ChatGPT is an Open Source AI
- Thread starter ZNP
- Start date
-
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.
Let's explain how big this is. The only viable argument I have seen saying AI will not be able to replace all workers is how much power is required to operate AI. Based on power consumption they argue that AI can only replace about 20-25% of workers. This chip obliterates that barrier, now you can replace all workers because the power consumption will be so much less.
What is also scary is they used AI to design this chip. Imagine how much money and energy is being directed right now by AMD, and Intel and all the other chip makers to have AI design new and improved chips that will be exponentially better.
What is also scary is they used AI to design this chip. Imagine how much money and energy is being directed right now by AMD, and Intel and all the other chip makers to have AI design new and improved chips that will be exponentially better.
The third-largest supermarket chain in China has collapsed, and a wave of bankruptcies is underway
Am I missing something or at the same time we are learning that China has a chip that will allow AI to replace 100% of their workers they are also having massive shutdown of food distribution which will lead to starvation and death of their workers?
Is this replacement theory in action?
Am I missing something or at the same time we are learning that China has a chip that will allow AI to replace 100% of their workers they are also having massive shutdown of food distribution which will lead to starvation and death of their workers?
Is this replacement theory in action?
Many people seem to think that Humans are smarter and more creative than AI will ever be and so they will never lose their jobs to AI.
Ask yourself this
UPS delivers packages. If they can do it faster, cheaper and more reliably with machines and AI is there any reason to think that we still need people for that job?
Adidas makes shoes. If they can be made with a 3d printer and other machines, and if AI can process the order, box up and ship out the order, is there any reason to think we need humans involved in that process?
McDonald's makes hamburgers. They have a fully automated restaurant that is drive through only. If they can service you faster, cheaper and more reliably with AI and robots is there any reason to have people involved in that process?
Already AI is performing better than doctors at diagnosing diseases, is there any reason why you would rather have a person do that if the AI is clearly faster, cheaper and more reliable?
You can say the same about data analysis, programming, translating, lawyers, cashiers, mechanics, waitresses, and CPAs.
Ask yourself this
UPS delivers packages. If they can do it faster, cheaper and more reliably with machines and AI is there any reason to think that we still need people for that job?
Adidas makes shoes. If they can be made with a 3d printer and other machines, and if AI can process the order, box up and ship out the order, is there any reason to think we need humans involved in that process?
McDonald's makes hamburgers. They have a fully automated restaurant that is drive through only. If they can service you faster, cheaper and more reliably with AI and robots is there any reason to have people involved in that process?
Already AI is performing better than doctors at diagnosing diseases, is there any reason why you would rather have a person do that if the AI is clearly faster, cheaper and more reliable?
You can say the same about data analysis, programming, translating, lawyers, cashiers, mechanics, waitresses, and CPAs.
Ukraine is saying they are going to produce 1 million drones this year. So I was wondering, if we were to be attacked by a coalition of Russia, Iran and China, how many drones and robots would they employ? Imagine a squadron of 24 drones, if you had a total of 100,000 squadrons you would have 2.4 million drones, that seems a reasonable amount. Imagine in the first hour of attack we were hit by 100,000 squadrons, most flying from the borders (two oceans, the gulf, Canadian border and Mexico border. Imagine how much chaos and carnage that could cause as well as major disruptions. You could take out every truck stop in the country, a thousand bridges, and all kinds of other infrastructure. They could riddle the streets and highways with burned out wrecks making it very difficult to have cars or trucks or trains moving in this country. They could destroy commercial airports causing all air traffic to cease.
China is making robot dogs with machine guns. They can be dropped behind enemy lines. Imagine 100,000 companies of these robot dogs (packs of 24 each) sweeping across the country. Even if we were able to destroy a million of them the amount of mayhem and destruction they could wreak would be astonishing. All in the first day of an attack. They would probably take out most if not all of our police cars, huge numbers of police and sheriffs, as well as local militias.
Has anyone considered that these wars in the Middle East and Ukraine are simply a dress rehearsal for attacking the US, testing their systems out in real combat situations?
China is making robot dogs with machine guns. They can be dropped behind enemy lines. Imagine 100,000 companies of these robot dogs (packs of 24 each) sweeping across the country. Even if we were able to destroy a million of them the amount of mayhem and destruction they could wreak would be astonishing. All in the first day of an attack. They would probably take out most if not all of our police cars, huge numbers of police and sheriffs, as well as local militias.
Has anyone considered that these wars in the Middle East and Ukraine are simply a dress rehearsal for attacking the US, testing their systems out in real combat situations?
We have good news and bad news concerning AI.
First, the good news is that AI is getting to the point where it will be able to solve our biggest problems of this planet. The bad news is that AI has determined that the cause of most if not all of these problems is humans. Eliminate humans and the problem is solved. Though to be fair Iranian AI only sees certain humans as the problem called "the Great Satan".
The A.I. Terminator | Billy Crone
First, the good news is that AI is getting to the point where it will be able to solve our biggest problems of this planet. The bad news is that AI has determined that the cause of most if not all of these problems is humans. Eliminate humans and the problem is solved. Though to be fair Iranian AI only sees certain humans as the problem called "the Great Satan".
The A.I. Terminator | Billy Crone
AGI in 7 Months! Gemini, Sora, Optimus, & Agents - It's about to get REAL WEIRD out there!
I would say that people thinking that when we have AGI that will be the end of society as we know it are way, way off. A fortune 500 company does not hire someone based on what they can produce tomorrow. They are looking at what they will produce for the company over the next ten years, perhaps over the next five years. The minute they think that in ten years we won't need this new hire, or in five years we won't need this new hire they are looking at ways to not hire them. There is a training cost, and there is liability. So you don't take on those costs and liabilities unless you think there is a high likelihood of a payoff.
This is why we have been seeing layoffs in the human resources department. If you aren't planning on hiring you don't need so many people in that department.
To me what signaled the end of society as we know it was learning that Tesla robots can be trained by watching a Youtube video.
Watson winning Jeopardy demonstrated that AI has the general knowledge to be in the top 0.1% of human knowledge. AlphaGo demonstrated that AI has the ability to come up with new, creative and revolutionary strategies and concepts. Deep Mind's program which does weather forecasts demonstrates the superiority and cost savings of using AI to work on very difficult and complex tasks. Meanwhile tesla robots can replace blue collar workers.
I would say that people thinking that when we have AGI that will be the end of society as we know it are way, way off. A fortune 500 company does not hire someone based on what they can produce tomorrow. They are looking at what they will produce for the company over the next ten years, perhaps over the next five years. The minute they think that in ten years we won't need this new hire, or in five years we won't need this new hire they are looking at ways to not hire them. There is a training cost, and there is liability. So you don't take on those costs and liabilities unless you think there is a high likelihood of a payoff.
This is why we have been seeing layoffs in the human resources department. If you aren't planning on hiring you don't need so many people in that department.
To me what signaled the end of society as we know it was learning that Tesla robots can be trained by watching a Youtube video.
Watson winning Jeopardy demonstrated that AI has the general knowledge to be in the top 0.1% of human knowledge. AlphaGo demonstrated that AI has the ability to come up with new, creative and revolutionary strategies and concepts. Deep Mind's program which does weather forecasts demonstrates the superiority and cost savings of using AI to work on very difficult and complex tasks. Meanwhile tesla robots can replace blue collar workers.
-
1
- Show all
It Started: A.I Advances 1,000X Today - Millions Will Lose their Job
Consumer debt passes $17 trillion for the first time.
Car payment defaults hit 29 year high
First Jobs to go
Graphic designers, lawyers, teachers, proof readers, customer service reps, market research analysts, bookkeeping, coding, writing, telemarketer, receptionists, doctors, paralegal, accountant, trader, even surgeons.
Consumer debt passes $17 trillion for the first time.
Car payment defaults hit 29 year high
First Jobs to go
Graphic designers, lawyers, teachers, proof readers, customer service reps, market research analysts, bookkeeping, coding, writing, telemarketer, receptionists, doctors, paralegal, accountant, trader, even surgeons.
-
1
- Show all
OpenAI's Simulator STUNS the Entire Industry! UNREAL Physics Model, Emergent Abilities and AGI.
Sora is a physics simulation.
They are now training on a synthetic training set, meaning AI is generating the pictures and videos that they train the new AI on. All of the complaints about stealing others work and not giving them credit has been eliminated, no one will be able to make any claim against AI going forward!
This also means that we are no longer running out of data with which to train AI as we are now creating data from AI on which to further train new AI models.
What is totally scary about Sora is that they thought using Synthetic data would cause the models to amplify mistakes, instead we are learning that we were able to successfully train a much better model using synthetic data. This means we have a positive feedback loop. If Sora is exponentially better than previous model, what happens when we train out next generation AI model using Sora created images? Also, by making this open source and available to all they can easily create millions of new images and videos based on prompts from around the world which in turn will train the next level AI. Not only on the images but also on the prompts.
Wow! So the legal issue is that you can't use someone's copyrighted material to train on without paying them and giving them the appropriate credit. What Open AI has done is use AI to create a data set on which they train the next model and then they use the next model to create a data set on which to train the next model from that model. So there is no longer any issue with copyright as they own all the copyrighted material that they use to train these models. Also, since this is a "free" service and you are required to say that the AI created these images or data you avoid most copyright issues because no one made money on your work and no one didn't give you the credit.
Sora is a physics simulation.
They are now training on a synthetic training set, meaning AI is generating the pictures and videos that they train the new AI on. All of the complaints about stealing others work and not giving them credit has been eliminated, no one will be able to make any claim against AI going forward!
This also means that we are no longer running out of data with which to train AI as we are now creating data from AI on which to further train new AI models.
What is totally scary about Sora is that they thought using Synthetic data would cause the models to amplify mistakes, instead we are learning that we were able to successfully train a much better model using synthetic data. This means we have a positive feedback loop. If Sora is exponentially better than previous model, what happens when we train out next generation AI model using Sora created images? Also, by making this open source and available to all they can easily create millions of new images and videos based on prompts from around the world which in turn will train the next level AI. Not only on the images but also on the prompts.
Wow! So the legal issue is that you can't use someone's copyrighted material to train on without paying them and giving them the appropriate credit. What Open AI has done is use AI to create a data set on which they train the next model and then they use the next model to create a data set on which to train the next model from that model. So there is no longer any issue with copyright as they own all the copyrighted material that they use to train these models. Also, since this is a "free" service and you are required to say that the AI created these images or data you avoid most copyright issues because no one made money on your work and no one didn't give you the credit.
Now consider this, we have flight simulators for airplanes, we could make them for drones and other objects. Yes, we originally train the AI on the flight of these drones that were flown by people. But now that we have AI that can fly drones imagine we train their flight based on AI generated data and take that AI to the next level. Then suppose we have dog fights between two AI models. Imagine using a supercomputer we are able to create ten years worth of data in a week and train our next level model on that and continue with this positive feedback loop. If our first model was operating at about a score of 65% and then created the data in one week to train our second model which jumped to 75% score and a week later we are at the 85% score level and a month later we are at the 95% score level and a month later we are at the 99% score level. It is conceivable that a year after having an AI that can fly the drone we have an AI that can fly the drones superior to any pilot we have, just like having it teach itself GO and able to defeat the best Go player where only a few months earlier it was only a competent Go player. I think we can confidently assume the military has done this and is already at this level. Now suppose instead of training it in a dog fight against one drone they add surface to air missiles and other defenses. These drones could potentially become amazingly difficult to stop. Especially since G forces are not limited by a human pilot, and with a smaller object stealth technology is even better at hiding it from radar.
Also consider this, they can train the drone for special missions by having it practice on a computer simulation of the terrain. With each practice run it can become faster, more elusive, and more hidden from the defenses. A human pilot might practice 20 times on this course, but the AI could train 200 times or even 2,000 times.
Also consider this, they can train the drone for special missions by having it practice on a computer simulation of the terrain. With each practice run it can become faster, more elusive, and more hidden from the defenses. A human pilot might practice 20 times on this course, but the AI could train 200 times or even 2,000 times.