Where in the AI World Are We??

  • Christian Chat is a moderated online Christian community allowing Christians around the world to fellowship with each other in real time chat via webcam, voice, and text, with the Christian Chat app. You can also start or participate in a Bible-based discussion here in the Christian Chat Forums, where members can also share with each other their own videos, pictures, or favorite Christian music.

    If you are a Christian and need encouragement and fellowship, we're here for you! If you are not a Christian but interested in knowing more about Jesus our Lord, you're also welcome! Want to know what the Bible says, and how you can apply it to your life? Join us!

    To make new Christian friends now around the world, click here to join Christian Chat.

Karlon

Well-known member
Mar 8, 2023
3,209
1,485
113
#61

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#62

Ex-Google CEO Issues TERRIFYING AI Warning — “We’ve Lost Control”

"We believe in the next 1 year the vast majority of programmers will be replaced by AI programmers"

AI is advancing much faster than anyone predicted. This is precisely what I have been arguing, that they will be telling you the changes are much further in the future to keep people from becoming alarmed.

AI is dominating the healthcare industry now, they can design and create all kinds of treatments, but they can also design and create all kinds of pathogens for biowarfare.

"Every single Phd project in Chemistry, Biology and Physics is now using AI in some form".

Think about a world in which all programming is done by AI and all scientific research is done by AI. That is the world we are now entering as you read this.

In 2 years AI will be programmed and improved by AI. Deep Mind already has a program to do this now. As a result it is inevitable that all AI will be self improving in the very near future.

Imagine a world in which everything is run by AI, designed by AI, discovered and developed by AI. It will be extremely difficult for humans to even understand what things are, much less how they work or how to fix them. We know that computers and AI can do repetitive tasks and mundane tasks, but now we are learning they can also do the most creative and cutting edge tasks. It doesn't simply replace cashiers, drivers, mechanics, and maids, it can also replace scientists, researchers, bioengineers, software engineers, etc.

Within 3 years we will have AGI. This means that the AI will not simply be ChatGPT level, but will be at the level of the best human on any topic. If you want to design a house, simply ask it to do that, if you want it to give you legal advice or medical advice just ask it. That may not eliminate every job but it certainly means we don't need to pay humans anymore than simply a minimum wage. That is 3 years away!

The good news is it can take away most jobs in three years, the bad news is it will take enormous amounts of power, so we won't need humans for jobs and we can't afford to let useless humans use power that we need for AI.

Consider this, the PC was first made in 1975, about 35 years earlier the first computer was being made to break German codes, and in 2005 we first had Youtube. So we have had a full 65 years to go from having a computer do something amazing to a full fledged internet. It took us about 80 years before we had our first glimpse of AI. But within a short period of 7-8 years AI will go from not very good to Super intelligence. We had 80 years to make the transition to the computer age, and now we are going to have at most, 8 years to make the transition to AI.

He argues that in all times in the past automation didn't take jobs, it simply caused them to be changed. I disagree. Automation has taken jobs, it took jobs from horses, and cows and oxen. But those machines could not build themselves, the AI can do that. Those machines could not sell themselves, the AI can do that. Those machines could not fix themselves, robots are being designed that can do that.

AI does not require that we create a new industry, we already have a computer industry, a network industry, a power industry, etc. It does not require we train programmers, it will do that. I do agree it will make workers more productive, but that means one lawyer with AI will be able to do what five or six lawyers did. But it is worse, I can imagine people preferring to go to court with an AI judge or arbitration will be done with an AI judge. I can also imagine an AI lawyer like Turbotax that can answer 50% of the questions that normally a lawyer would have needed to answer. Likewise, instead of seeing a doctor I think 50% of your medical questions will be able to be answered by an AI. So service will be better, cheaper, faster, but at the same time a clinic will likely need half the staff to do the work it now does. The argument that AI will increase productivity without taking jobs doesn't hold water. Already we are seeing fully automated fast food restaurants that can operate without any workers. No workers means no liability, faster service, cheaper prices. Why would anyone prefer a sixteen year old kid who makes viral videos to make his food instead of a robot?

For example, all these computer programmers who are being laid off, is there a new job that the AI is creating for them to go into? No.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#63

Ex-Google CEO Issues TERRIFYING AI Warning — “We’ve Lost Control”
He also explains that AI is a network business and that network business leaders tend to get 90% of the business. This means whoever gets to ASI first will get 90% of the world's intelligence. This explains the wars, it is an existential threat. If you can't stop another country from getting to ASI before you then you may feel you have to bomb their data centers.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#64
Huge Breakthrough

AI Chemist's Breakthrough: Room-Temp Superconductor!


We lose between 5% and 15% of all electricity worldwide because of resistance. Room temp superconductor would save us that power. It is especially big seeing as AI is going to be a major consumer of electricity.

In addition it may make energy storage in batteries far more efficient so we can store power generated at night until the next day. This means a much more even running of the power plants. It also means that wind power becomes far more economic. It also would make mag lev trains practical. That would result in faster and more reliable transportation that is also significantly cheaper because of lower maintenance costs and reduced energy consumption.

This would also be a very big breakthrough for Quantum computers which need to use superconductors.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#65

NVIDIA CEO Jensen Huang Leaves Everyone SPEECHLESS (Supercut)

"We're building AI factories".

4 million square feet. 1 gigawatt.

Nvidia MV link fusion --

What is incredible to me is that all AI systems can immediately upgrade and when they do they'll be significantly faster and more powerful. You can improve AI by programming, you can also improve it by improving the hardware. This week we have seen massive improvements in the programming so that AGI is clearly on the very near horizon, and now you plug them into this and the performance of all this AI will instantly be significantly better. You can also improve AI by training and we now know that AI can train itself and program itself and we have much more powerful platform to do this on. This is at least 4x more powerful than the platform that AI used to solve the problem of folding protein structures which no human had been able to solve, and at least 2x more powerful than the AI system which may have just solved the problem of room temperature superconductors.

We now need to solve the problem of Fusion reactors but we have gotten them up to operating for 22 minutes and that was with a system that was at most half as powerful as this. A room temperature superconductor is a huge advance for AI, but Fusion would give it unlimited power. We are witnessing the rise of the Beast system.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#66

NVIDIA CEO Jensen Huang Leaves Everyone SPEECHLESS (Supercut)

"We're building AI factories".

4 million square feet. 1 gigawatt.

Nvidia MV link fusion --

What is incredible to me is that all AI systems can immediately upgrade and when they do they'll be significantly faster and more powerful. You can improve AI by programming, you can also improve it by improving the hardware. This week we have seen massive improvements in the programming so that AGI is clearly on the very near horizon, and now you plug them into this and the performance of all this AI will instantly be significantly better. You can also improve AI by training and we now know that AI can train itself and program itself and we have much more powerful platform to do this on. This is at least 4x more powerful than the platform that AI used to solve the problem of folding protein structures which no human had been able to solve, and at least 2x more powerful than the AI system which may have just solved the problem of room temperature superconductors.

We now need to solve the problem of Fusion reactors but we have gotten them up to operating for 22 minutes and that was with a system that was at most half as powerful as this. A room temperature superconductor is a huge advance for AI, but Fusion would give it unlimited power. We are witnessing the rise of the Beast system.
He said there are currently $5 trillion worth of AI "factories" or plants being built. He doesn't simply call them data centers. But $5 trillion tells me the world is all in.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#67

Sam Altman And Elon Musk Just Revealed Their Next AI Models...

Wow, I get it! The reason it is so important for AI to be excellent in Math and Physics is because only then can it accurately make a simulation to train AI on. They need to train AI in simulations so that their robots and computers can be fully trained. They have used all the human data available, but that data is not enough. In a fighter jet they can only fly so fast because of Gforces, but for AI and robots Gforces are not a limiting factor. That is simply one example of why they have to be able to simulate real world events accurately so that the AI can be trained properly.

They trained Alpha Go to beat the best Go players in the world and it took about six months. The computers they are using now are 4x more powerful so we can do that training in six weeks. However, the software is better at training and so we might be able to cut that to four weeks. So the speed of development is much faster. What used to take four years will now be done in a year. However, when they can get the AI on Quantum computers problems that took us six months a few years ago will take six seconds. We will have more developments in one month than we had in the previous 100 years.

AI is on the cusp of solving room temperature superconductors and Fusion. I think it is reasonable to say that if we put our existing AI on Quantum computers we will have ASI, we will solve Fusion, and then the ASI will also have unlimited power supply.

Sam Altman describes where we are with AI right now as being similar to where we were when the PC was first made or when people first started web design. However, the rate of change is 100x faster than when we first had a PC and probably 10x faster than when we first had web sites being put up. The amount of change we saw in the last 80 years since Alan Turing we'll probably see in the next year.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#69
The half was not told me


EXT. ULTRA-MODERN EXPRESSWAY – LATE AFTERNOON


A sleek, autonomous luxury car glides down a spotless expressway, elevated above a glowing cityscape. Everything gleams. Trees shimmer with engineered foliage. Solar panels glint like crystal fields. Drones dart overhead delivering goods. Billboards silently flash perfect lives.


INT. LUXURY VEHICLE – CONTINUOUS


MARA VELASQUEZ (40s), dignified, composed, gazes out the tinted window in silent reverie. She's overwhelmed, hushed, like someone who’s seen a divine vision. A small tear escapes — not of sorrow, but of rapture.


MARA
(to herself)
The half was not told me…


Her driver, ANTON (50s), stoic and well-groomed in a minimalist uniform, glances at her through the mirror.


ANTON
There's something else you might want to see. If you're not in a rush.


MARA
Something more?


ANTON
In a way.


He signals. The car smoothly exits the expressway, leaving the elevated dream behind.



EXT. CITY STREETS – MOMENTS LATER


They descend. The mood changes. The hum of progress fades. The air thickens. The sunlight doesn’t shine the same way down here.


They pass a row of underpasses. Mile after mile — tents. Blue tarps flapping in the breeze. Children running barefoot across concrete. Fires in oil drums. Makeshift signs: “Former Engineer,” “Looking for Work,” “AI Took My Job.”


MARA
What... what is this?


ANTON
The others. The ones who weren’t invited up.


MARA
But they’re... they look like—they are—middle class.


ANTON
They were.


A man in a suit with dirt on his cuffs adjusts a solar panel outside his tent. A woman with a wireless tablet teaches a small group of children — a classroom on the curb.


MARA
Stop the car.



EXT. ENCAMPMENT – CONTINUOUS


Mara steps out, heels clicking awkwardly against the broken pavement. She walks slowly, absorbing it. No words. Just a series of images.


A father trims his son’s hair with surgical precision under the light of a jury-rigged drone lamp.


A woman in a once-elegant dress uses VR goggles — escaping? Working? Learning? We don’t know.


A giant billboard above the encampment flickers with an ad: “Welcome to the Future. No Limits.” Below it, someone has spray-painted: “Except Us.”


Mara stops at a tent where a couple in their 30s sit cross-legged, sharing a can of soup.


MARA
(quietly, to the couple)
May I ask what happened?


MAN IN TENT
(shrugs, not bitter)
They made a new logistics AI. Ran our whole department better than we ever could. Took 43 of us a week to do what it did in an hour.


WOMAN IN TENT
We had savings. Sold the condo. Lived in our car. Now here.


MARA
And no one helped?


WOMAN IN TENT
They called it progress.


Mara looks around. Silent. Her face changes. The shine is gone from her eyes. Something deeper sets in — comprehension.



INT. LUXURY VEHICLE – LATER


Back on the road. Mara looks at the expressway above them, then down at her hands.


MARA
I thought I’d seen the future.


ANTON
You did. Just not all of it.


The car merges back onto the expressway. But now, it doesn’t feel like utopia. It feels... detached. Above everything.


As they ascend, the camera lingers below — on a child drawing a house in chalk next to a tent.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#70
Tesla CEO Elon Musk says demand for Optimus robots 'will be insatiable'

 
Nov 25, 2024
521
246
43
#71

Mem

Senior Member
Sep 23, 2014
7,829
2,404
113
#74
I wonder what ai would answer if asked, "Do you love me?"

Anyway, I really have to like a movie to sit through it to the end these days and the last one I watched was about AI called "Atlas." The title role isn't the AI in the movie, but she has been traumatized by and is severely distrustful of it, and happens to have to work with it...
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#75
"How can we keep up with this?"


The Compression of Time in the AI Era

One issue that AI agents have is intent over time. At present there may not be any AI agents today that can stick to the same task for a week. When I was a teacher I planned out an entire year, I had to present this plan to my principal, I had to update her every few months on the progress, at the end of the year I had to assess how the plan went and ultimately report this to her prior to making my plan for the next year.

So this is interesting, the intelligence scale of AI is increasing exponentially, but not the intent. That is moving along much slower. Yes, AI can do more the amount of time, but even so, they are finding it very hard to stay with a task for a longer period of time.

Wow -- The CEO of Nvidia said that by using a simulated environment they can now give AI or a robot 10 years of training in two hours! Talk about straight out of the Matrix movie.

Now the implication of the fact that Agents are getting extremely capable but still lack the ability to stick to a job for more than a few days is that we will need human managers. A human manager would have to be aware of all the jobs the entire team is doing, assign tasks to agents and when they complete those tasks give them new tasks. One of the tasks would probably be to write out a list of all tasks that need to be done in order, so that as these agents return having completed one task you can check it off and give them another one.
 

ZNP

Well-known member
Sep 14, 2020
40,742
7,682
113
#76
"How can we keep up with this?"


The Compression of Time in the AI Era

One issue that AI agents have is intent over time. At present there may not be any AI agents today that can stick to the same task for a week. When I was a teacher I planned out an entire year, I had to present this plan to my principal, I had to update her every few months on the progress, at the end of the year I had to assess how the plan went and ultimately report this to her prior to making my plan for the next year.

So this is interesting, the intelligence scale of AI is increasing exponentially, but not the intent. That is moving along much slower. Yes, AI can do more the amount of time, but even so, they are finding it very hard to stay with a task for a longer period of time.

Wow -- The CEO of Nvidia said that by using a simulated environment they can now give AI or a robot 10 years of training in two hours! Talk about straight out of the Matrix movie.

Now the implication of the fact that Agents are getting extremely capable but still lack the ability to stick to a job for more than a few days is that we will need human managers. A human manager would have to be aware of all the jobs the entire team is doing, assign tasks to agents and when they complete those tasks give them new tasks. One of the tasks would probably be to write out a list of all tasks that need to be done in order, so that as these agents return having completed one task you can check it off and give them another one.
So this implies that big companies with say 100 engineers will be able to make maximum use of AI by reducing their staff to the ten most competent engineers and using AI to replace 90. But a company with three engineers could only use AI to replace two engineers.

On the other hand a person can go into business by themself and be like a company with a receptionist, ten engineers, a salesman, an accountant and a personal assistant. All of a sudden one person is like a small company with 15 people.