This is brilliant! I’ve had a feeling something like this had to be around the corner for a while now. That is, the idea that we can actually effectively resist our shared “algorithmic oppression” in our comprehension of how it is empowered in the first place. It’s not like we aren’t capable of making the algorithms “dumb” as well. We can make the gradients melt away just as much as we can make them emerge in the first place.