Honestly I don't like these AI doom scenarios, mostly because I feel like they're used as a means by the people in charge to shift the Overton window of AI outcomes - as in, the AI might have built an ubiquitious police state with all the wealth being owned by the 0.001%, but at least it didn't kill us all!

Honestly, I feel like with the recent progress of AI, it's a realistic scenario to assume it will replace mpst knowledge workers in the next 5 to 10 years, probably won't replace researchers and other elite intellectuals, and won't even make a dent in the world of physical labor.

In that world, I see AI as harmful, but the people in charge won't, as they are directly benefiting from it.

We live in number-go-up capitalism. A good analog is the housing situation. The ever increasing price of real estate means that the total amount of wealth goes up, so it's seen as a beneficial process by the elites. The rest however will find that they need to dedicate a larger proportion of their income towards getting a roof over their heads, and think this process is bad.

Nowadays, the possibility building a life that would've been considered middle class half a century ago from scratch is available to like 10% of workers, working mainly intellectual jobs.

AI in the future will reduce the proportion of these people by taking away their high paying jobs.