A Question of Manmade "Intelligence"

             I'm sure you remember that on Mew-onday, just a couple of days back, we had an opportunity to listen to Emad M. discourse on the end of capitalism and how AI is going to change every thing in 1000 days!
           It was a very long interview. I listened to all of it. 
          Of course, I am not qualified to judge the value of the information presented. So, I decide to forward it to an expert, my son, who has been doing all sorts of such work as internet security and like that since he started doing IT for a cardiologist's office when he was 18. I think it might be more accurate to say that he is an internet engineer? 
        Anyhow, below I have included his review of the video. I have also included the video itself, so that you can compare the two. 

_______________________________________________

Okay, so i watched most of that video. They ramble a bit, so i don't remember it all.

AI replacing humans? There are still some big gaps to cover. We context switch constantly and i don't see any indication of them doing that well. We constantly learn, and as i understand it, in the big wonderful complex models learning and running are distinct phases and once the model is 'taught' it doesn't learn more from its day to day operations. 

I expect that some highly specialized tasks could be done by machines, but at least for a while they will still need human overseers for when they screw up. I can envision certain classes of work being done poorly, but at ever more frenetic pace, by machines.

 Of course a lot of their economics talk is well over my head. but i do have a fairly strong contempt for economics at all. I don´t think economics as a 'science' describes some objective reality. I think their 'reality' is a shared human-created consensus and essentially arbitrary. Of course a lot of people make their gelt by reinforcing the consensus, so it probably isn't going anywhere. I don't have the wisdom or charisma to immanentize my particular eschaton. the guest's utopia seemed to be based on paying people for being human -- right after he's been talking about how AI's will be able to do everything we can do. So by that argument, the AI's should be able to fraudulently claim to be human and short circuit his particular secular prosperity gospel.

Towards the end when the host is groping for some actionable advice from his guest, the Brit says people should learn how to use AI. No matter what I think about their convo, that seems to be where i am. I'm trying to get stablediffusion running on one of my computers for practice. Then maybe try jan.ai next. Not so i can talk to them, but so can learn the mechanics of running their software.

_______________________________________________

           I would be very interested to know what any of you think of the video, or the review.
           My thanks, also, to LoneStar for posting it. It's really something to think about.  
           Could our whole economy change like that? What say you?     

❓❓❓