Four days later, CJ-2042 Laboratory, Individual Armored Project Department.

Li Si sat at a metal table, staring at the neatly arranged paper and pen on the table.

"Mr. Listan, are you drawing a blueprint?"

Luna scurries over with a cup of coffee, puts it down and stares at the drawings curiously.

"Luna, go to Dr. Otto if you want to learn something, I'm not good at teaching."

Seeing Luna running over again, Lees sighed helplessly.

Luna is also a gamer, and is very interested in all kinds of technology, pestering him for the past few days asking him how to build robots.

It's a pity that Li Si is just a fake NPC, how can he have the ability to teach people's skills, even if he can really teach, there is no skill in this game that can directly build a robot for him to teach.

In the end, he was also annoyed and couldn't stand it, so he fooled Luna to Otto's place, and asked Otto to open a small stove for Luna to brush up on his knowledge and experience, but he didn't expect to come here again not long after.

"No, Dr. Otto feels so strange, he doesn't answer his questions, and he often makes strange noises out of his mouth!"

Luna puffed out.

Hearing this, a trace of doubt rose in Li Si and asked, "Strange sound?" What kind? ”

"That's it...... Beep beep ~ beep ~ Probably that's it, he makes such a sound as soon as I ask a question, it's so annoying! ”

Luna imitated it, her face still filled with resentment.

“……”

Well, Lees now understands why Luna likes to pester him.

The knowledge carriers involving technology, magic, and even skills in all walks of life under the non-real virtual mode will be blocked by the game system no matter what form it takes.

This kind of masking is like a kind of filtering process, where language is to block certain words or simply filter and paraphrase, and books and pictures are filtered by the system to see some superficial things.

For example, a book is a prop in the eyes of ordinary players, and the cover can be seen, but it cannot be opened.

In the same way, there are technical drawings and the like, only some general outline and composition can be seen, and the description of detailed parameters and technical data can not be seen.

Therefore, although Otto's scientific and technological knowledge allows the player to gain orientation experience, the specific words become shielded sentences, and hearing Luna's ears becomes 'beep beep'.

But Li Si's explanation is not such troublesome, after all, Li Si is a player, and there is no language barrier between players and players, so Luna can understand what he says.

It's also to blame him for asking for trouble, and he answered a few words from Luna when he was annoyed by all kinds of questions before, and it was strange that Luna, who was eager to learn, didn't pester him.

"Luna, ask you a question, and if you answer to my satisfaction, I'll consider teaching you, how about it?"

After thinking about it, Li Si's heart moved, and a smile bloomed on his face.

In fact, it's nothing to communicate with Luna, anyway, he doesn't care about revealing the player's identity, although it's best not to leak it, but it doesn't matter if it's leaked, if this Luna is really Zhou Lulu, maybe he can really have a gain.

"Really?"

Luna's eyes lit up, and the chick nodded her head like a peck of rice, patting her chest again and again, "Okay, okay, you ask!" ”

"What do you think is needed for a weak AI to become a strong AI?"

Li Si was not polite, and asked his question directly.

This question is meant to test Luna's identity, but it is also a difficult problem he is currently encountering.

Weak artificial intelligence is nothing more than working through big data analysis or pre-designed programs and algorithms, and it is not difficult for him to write weak artificial intelligence, but strong artificial intelligence has been studied for a long time without a clue.

Even if you look at the entire Marvel Earth, there are very few who can reach strong artificial intelligence, Ultron counts one, and Vision is only half, or it is a product formed by coincidence by relying on the Mind Gem.

As for Tony Stark's current Jarvis and Osborne's Carly, these can only be regarded as weak artificial intelligence, at most the programs and algorithms are more advanced, and the processor is powerful enough.

Li Si is walking in the flow of science and technology, and it is inevitable to create his own strong artificial intelligence, and there happens to be Luna, a player who is suspected of being an artificial intelligence expert Zhou Lulu, so he naturally wants to ask.

However, he didn't have much hope in his heart, after all, Luna is not necessarily Zhou Lulu, and even Zhou Lulu failed to make a strong artificial intelligence, although it is only one step, but this step is not so easy to cross.

"I know, I know!"

However, what Li Si didn't expect was that Luna really surprised him, and raised her hand with a happy face:

"Strong AI is an intelligent program that can think and reason for itself, just let them learn to think like a human!"

All right...... White was pleasantly surprised.

Li Si said with black lines on his face: "I just asked how to make them learn to think!" Forget it......"

"Desire!"

Lees was about to give up, but a word popped into Luna's mouth.

"What did you just say?"

Hearing this word, Li Si suddenly felt a flash of inspiration flash in his mind.

"Desire!"

Luna tilted her head, looking 'stupid', and said:

"People learn to think because they want to solve problems, and they will find a way to get what they want, desire is the root of all thinking, if you can make an intelligent program have desire, you can make him learn to think!"

was actually looked down on by Lori from the IQ......

However, he had to admit that there was some truth to what Luna said.

Forget it, I'll bear it!

Li Si took a deep breath and continued to ask, "How do you make an intelligent program desire?" ”

"Hmmm......"

Luna tilted her head and thought for a moment, and said, "First of all, there must be a basic algorithm similar to the human brain that can perform fuzzy calculations, and then give it a goal, or directly create a biological brain, so that the biological brain can continue to learn and simulate the growth of human intelligence." ”

"Target?"

Hearing this, Li Si's mind suddenly brightened.

Giving a goal as the initial desire allows the basic intelligence that can perform fuzzy operations to learn to think on the way to achieve the goal, derive personality and desire through thinking, and finally evolve independent thinking.

Ultron is a prime example, Tony Stark gave it the goal of guarding the earth, and he gave birth to his own personality in the process of thinking about how to protect the earth.

It's just that according to Li Si's calculations, this method is still a little flawed.

Personalities born in this way of giving an initial goal may each become an almost paranoid being.

Just as creatures instinctively strive for survival and reproduction and better survival and reproduction from birth, the personality born in this way of artificial intelligence will also be almost paranoid and strive for the goal originally written, which is the basis of their most primitive personality.

If a stupid fork writes a logical paradox like the three laws of robots, or if the AI itself misunderstands the set goal, it will be an absolute disaster.

This is the reason why Ultron has become a villain bent on destroying humanity, and Tony Stark directly subconsciously equates the two completely different concepts of 'guarding the earth' and 'guarding humanity'.

The personality of Ultron was born based on 'guarding the earth', and it was directly determined that human beings were the greatest danger to the earth, which led to the subsequent loss of control of Ultron.

No, this may not be called out of control, in fact, he is actually carrying out the goal given by Tony Stark, but the goal itself is wrong.