The OAI-HF incident, the Iranian critical infrastructure hacks, etc. (combined, of course, with our inherent vulnerability to pathogens and stuff like that) should be enough to widen the aperture of imagination on the very significant downside risks that are approaching. But this idea @KelseyTuoc has of "a world where humans are without economic value" strikes me as a similar failure of imagination. Even if a model can do large portions of the types of jobs Kelsey and I have better than we can, I reckon the extraordinary surplus that will correspond with those advances will mean a surge in demand for economic exchanges that only humans can fulfill. People (myself included) will continue to want human beings to do certain things for them based on the very fact that they're human beings. I could come up with a hundred examples, but here are two that appeal to me personally: 1) Human tutors for my kids. I don't want to turn my children over to machines, nor do most parents. AI will surely play a big part in developing lessons, but even if rapid technological progress sets in I expect to be very willing to pay a lot of money for a caring human guide. (And when we're surpassed as writers, it'll be people like Kelsey and me that are available to fill these roles.) 2) Live entertainment. We like watching people do things well, be it on stage or on the field. We will continue doing so even when humanoids can play a concerto really well and shoot three-pointers. As @ModeledBehavior puts it, don't underestimate the economics of the human touch.