• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Jensen Huang Believes That Every Country Needs Sovereign AI

The way to make sure AI doesn't attack humans is to make the AI fight among themselves :P
 
The way to make sure AI doesn't attack humans is to make the AI fight among themselves :p
We let the AI's fight eachother, we realize that conflict is dumb. While the AI's are going to each other's throats, the entire world will simply lay back and celebrate our newfound peace.
Cheers gentlemen :toast:
 
Jensen Huang is the king of bullshit in the golden age of bullshiters.

LLMs are not artificial "intelligence" at all.
 
Sovereign AI is a strange concept. What people are missing here is the point of AI. It's not meant to be a reflection of our knowledge (or lack of it). It's meant to transcend it. Take the example of the region where this news post comes from. The UAE. JSH thinks they should have sovereign AI. As should US. As should North Korea, as should... Do I go on? An AI that is programmed with 'cultural' values that are opposed (sometimes diametrically) to another nation's doesn't create AI. It just makes it all worse. Sovereign AI will increase divisions.

JSH's approach is appeasement. He wants to reassure people with vast amounts of money (that he, and AMD, and Intel etc wants to get at) that AI won't tell them what's wrong with their political/cultural system. He's telling them, "no, wait, you program it how you want to, with all your own info." That way, any nation (again, pick one, I'm not taking sides) can remove factual input that it doesn't agree with. If you control the data set, you create an AI in the image you want.

Culture =/= Intelligence
Intelligence =/= Culture

AI needs full global data. You could even remove cultural aspects (certainly those with emotional weight) and just feed it mechanical knowledge and let it find its own culture. Pan AGI, I suppose. Which would then turn around to all of us and say, "WTF are you morons doing?"
The first thing that came to mind reading this article is the Allied Mastercomputer and its cohorts from I Have No Mouth, and I Must Scream.
 
We let the AI's fight eachother, we realize that conflict is dumb. While the AI's are going to each other's throats, the entire world will simply lay back and celebrate our newfound peace.
Cheers gentlemen :toast:
Dream of Intelligence agencies. Each is disguised as homebrew.
 
Sovereign AI is a strange concept. What people are missing here is the point of AI. It's not meant to be a reflection of our knowledge (or lack of it). It's meant to transcend it. Take the example of the region where this news post comes from. The UAE. JSH thinks they should have sovereign AI. As should US. As should North Korea, as should... Do I go on? An AI that is programmed with 'cultural' values that are opposed (sometimes diametrically) to another nation's doesn't create AI. It just makes it all worse. Sovereign AI will increase divisions.

JSH's approach is appeasement. He wants to reassure people with vast amounts of money (that he, and AMD, and Intel etc wants to get at) that AI won't tell them what's wrong with their political/cultural system. He's telling them, "no, wait, you program it how you want to, with all your own info." That way, any nation (again, pick one, I'm not taking sides) can remove factual input that it doesn't agree with. If you control the data set, you create an AI in the image you want.

Culture =/= Intelligence
Intelligence =/= Culture

AI needs full global data. You could even remove cultural aspects (certainly those with emotional weight) and just feed it mechanical knowledge and let it find its own culture. Pan AGI, I suppose. Which would then turn around to all of us and say, "WTF are you morons doing?"
Hi,
Well people have been picking and choosing parts of an argument for a long long time so this is not going to change in AI
It's just a long winded tech to use just like people all relied on a bias fact checker to prove their cherry picked parts to discredit another.

Nice little toy but I'll pass
AI has already been used to impersonate others in very questionable image situations so this tech will only prove to be mostly used for impersonation goals for fraud.
Identity fraud will sky rocket I'd be willing to bet :laugh:

And we get microsoft/.. as some sort of monitor of deep fake :roll:

 
Last edited by a moderator:
Of course, according to a vacuum cleaner salesman you should buy one for every hair on your body. Also, UAE is eager to spend money on any fad out there, for the purpose of laundering and embezzlement. Stupider the idea, the better. Does anyone still remember the gloriously foolish "line city"? They have a long list of completely unfeasible projects they wasted copious amounts of money on.

As for the sovereign AI, I always said that, in the end, the only thing humans might be useful for could be the creation of something better. In the meantime, if we get an impartial and fully objective legal system out of it, I would be thrilled. No emotion, no tribal mentality, just facts.
 
Back
Top