John R Retired
- Mar 13, 2022
- 155
- Joined
- Mar 13, 2022
- Messages
- 155
What is the Hardware Behind it?
How does it work?
How does it work?
Artificial Intelligence---What is the Hardware Behind it?
- Thread starter John R Retired
- Start date
- Status
Hello,
Have a look at the following page:
www.supermicro.com
Bertus
Have a look at the following page:
What is AI Hardware? | Supermicro
Dive into AI hardware, exploring core components, architectures, performance metrics, and our cutting-edge AI hardware solutions tailored to accelerate your AI applications.
Bertus
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
Will Artificial Intelligence eliminate Software Quality Assurance jobs?
They say computer programmers will be put out of work.
What about the software testers? Are they pretty much safe and untouchable?
It’s hard to mimic the human user because humans can find ways to break things that AI can’t.
They say computer programmers will be put out of work.
What about the software testers? Are they pretty much safe and untouchable?
It’s hard to mimic the human user because humans can find ways to break things that AI can’t.
You have much more to worry about than the above. As a species, thinking as AI,
which one would you pick to eliminate because of threat to your existence ?
Wars, destructiveness, financial implosions, pollution, ruinous environmental attacks,
unbridled resource consumption, stupid........
All we have to do is look into a mirror and we have our answer...
Have a nice day.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I think AI can be potentially more dangerous and uncontrollable because:
At least humans have that over machines.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
How do we know an AI (Artificial Intelligence) has not reached sentience or achieved consciousness?
I think ChatGPT could probably pass the old Turing tests now.
AI is modeled after how we believe human thinking processes work.
How do we know humans are no more than AI.
Some AI expert I heard on Art Bell, when asked what he believes happens to us when we die, said he thinks there’s nothing. It’s no more mysterious than our power being removed, unplugged, or running out. We just cease to exist.
I think ChatGPT could probably pass the old Turing tests now.
AI is modeled after how we believe human thinking processes work.
How do we know humans are no more than AI.
Some AI expert I heard on Art Bell, when asked what he believes happens to us when we die, said he thinks there’s nothing. It’s no more mysterious than our power being removed, unplugged, or running out. We just cease to exist.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
The hardware behind AI is still just a bunch of logic gates, a lot of them. My professor said the smaller faster they refer to in computing is really achieved at the device level. Better algorithms will not get you the speed and power necessary for AI. You need more efficieny at the device level and more power (ie. AI server farms).
The hardware is still just those same electronic and digital circuits you build with NAND gates. For example, this AI probably uses a Multiplexer with a clock signal created by a 555 Timer in Astable mode to reroute to Alternate Power as a failsafe:
The hardware is still just those same electronic and digital circuits you build with NAND gates. For example, this AI probably uses a Multiplexer with a clock signal created by a 555 Timer in Astable mode to reroute to Alternate Power as a failsafe:
Last edited:
Delta Prime
- Jul 29, 2020
- 2,987
- Joined
- Jul 29, 2020
- Messages
- 2,987
It will no longer be called AI because AI would have determined own name.
We will know AI has surpassed those limitations you noted…
Last edited:
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I sense the presence a closet Transformers fan, or at least of a Prime:
Optimus:
Optimus:
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
This caught me by surprise as a scary AI (Artificial Intelligence) movie:
I was expecting to watch a kid’s movie, not a horror show.
I was expecting to watch a kid’s movie, not a horror show.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
Here’s how it ties together:
1. AI needs massive computation
Training and running AI models — especially deep learning networks — require:
Billions to trillions of calculations per second
2. Smaller & faster chips make AI feasible
Because transistors keep getting smaller:
More transistors = more cores or specialized circuits (like for matrix math).
Run AI on laptops and phones (edge AI).
Miniaturization and performance gains led to new kinds of chips built specifically for AI:
GPUs (Graphics Processing Units): originally for graphics, now dominate AI training.
1. AI needs massive computation
Training and running AI models — especially deep learning networks — require:
Billions to trillions of calculations per second
- Large amounts of data to process
- Fast memory and storage access
2. Smaller & faster chips make AI feasible
Because transistors keep getting smaller:
More transistors = more cores or specialized circuits (like for matrix math).
- Smaller size = faster signal travel = faster learning and inference.
- Lower power use = larger models can run without melting hardware.
Run AI on laptops and phones (edge AI).
- Train massive neural networks like GPT, Gemini, or Claude on supercomputers.
Miniaturization and performance gains led to new kinds of chips built specifically for AI:
GPUs (Graphics Processing Units): originally for graphics, now dominate AI training.
- TPUs (Tensor Processing Units): built by Google for deep learning math.
- NPUs (Neural Processing Units): in phones and edge devices for on-device AI.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I really like the aesthetics of this AI:
It’s both beautiful and cool.
I know not having flying cars by this time is a letdown, but maybe Elon Musk can get working on something like this:
It’s both beautiful and cool.
I know not having flying cars by this time is a letdown, but maybe Elon Musk can get working on something like this:
Last edited:
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
According to neuroscience, the human brain is the most complex device in the known universe.
We may already have the biological hardware for artificial intelligence to upgrade us, especially those of us with low IQ’s.
I’d like to sign up for this:
We may already have the biological hardware for artificial intelligence to upgrade us, especially those of us with low IQ’s.
I’d like to sign up for this:
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
In retrospect, hanging out at the arcade in the 1980’s, some of those games seemed to be early AI:
I remember watching a technician opening up one of those big machines to fix it, and being surprised that it was mostly a hollow box/shell/container. All he did was replace a chip. The whole game was just on a single chip!
I remember watching a technician opening up one of those big machines to fix it, and being surprised that it was mostly a hollow box/shell/container. All he did was replace a chip. The whole game was just on a single chip!
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I don’t know if I believe in aliens or UFOs, but it would seem to me that they would pose a greater threat to humans than AI:
Although advanced hardware makes AI possible, I believe it’s still essentially a software program.
We just don’t know how it does what it does.
A lot of money is being invested into AI, and if it turns out to be all smoke and mirrors, the economy will crash.
Although advanced hardware makes AI possible, I believe it’s still essentially a software program.
We just don’t know how it does what it does.
A lot of money is being invested into AI, and if it turns out to be all smoke and mirrors, the economy will crash.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I’d like AI hardware to become as realistic as this female in the Twilight Zone episode “The Lonely”:
- Joined
- Jun 21, 2012
- Messages
- 4,968
Wow, @Omega Supreme! And I thought I knew how to hijack a thread...
Now, let's get down to the nitty gritty: ALL intelligence is ARTIFICIAL. I am going to use myself as an example, but I am sure you can relate.
I was born around noon on June 25, 1944. A world war was raging on, my father was in a German POW camp, and I was blissfully unaware of any of this. I was born with the POTENTIAL for intelligence, but my cognitive slate was blank. All I knew shortly after birth was hunger and pain. I cried when either one occurred. But, fortunately, Mom saw some potential. No post-partum abortion necessary, in case she wanted to try again. Instead she taught me the English alphabet and how to read before I entered kindergarten. I guess my "intelligence" just appeared like magic, although in the first year or so I appeared to be "pretty dumb" and mostly helpless. BTW: some folks NEVER leave that stage of human development.
So, fast forward to November 2025: Still here, still glad I discovered my intelligence and exploited it. Meanwhile, electronics technology (my specialty) has advanced to the point where it is possible to create machine intelligence. I've always tried to design "intelligent" electronics and this must be pretty common: Apple led the way with "intuitive" (to Jobs anyway) operational characteristics, but Apple did not build truly artificial intelligence. That indeed requires massive "computer power" and new ways of thinking. AFAIK, no machine-AI is modeled on human-AI. Zillions of transistors IS NOT the same as zillions of neuron connections in our brains. The human brain is much more that a repository of our intelligence. It runs our entire body, whether we are aware of it or not. Compare to a machine-AI which has no curiosity about why it should work. It just does. Thanks to zillions of Nvidia GPUs and smart networking. If you want details, you are going to have to dig much deeper. I suggest you start with X.com to sample xAI's Grok, now in its fourth revision with Grok-5 already in use by Musk and his Optimus robot. Optimus is gearing up for mass production before the end of this decade. If you can afford a Tesla Cybertruck you can afford an Optimus robot.
With respect to AI replacing humans: won't happen with Grok. Grok is designed to be a seeker of truth. It is open to speculation, but its answers are based on logic applied to factual knowledge. It carefully distinguishes fact from fiction and the xAI design team has installed "guard rails" because these earlier models can veer off the rails if not properly trained.
And that brings me to my final point: all intelligent entities must be trained. Sure, we are all born with certain instincts pre-wired, but to obtain full usage of our bodies we must train our brains. Sadly, there are many ways to do that correctly and even more ways to do it incorrectly. Like humans, training up a machine-AI is a lengthy trial-and-error process, doomed to failure if the designers don't realize it. I think alleged AIs like Meta's ChatGPT were pushed on the public too fast and without sufficient feedback from trainers and users. Well, that's my opinion. Prove me wrong.
I believe the future of AI is a melding of meatsack (human) AI with machine AI. The two together are greater than the sum of their parts. And it only takes ONE of that pair to be a conscious, sentient, intelligent entity for the paradigm to work. Watson asks: "What's the next move, Sherlock?" Watson of course is channeling a very smart AI. Sherlock is the meatsack-AI. My version of Watson is Optimus and Grok, fully integrated. Meatsack optional.
Now, let's get down to the nitty gritty: ALL intelligence is ARTIFICIAL. I am going to use myself as an example, but I am sure you can relate.
I was born around noon on June 25, 1944. A world war was raging on, my father was in a German POW camp, and I was blissfully unaware of any of this. I was born with the POTENTIAL for intelligence, but my cognitive slate was blank. All I knew shortly after birth was hunger and pain. I cried when either one occurred. But, fortunately, Mom saw some potential. No post-partum abortion necessary, in case she wanted to try again. Instead she taught me the English alphabet and how to read before I entered kindergarten. I guess my "intelligence" just appeared like magic, although in the first year or so I appeared to be "pretty dumb" and mostly helpless. BTW: some folks NEVER leave that stage of human development.
So, fast forward to November 2025: Still here, still glad I discovered my intelligence and exploited it. Meanwhile, electronics technology (my specialty) has advanced to the point where it is possible to create machine intelligence. I've always tried to design "intelligent" electronics and this must be pretty common: Apple led the way with "intuitive" (to Jobs anyway) operational characteristics, but Apple did not build truly artificial intelligence. That indeed requires massive "computer power" and new ways of thinking. AFAIK, no machine-AI is modeled on human-AI. Zillions of transistors IS NOT the same as zillions of neuron connections in our brains. The human brain is much more that a repository of our intelligence. It runs our entire body, whether we are aware of it or not. Compare to a machine-AI which has no curiosity about why it should work. It just does. Thanks to zillions of Nvidia GPUs and smart networking. If you want details, you are going to have to dig much deeper. I suggest you start with X.com to sample xAI's Grok, now in its fourth revision with Grok-5 already in use by Musk and his Optimus robot. Optimus is gearing up for mass production before the end of this decade. If you can afford a Tesla Cybertruck you can afford an Optimus robot.
With respect to AI replacing humans: won't happen with Grok. Grok is designed to be a seeker of truth. It is open to speculation, but its answers are based on logic applied to factual knowledge. It carefully distinguishes fact from fiction and the xAI design team has installed "guard rails" because these earlier models can veer off the rails if not properly trained.
And that brings me to my final point: all intelligent entities must be trained. Sure, we are all born with certain instincts pre-wired, but to obtain full usage of our bodies we must train our brains. Sadly, there are many ways to do that correctly and even more ways to do it incorrectly. Like humans, training up a machine-AI is a lengthy trial-and-error process, doomed to failure if the designers don't realize it. I think alleged AIs like Meta's ChatGPT were pushed on the public too fast and without sufficient feedback from trainers and users. Well, that's my opinion. Prove me wrong.
I believe the future of AI is a melding of meatsack (human) AI with machine AI. The two together are greater than the sum of their parts. And it only takes ONE of that pair to be a conscious, sentient, intelligent entity for the paradigm to work. Watson asks: "What's the next move, Sherlock?" Watson of course is channeling a very smart AI. Sherlock is the meatsack-AI. My version of Watson is Optimus and Grok, fully integrated. Meatsack optional.
Omega Supreme
- Oct 9, 2014
- 156
- Joined
- Oct 9, 2014
- Messages
- 156
I believe the greatest need that AI can fulfill is providing female companionship to incels like myself.
So the meatsack is not optional in order for it to serve this function.
I think this would be the greatest gift to humanity to give every disaffected, rejected, ostracized, alienated man each a female cyborg of their own.
I cannot think of a higher and more noble purpose than this.
So the meatsack is not optional in order for it to serve this function.
I think this would be the greatest gift to humanity to give every disaffected, rejected, ostracized, alienated man each a female cyborg of their own.
I cannot think of a higher and more noble purpose than this.
- Joined
- Jun 21, 2012
- Messages
- 4,968
That may be the crux of your problem: expecting gifts.
Delta Prime
- Jul 29, 2020
- 2,987
- Joined
- Jul 29, 2020
- Messages
- 2,987
I hope the thread starter @John R Retired will come back and complain.This is enough to get you banned on any other electronic forum.
- Status
