future dominated by artificial intelligence.

According to Abram, our current attempts at recreating an animist world (using tech that can talk to us and seem to perceive us) are doomed by solipsism. A smart refrigerator can't say anything new.

But when AI gains the ability to consistently surprise us in intelligent ways, will our tech regain the 'radical Otherness' of nature? It seems awfully optimistic, but I find the idea exciting.

Maybe machine learning will 'think' in a more natural way than we do. It's easy to imagine AI as a super-powered extension of our own thinking styles, but the mechanics of machine-learning are more like nature selection than human abstraction-- they take a ton of time and experience, brute-forcing their way to understanding.

What might come with those encounters with superintelligence, then, is a sense that the non-human world is clocking us, observing our nakedness-- a connection with a very old feeling.

The chess games of AlphaZero provide this kind of mystery.



Few people today, when they’re cycling past a stand of oaks, sense that those trees are sensing them; we don’t feel the breeze gusting around us as a sensitive and sentient presence, and upon arriving...
Maybe we should aim for a goal that is the opposite of that of Enlightenment philosophy: to fragment the world according to difference instead of universalizing through a presumed absolute. A new worl...