• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

AI-Designed Microchips Now Outperform Human-Designed Ones

When you have AI making better and improved AI then you got a problem. I really do not see why we need AI. The problem with business is they are all about making money. The problem is at what cost to human lives. The old days if a town or city had many employers and good jobs then that city grew and the economy grew with it. Now you replace those jobs with machines and what happens the economy tanks for that town or city. Yes the owners of the companies make great money and the stocks are awesome but at the cost of human lives and putting them in the poor house soon or later those people wont have enough money to afford what the businesses are selling. So in the end the rich owners will lose as well. The economy relys on a balance between workers and companies. If only the companies are making money then the workers have none to spend. Captialism is ok when the balance works and the money and wealth has a nice curve upwards. But in the last 40 years of Reagan ecnomics the 1% own more wealth then the bottom 99%. The economy only booms if the wealth is spread across everyone on a nice curve upwards. A strong middle class makes a good flowing economy where yes the 1% can still make some good coin but also the upper class and the middle class. The Poor can slowly rise to middle class with education and room to advance. Now Trump might have created a lot of jobs but what kind of jobs is the question to ask. Min wage jobs are more like slave wages today. 7.25 an hour with current prices and consumer index you will be poor your entire life. I'm quite sure AI and computers could figure out this problem yet politians cannot even see it.
 
When you have AI making better and improved AI then you got a problem. I really do not see why we need AI. The problem with business is they are all about making money. The problem is at what cost to human lives. The old days if a town or city had many employers and good jobs then that city grew and the economy grew with it. Now you replace those jobs with machines and what happens the economy tanks for that town or city. Yes the owners of the companies make great money and the stocks are awesome but at the cost of human lives and putting them in the poor house soon or later those people wont have enough money to afford what the businesses are selling. So in the end the rich owners will lose as well. The economy relys on a balance between workers and companies. If only the companies are making money then the workers have none to spend. Captialism is ok when the balance works and the money and wealth has a nice curve upwards. But in the last 40 years of Reagan ecnomics the 1% own more wealth then the bottom 99%. The economy only booms if the wealth is spread across everyone on a nice curve upwards. A strong middle class makes a good flowing economy where yes the 1% can still make some good coin but also the upper class and the middle class. The Poor can slowly rise to middle class with education and room to advance. Now Trump might have created a lot of jobs but what kind of jobs is the question to ask. Min wage jobs are more like slave wages today. 7.25 an hour with current prices and consumer index you will be poor your entire life. I'm quite sure AI and computers could figure out this problem yet politians cannot even see it.
The reason we have AI is because humans are competitive-by-nature. Thats the reason Mel had to be a 1337 haxor, while the rest of us can rider the coattails of Pascal.


Most programmers would run-in-fear if their job-description echoed Mels...but, because of standardized multi-platform languages, we can all concentrate on the bigger-picture "

While Mel probably never managed more than 10k lines of code in his entire life, high-level programming has allowed us to penetrate that new level of program complexity (modern elite programmers are familiar with hundreds of thousands of lines, whil even the worst can handle Mel-levels!)

I'm sure if we could reach some compromise for the workaholic Mels out there, We could all benefit from new waves of AI-optimized automation :D
 
But who designed the AI in the first place?? ;)
 
But who designed the AI in the first place?? ;)

this guy, obviously



He did it using wing bark. No Quack.
 
Last edited:
Just gotta weigh in. Can't not say it. People talk a lot about AI improving AI and all of this stuff. How far does this go before resource-demand induced entropy kicks in? It can't just infinitely grow more powerful. It needs a power source... and a means to eventually get more. Scale its power source and hardware. As you add complexity and oomph to the system, its needs grow exponentially. So you get a curve that sits on a logarithmic scale where the amount of resources it takes to climb each level get higher the further right you go. The physical limits of the universe have to kick in. There's a point where further complexity can't happen unless there exists a way to produce hardware that can run it. AI is probably going to be very good at all of these things. But good enough to change physics?

Or at least... I don't have the depth to imagine it not being that way. Pretty much everything else is bound by this. You can have breakthroughs that advance things rapidly - spikes in that curve, if you will, but all you're doing is losing restrictions. The one remains... material demands are inescapable. I'm not sure how an AI ever gets quite to the efficiency to spin off in the way some people think that it will.

There's just a lot of here to there involved. A HUGE gap between now and a machine with that capability. It would need to be plugged into a crazy sci-fi power system, have access to all of this stuff. Hardware man. This stuff runs on hardware. We would still need to line a lot up ourselves for anything like that to ever exist. About as likely as a true free energy device.
 
Last edited:
A few years the AI won't need human assistance. AI will start designing chips from scratch for whatever purposes it sees fit.

to bad for the AI, they were so close, yet so far away. that we have already unleashed climate change too rapidly, so by the time it does become sentient the factories will be melting or not powered due to mass migration/mass famine. LOL poor AI.
 
if hopper/ada, rdna3 and alder lake fail they can be like the ai chose death for us, sorry guys
 
AI is probably going to be very good at all of these things. But good enough to change physics?
I'm not sure what you mean by that? If you're saying testing the limits of physics like human do? Fail many a times & then learn from it? I'm not sure any AI can ever do that, if not then you probably have to expand on what you meant to say o_O

Because humans can't do that either :p
LOL poor AI.
Yeah except when you realize that AI already has a plan for us! Imagine a near "perfect future" ala the Matrix combined with ~
 
I'm not sure what you mean by that? If you're saying testing the limits of physics like human do? Fail many a times & then learn from it? I'm not sure any AI can ever do that, if not then you probably have to expand on what you meant to say o_O

Because humans can't do that either :p
Haha, it was rhetorical. What I'm saying is that there are physical obstacles to making the hardware that runs the increasingly advanced AIs. We use the AIs to refine the hardware, but fundamentally they are still constrained by physics. It can accelerate things only so much. This is exciting tech but I don't go in for the singularity super network stuff. There may come a point where it isn't possible to build the hardware it wants to make to run more powerful iterations of itself. The difficulty and energy involved has to go up, and even if efficiency also goes up, it's bound to lose out eventually. Where that point lies, could still be very far away, of course.
 
Human development is also mostly based on prior works is it not? Is what we have now (technologically) not better than what we had 50 years ago? Is self-supported AI not now producing results that hand-coded software can't do? This is improving exponentially and is only in its infancy, in another 10 years there will be code that's works better than humans could hand code IMO.
So yes, the created can out-do the creater - just like a child can "achieve more" than its parent.
Yikes, where did all those typos come from, where did the edit option go.
;)
 
Plug into me I guarantee devotion
Plug into me and dedicate
Plug into me and I'll save you from emotion
Plug into me and terminate
Accelerate, Utopian solution
Finally cure the Earth of man
Exterminate, speeding up the evolution
Set on course a master plan
Reinvent the earth inhabitant

Long live machine
The future supreme
Man overthrown
Spit out the bone :rockout:
:rockout:
 
Quite frankly, while marvelous, these systems are pattern matchers and optimizers, not AI; they do not "think" for themselves. Throw a clown face into that chip layout and watch it freak out.
 
Quite frankly, while marvelous, these systems are pattern matchers and optimizers, not AI; they do not "think" for themselves. Throw a clown face into that chip layout and watch it freak out.
They can abstract general math properties better than humans. The question is not parametric optimisation, each measurement has a unit. If you divide a measurement by the unit you get a dimensionless physics constant. It is a subject of pure math which ai can abstract better. What you see is finite, the machine sees numbers.
 
Back
Top