Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/how-long-until-ai-dominates-everything.17575/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

How Long Until AI Dominates Everything ?

A

Arthur Hanson

Well-known member
With multiple AIs out there and more coming on fast, all the while becoming ever more powerful at a blistering rate, how long will it be that AI becomes the dominant force and disruptor in our society? It looks like AI has reached critical mass and is about to go nuclear in a social, business, government and conflict sense. This is being propelled by ever more powerful processors, memory, communications and sensors. Will pitting AIs against each other speed up this process even more? This will be the most disruptive force in human history dominating automation, business, government and conflict of all types including war. All thoughts and observations sought and welcome. Is there any chance AI will also achieve a sentience of some type?
 
Computers aren't smart, they're high speed idiots.

AI has been around for decades and decades. Processors are getting faster at a slower pace, anyone who hopped on a 286/486 or Pentium had a different experience than getting a new generation today.

So, what suddenly changed? Sure, more parallelism, but that's been around for decades too. Now, it should keep getting better, especially since it's the latest craze and getting so much attention. And we do have more processing power than before. But, it's just difficult for me to believe that there's going to be this inflection point where it's suddenly going to go from sucking (which it does now), to being able to achieve something approximating real intelligence on a scale far beyond a human.

Give me a computer and I'll kill it in any game if we both have equal units. They are better now, but even with that, they aren't even competitive in anything requiring reasoning, fuzzy logic, intuition, etc.... If it's easily quantifiable, yeah. Anything requiring real intelligence, a computer isn't competitive in strategy games. Maybe they will be, but they are so far behind, I don't see if happening any time soon.

I'm just not sold on it. A quick Google search where I rarely get what I want, indicates they can't seem to train their search engine to work better. Whereas, if I asked a human the same thing, they would easily understand me. The human brain is imprecise, but it works on so many levels, I don't see AI replacing it any time soon. Of course, computers are better, and have been for quite some time, at many things. I'm not sure that's expanded dramatically lately, nor do I see why it would have based on any technological barriers being removed.
 
ta152h said:
Computers aren't smart, they're high speed idiots.

AI has been around for decades and decades. Processors are getting faster at a slower pace, anyone who hopped on a 286/486 or Pentium had a different experience than getting a new generation today.

So, what suddenly changed? Sure, more parallelism, but that's been around for decades too. Now, it should keep getting better, especially since it's the latest craze and getting so much attention. And we do have more processing power than before. But, it's just difficult for me to believe that there's going to be this inflection point where it's suddenly going to go from sucking (which it does now), to being able to achieve something approximating real intelligence on a scale far beyond a human.

Give me a computer and I'll kill it in any game if we both have equal units. They are better now, but even with that, they aren't even competitive in anything requiring reasoning, fuzzy logic, intuition, etc.... If it's easily quantifiable, yeah. Anything requiring real intelligence, a computer isn't competitive in strategy games. Maybe they will be, but they are so far behind, I don't see if happening any time soon.

I'm just not sold on it. A quick Google search where I rarely get what I want, indicates they can't seem to train their search engine to work better. Whereas, if I asked a human the same thing, they would easily understand me. The human brain is imprecise, but it works on so many levels, I don't see AI replacing it any time soon. Of course, computers are better, and have been for quite some time, at many things. I'm not sure that's expanded dramatically lately, nor do I see why it would have based on any technological barriers being removed.
Click to expand...
This may be true now, but at the rate AI/ML is improving, it won't be long before it outruns all but the very most intelligent people in many areas. This presents a real threat to many peoples occupations.
 
AI has been around since I was an undergrad, there was a language called LISP we used for AI like programs. We used mostly Fortran back then and some C. Today's AI is far more advanced so there is really no comparison.

AI is showing up big time in the semiconductor ecosystem and on wafers. AI eats up CPU speed and memory so I would say it is a top driver of new semiconductor processes.

As far as dominating everything, I would say that is probable in our lifetimes but it depends on the application. Automotive and transportation, definitely. Medical for sure but it will take longer. It is already a part of chip design and that will increase. The chatbot stuff is still a toy but I can see it maturing into something disruptive. But again it is going to take a LOT of silicon so we are in the right place, absolutely.
 
Daniel Nenni said:
AI has been around since I was an undergrad, there was a language called LISP we used for AI like programs. We used mostly Fortran back then and some C. Today's AI is far more advanced so there is really no comparison.

AI is showing up big time in the semiconductor ecosystem and on wafers. AI eats up CPU speed and memory so I would say it is a top driver of new semiconductor processes.

As far as dominating everything, I would say that is probable in our lifetimes but it depends on the application. Automotive and transportation, definitely. Medical for sure but it will take longer. It is already a part of chip design and that will increase. The chatbot stuff is still a toy but I can see it maturing into something disruptive. But again it is going to take a LOT of silicon so we are in the right place, absolutely.
Click to expand...
Since we are all getting old here, that won't be too long.
 
We are using AI. It works well. You have to change your system to take advantage of it. More parallel threading, and you need to create an overall goodness factor. In fact, once you have think you have a solution and think you are done, you can get notified later that there is a better solution to consider. $200B (or so) per year in semis is speeding up the calculations.

Arthur, together, I believe we can knock everybody out of work by 2025q3, but what do I know. Hist78, what does ChatGPT say?
 
It is getting better in leaps and there is an explosion in the number of groups who see uses for the tech and are learning how to do better with existing models, not just future bigger models.

Forget replacing us. It is going to replace politicians in the next election cycle, just like Hatsune Miku for politics. You won't need to stomach watching Mitch for the republican spokesperson.
 
You must log in or register to reply here.
Back
Top