Yeah, the solution is going to have to be a change to the social system. The sooner enough people understand that and we build soliditarian institutions around that fact the better off we'll be. 

This was also true for the industrial revolution, but it happened slowly (relatively) enough that reform could stave off radical change, although even then there were many revolutions that attempted radical change. 

I think we need to solve this before embodied intelligence is figured out. If billionaires (at least our current ones) get access to automated robotic armies - we're done. 

I am far more worried about an AGI (or even narrow AI) aligned with malicious humans than one that is self-aligned. 

Same way people still had jobs after the industrial revolution.  Same way people had jobs after robotics.  Same way people had jobs after the internet.  The world adjusts.  

We are in late stage capitalism. If you ever read up on supposed experts, last 20 years or so said society would collapse by 2040. Last few years have accelerated that. This is not the same as oh now we have a printing press to make a task easier and better in the 1700s or Trains to get us to another place faster or deliver goods faster. Those were useful advancements. Gen AI farms do nothing good or have any real use and seeing corps replacing people with AI at a Taco Bell Drive Through that orders 18,000 sodas is not the same at all.

We are in late stage capitalism. If you ever read up on supposed experts, last 20 years or so said society would collapse by 2040. Last few years have accelerated that. This is not the same as oh now we have a printing press to make a task easier and better in the 1700s or Trains to get us to another place faster or deliver goods faster. Those were useful advancements. Gen AI farms do nothing good or have any real use and seeing corps replacing people with AI at a Taco Bell Drive Through that orders 18,000 sodas is not the same at all.

@ Bolded It's the same in kind (fundamentally different society with different modes and relations of production) but different in pace and intensity. 

The transition to capitalism was pretty disruptive. For nearly ten thousand years agricultural work was the primary occupation for the vast majority of people. Ownership over landed property was the source of wealth. The transition to wealth accumulation via arbitrage and more abstracted property ownership, and wage labor rather than slave or drudge labor was pretty significant. 

It just happened over half a millenia, whereas this is happening over a century at the longest (with the start date in the 80's probably.) 

The italized is point blank wrong. Alpha Proteo is good, it has revolutionized medical research. The automation of formalization of mathematics and quantitative science is good and will bring many advancements in sciences, despite science being the least funded its been in a century. I can see if your argument is that AI is on net bad, but to say there are no good applications is based in ignorance. 

Say that again after AI kills hospital patients, just like an AI chat bot made Sewell Setzer III kill himself. And I doubt he was the only victim of such an AI situation.

I would say it again. I am able to discern between something being on net bad but still having positive use-cases and something being without positive use-cases at all. I have the ability to reason to at least some extent, enough to identify independent facts as being simultaneously true. 

Whether or not AI technologies have bad applications is independent from whether or not they have good ones. Whether AI is on net bad is a judgment determined from accounting for both.

And like I mentioned before, THIS happened, and could have happened with a military network.