I Became a Genius Engineer - I Became a Genius Engineer chapter 104
Humanoid Developer Cheolwoo Lee (5)
I thought hard about what the biggest difference between humans and Ark’s intelligence was.
Again, inflexibility and inability to work efficiently.
And as I dug deeper, I tried to think about why I was inflexible and couldn’t work efficiently.
As a result.
“Ark has no stimulation.”
Stimulus, humans are stimulated when they do something. It’s fun when you enjoy a hobby that you enjoy. If something sad happens, it’s sad. If you work, you are rewarded.
emotions and stimuli.
Emotions were something that could not be planted in Ark right now. But stimulation is different.
All you have to do is to create a stimulus by sending an electrical signal.
“Are you talking about provocation?”
“Yes. There is no stimulation. The Ark simply moves according to commands and does not feel any stimulation. Just like a machine.”
“The machine is right, but…”
From noble mtl dot com
In the heavy atmosphere, Researcher Song took a stand.
“That’s right… but do you really have to walk away?”
“hahahaha, I think the atmosphere is too heavy. It’s a joke.”
“Well, anyway, we keep talking about it, but people get stimulated from time to time. But Ark doesn’t have that. And we need basic instincts to create that stimulation.”
Ark and people are different. Ark only listens to people, understands them, and practices them.
The natural things that humans have, such as emotions, instincts, and stimulation, are lacking in Ark, so he had to fill them up one by one.
“Hmm… do robots need something like that?”
said Dr. Nam.
“Dr. Nam seems to know, but do you know Maslow’s hierarchy of needs?”
“Are you referring to the five human needs?”
“yes.”
“Starting with physiological needs, safety needs, social needs, esteem needs, and self-actualization needs. These are the five stages.”
“Yes. We have to make Ark that kind of desire.”
“Um…. Do Ark need respect or self-realization?”
“hahahaha, of course there must be a desire different from that of a human being. We need that model. Um… Rather, we should create a desire similar to that of a dog.”
Ark is a robot that will work in place of humans.
It wasn’t about creating a new human being. In order to do that, it would be better for Ark to resemble the needs of a puppy.
“……? A puppy?”
“Yes. How about a puppy? He tries to be liked by the dog owner and acts cute to eat one more snack.”
“Ark is a puppy…”
Researcher Song felt frustrated.
“Why is that?”
“No, it’s such a cool humanoid arc, but imagine it acting like a puppy. How cool is that.”
“hahahahaha, that’s right.”
“Well, romance is romance and reality is reality. There’s no choice. That’s why we have to create an arc desire model and even develop it, right?”
“Yes. That’s right. And in the process, various emotions will be added.”
“What do you think will be added?”
“Two things I said at the beginning. Carrot and stick.”
“What is it?”
“Happiness and pain. First we have to add those two.”
If the dog’s needs are applied to Ark, Ark will feel happy when he does well with human commands.
Then, in order to feel that happiness, you will follow orders better, and you will ‘think’ yourself how to better follow human orders.
Because you can still think on your own.
And conversely, if the owner scolds them, they will suffer. Then, in order not to suffer from that pain, Ark will not do anything that the owner hates.
It was carrot and stick.
“It’s difficult.”
“You didn’t think it was easy from the beginning, right? hahahahaha.”
“Still, it’s amazing. Ark can already move on its own. It must have learned how to move.”
“Yes. However, stimulation is needed to make the learning effect more pronounced.”
“Yes. That’s right. Ark will think for himself in order to move more efficiently.”
“Huh, there are more important issues than that.”
“A significant matter?”
“Yes. The delay of 1 minute, I’ll have to solve that first.”
I just gave an order to the arc, but there was a delay of 1 minute, so it was very frustrating.
I ordered an antenna from NF Network, the so-called NF Network Amplifier, but I haven’t heard from you yet.
I guess I’ll have to visit it sometime.
“Yes. That’s a mountain to climb.”
“Please do various tests on Ark. I think I will have to go to the NF Foundation.”
“Yes. I understand. I will test it in various ways.”
“yes.”
Creating a humanoid was a lot of thinking.
* * *
China was thinking about how to proceed with the development of the universal intelligence and strong artificial intelligence that China was trying to create, and eventually decided to outsource to Tensort.
Because it was judged that private companies had better technology than national research institutes.
And to see the progress, Minister Wang from the Ministry of Industry and Information Technology visited Tensot.
“Hello, Minister Wang. Long time no see. How have you been?”
Tensort is the number one IT company in China by market capitalization. The technical director there, Xiao Huan-ze, greeted Minister Wang.
“Yes. Hey, I’m doing well.”
“Thank you. Then I’ll report right away.”
“yes.”
Xiao Huan-ze started his presentation right away.
“First of all, heaven and earth are being created with universal intelligence and strong artificial intelligence. They can think for themselves and judge for themselves. And…”
“Um… let’s be honest.”
Xiaohuandi was about to continue his presentation, but Minister Wang interrupted him.
“yes?”
“Can you make that universal intelligence? Strong artificial intelligence. Is it possible to have artificial intelligence that has the same way of thinking as humans even if it is not universal intelligence?”
Xiao Huan Di broke into a cold sweat at Minister Wang’s key question.
‘How am I supposed to answer this?’
And he thought of how to answer.
In fact, the development of Cheonji was difficult. This is because the existing deep learning algorithms showed limitations.
So, the innermost thought was, ‘Let’s just pretend to be developing and consume only development costs.’
There must be some penalty, but it was too much money for that. It would be over 180 trillion won in Hanwha.
“Didn’t you just pretend to be developing and eat the development costs?”
Minister Wang said as if looking into his mind.
“…… No. How can we eat persimmons only on development costs?”
“So you mean you can make it?
he threatened
Receiving the party’s wrath in China meant that in China, you couldn’t even get your feet wet. I was lucky if I couldn’t put my feet on it.
If you do it wrong, your neck will fly off.
Xiaohuandi was at a crossroads of choice.
‘… … . Can I develop it?’
It didn’t matter if it was developed. But what if you can’t develop it? I had a hunch that things were getting bigger.
180 trillion won.
Its weight was enormous.
‘no… … . No matter how much money we have, we don’t have the skills to make it.’
And he made up his mind.
“I’m sorry. In fact, I’m having trouble right now. No matter how much money you have, you can’t come out with a genius like Lee Cheol-woo.”
And he spoke frankly.
Creating strong artificial intelligence was a problem that simply could not be solved with money.
“……Why is the story of Lee Cheol-woo being brought up here?”
“Sorry. I couldn’t think of anyone.”
And, in fact, Prince Wang knew that too. Because they knew that they couldn’t make strong artificial intelligence even if their technology, no, Professor Jeffrey Hinton, who is called the father of artificial intelligence, came.
“Ha, too.”
“……. sorry.”
“No. I knew. That’s why I came.”
“Is there a way?”
He confirmed again that no one was in the conference room and opened his mouth.
“We have already invested heavily, and there are eyes from the US and South Korea. If we don’t show it at this opportunity, China will tarnish our image as a great country.”
“…”
“So, make it. Somehow.”
“… even if you say so.”
“Is it called artificial intelligence? Inchen, an artificial intelligence development tool, and other artificial intelligences have followed us a lot.”
“Yes. I know that.”
In China, they really spent a lot of money, and the weak artificial intelligence technology using deep learning, a technology that is already out there, has increased a lot.
That was all.
“You can’t be second place forever. So please show it this time.”
“Even if you say so…”
Xiaohuanji wondered why he told me to be honest if this was going to happen.
“Don’t you understand what I’m talking about yet?”
“…?”
When Minister Wang said that, Xiaohuandi had a clue.
“Prepare properly. So that everyone is surprised.”
“What does that mean?”
“I think you’ll understand even if I don’t say more.”
At those words, Xiaohuanji was convinced.
‘Ah, it’s a lie. That too properly.’
He originally planned to do that. ‘Pretend to be developing’ is only a matter of tricking it when announcing. But the King’s words… … .
‘My opponent is not the government, the United States, Korea and other powers.’
He could read the King’s true thoughts.
The government would turn a blind eye to it, so it was to deceive them properly. astonishing the world.
“I understand. I will prepare properly like that.”
“Yes. I think you understand enough.”
* * *
I visited the NF Foundation to check how much the development of the antenna, the so-called NF network amplifier, is progressing.
“Chairman! Hello!”
“hahahaha, you can call me comfortably when we’re together. Sangcheol hyung”
“hahahahaha…. How awkward it was to keep being the chairman of the foundation. Is that so?”
“Do it.”
“So, without saying anything, what are you doing today?”
“The NF network amplifier I ordered last time, I came to see how far it has progressed.”
“… uh, um… uh.”
Judging from the reaction, it seems that it is not being developed properly.
“What’s the problem?”
“Is there a problem? The problem is that it’s just too difficult. The network amplifier you mentioned makes communication faster in NF networks all over the world, right? How is that possible? After all, communication technology needs to be developed.”
“Um…”
It’s definitely a problem.
Critique learning theory is a form in which each shard network becomes one synapse. After all, it has to go through all networks, whether they have a lot of computing power or not.
Of course, Ark’s intelligence is not widely used to that extent, but it still has to go through all shards in the country.
“Why don’t you make a base and put a lot of NF network on it?”
“Well, I haven’t thought of that way.”
“Is there a problem?”
“Yes. The x2xx algorithm, which is an algorithm of review learning theory, simply does not work because there are many shard networks.”
Going into a little more detail is too complicated, so I omitted the details.
“then?”
“The point is to keep the distance?”
“Hmm…? What is the correlation between distance and artificial intelligence…?”
Sang-cheol hyung said as if he couldn’t understand.
“So, isn’t the way the theory of critical learning works is to create neurons and synapses? In the process, you have to be physically distant so that quantum interference can occur and you can think critically.”
“……? Quantum interference?”
“Yes. There is a thesis that I have read while researching, and it was a thesis that the human brain can be explained by quantum mechanics. The theory of critical thinking was created based on that thesis. Quantum interference must occur.”
From noble mtl dot com
“What quantum mechanics did you get into?”
“Somehow.”
“Crazy, a real genius must be a genius.”
“Anyway, so that’s why NF network amplifiers weren’t made at all?”
“…”
Well, I expected. I just ordered it, but I didn’t tell you how to make it specifically.
How to make an amplifier… … .
No, should I change my mind a little? Not an NF network amplifier. Isn’t it enough to make quantum interference happen even at a short distance?
“I need to make a quantum antenna.”