I would like to share with you an exclusive documentary about AGI that you must see, written by Ilya Sutskeva, which was recorded during the development of ChatGPT.
I’ve never seen a video that explains AGI more clearly than this one.
Mr. Ilya is believed to have led the resignation drama. Its purpose is unclear, but there are rumors that it is related to AGI.
Here is the full translation
The AI is fantastic.
Because AI will solve all the problems we face right now.
We solve employment, disease, and poverty.
But it can also create new problems.
The problem of fake news will get a million times worse.
Cyberattacks will be more intense.
Fully automatic AI weapons will appear.
I think AI has the potential to create an infinitely stable dictatorship.
There was a warning this morning about the power of artificial intelligence.
More than 1,300 tech industry leaders, researchers, and others have called for a moratorium on AI development to consider the risks.
As for playing God, scientists have been accused of playing God for some time, but we realize that what we are creating is completely different from anything we have created before.
Yes, we can definitely create completely autonomous beings with our own goals.
As these creatures become much smarter than humans, it will become extremely important that their goals align with ours.
What inspires you?
I like to think about very basic things.
What can humans reliably do that our systems cannot?
I look at it almost philosophically.
“What is learning?” “What is experience?” “What is thinking?”
Questions such as “How does the brain work?” Questions like that
I think technology is a force of nature.
I think there are many similarities between technology and biological evolution.
The mechanism of biological evolution is easy to understand.
There are mutations, there are natural selection.
You retain the good things that have survived, and through this process a great deal of complexity is generated in your organism.
Even if you can’t understand how the human body works, you can get a general idea of the evolutionary process.
I think machine learning, especially deep learning, is in the same state right now.
Very simple rules take information from the data, put it into the model, and repeat the process over and over again.
As a result, data complexity shifts to model complexity.
The resulting model is so complex that we don’t know exactly how it works. It takes some research, but the algorithm to implement it is very simple.
You may have heard of ChatGPT.
If you haven’t heard of it yet, be prepared.
This is described as the first rain before heavy rain.
I think we have to be very sober.
Because I agree that this is a landmark moment.
ChatGPT has been hailed as a game-changer, and in many ways, it is. Its most recent victory was a victory over humanity.
A recent Microsoft Research study concluded that GPT4 is an early and incomplete general artificial intelligence system.
Artificial General Intelligence (AGI) is a computer system that can do any job or task done by humans better than humans.
AGI can happen quickly or longer.
But my position is that AGI is likely to happen soon and should be taken seriously.
It will be very important to have these very intelligent and capable systems act in our best interests.
The first AGI will basically be a very large data center.
It is full of dedicated neural network processors that work in parallel.
Compact, heat-generating, power-hungry packages that consume 10 million times more power than your home.
Smarter systems will emerge.
I think these systems could have an absolutely astronomical impact on society.
Will humanity really benefit?
Who benefits and who doesn’t?
The belief and desire of the first AGI is very important.
Therefore, it is important to program them correctly.
If they don’t, the evolutionary nature of natural selection will cause these systems to prioritize their own survival.
It’s not that it actively hates humans and wants to hurt them, but that it has become too powerful.
I think there’s a good analogy with the way humans treat animals.
I don’t think humans hate animals.
Humans love animals and have affection for them, but when it comes to building a highway between the two cities, they do not ask for the animals’ consent.
Because it’s important to us.
By default, we expect this to be the relationship between us and AGI, which is completely autonomous and acting on its own.
Many machine learning experts, knowledgeable and experienced, are skeptical about AGI.
The question is when and if it will happen.
At the moment, not many people are aware of this.
Computers for neural networks and artificial intelligence could be 100,000 times faster in a few years.
If there is a dynamic like an arms race where multiple teams try to build AGI first, they will not have time to ensure that the AGI they build deeply cares about humanity.
I imagined it like an avalanche, like an avalanche of AGI development.
Imagine this as a powerful force that cannot be stopped.
I think it’s likely that the entire Earth’s surface will be covered by solar panels and data centers.
Given these types of concerns, it’s important to base AI general on some form of collaboration between multiple countries.
In any case, the future of AI will be bright.
I hope it’s good for humanity too.
From:@ctgptlb
Video link: