The reward function is already how well you manipulate humans
We have many, many parables about dangers arising from super intelligent or supremely capable AI that wreaks ruin by maximizing a simple reward function. "Make paperclips" seems to be a popular one[1]. I worry that we don't give a more realistic scenario deep consideration: the world that arises from super...
I had a similar experience with midjourney. The question now is, how do you change your life once you have the more visceral understanding of the near term future? Seriously, this is my biggest problem. I deeply believe change is coming, fast, but I'm still stuck in so many patterns that only make sense if the status quo continues.