Spark

After arranging Wang Jiancheng's itinerary, Chen Nian also completed the preliminary arrangement of general artificial intelligence technology.

In fact, unlike many people’s first impressions, although general artificial intelligence is conceptually easier to be accepted by ordinary people and has a broader romantic imagination in terms of prospects, in fact, the value it can exert is absolutely It's not just a simple humanoid robot.

Chen Nian has always felt that the description of general artificial intelligence in countless film and television works in the past is not so accurate.

Whether it is I, Robot, which made the Three Laws of Robots popular, or Skynet in Terminator, what can be reflected is actually only one aspect of general artificial intelligence.

Will artificial intelligence be a humanoid machine?

Will artificial intelligence be able to obtain the supreme data authority?

Maybe not.

Theoretically speaking, due to the perception of danger, human beings will limit general artificial intelligence to the level of tools, and its form may be closer to smart assistants, that is, vassals of a certain biological person .

And what Chen Nian intends to do is this direction.

Or in other words, he does not intend to create a new person as a creator god, but to make a fuss about the keyword universal to create a class that can adapt to most working environments and expand in specific fields. Self-learning smart machines.

Expressed in more academic language, it is weak general artificial intelligence.

Compared with GAI in the absolute sense, the difficulty of this idea will be slightly lower, but it is not so low.

After all, even in the last life, the development of artificial intelligence was completely a blind box. It was completely different from nuclear fusion technology, and basically there was no fixed and clear path. Instead, it was more similar to the exploration of room temperature superconductivity at that time.

A group of people explored in the dark along what they thought was the right direction, and they might occasionally touch a few gems in the dark, but no one dared to say whether the road to touch the gems directly led to the mine vein.

What do we need is a bigger model? more data? A more ingenious trick? More reasonable hyperparameters?

Maybe not, these things do not make artificial intelligence more powerful, on the contrary, it only makes it more bloated.

If you want to truly solve all the problems faced by GAI, you must first solve the differences in the underlying principles between the human brain and computers.

quantum computer.

This is the closest form of computer composition to how the human brain works.

Traditional computers only deal with 0 and 1, and the way they process data is linear and one-dimensional.

Even a multi-core computer is nothing more than performing multiple linear calculations at the same time.

But the way the human brain handles it is completely different. Under the action of enzymes, neurons produce many different changes and results in response to chemical stimuli.

There are as many different states and responses as there are different enzymes and chemical stimuli.

Moreover, the human brain can have hundreds of billions of neurons functioning at the same time per second, and can complete multiple tasks at the same time. For example, we often drink water and watch movies while thinking, which is impossible for computers.

To express it concretely, the computer intelligence takes a sip of water, watches a movie, thinks about a question, and then through ultra-high-frequency calculations, shortens the time of drinking water and watching movies to the extreme, creating the illusion of synchronization.

Only in the quantum state, verifying the working methods of all possibilities at the same time can best reproduce the underlying logic of the human brain.

——

However, under the premise that the equipment changes, the traditional deep learning theory seems to have to be adjusted accordingly?

That was Wang Jiancheng's business.

I am only responsible for creating the foundation.

Thinking of this, Chen Nian opened the system panel and checked the analysis source requirements of the quantum computer.

As he expected, in this field where the accumulation is not too deep, the source of analysis is extremely demanding.

The source point consumption of a general-purpose quantum computer is directly a series of question marks. Even if it is a special-purpose quantum computer, under the premise of the requirement for quantum volume, the source point consumption has basically reached more than four digits.

——

The so-called quantum volume represents the number of qubits that can successfully interact and produce correct results with a high probability under the joint action of several indicators such as the number of qubits, qubit lifetime, high-fidelity operation, and high connectivity.

According to Chen Nian's experience, when the quantum volume is below 100, there is basically no analytical value.

After all, in 2010, the University of Science and Technology of China had already achieved four-qubit entanglement. If it wants to reach 100, it just needs to maintain the current technical level and continue to pile up hardware.

Not worth it.

Just one computer needs thousands of source points, and this has not considered the operating system that relies on quantum algorithms.

Thinking of this, Chen Nian reluctantly exited the system interface, picked up the phone on the desk again, and called Chen Guo.

I might have a problem here.

Chen Nian said bluntly.

what is the problem?

Chen Guo, who was on the opposite side, asked in a slightly puzzled tone.

Well, it should be said that I overestimated the development level of domestic quantum computing.

Under the current circumstances, it is basically impossible or uneconomical to create a basis for the realization of GAI.

So, I need to invest more funds and resources in this area in China, and it is best to pull out a special project team.

At the same time, considering our current talent accumulation, in order to solve the short-term efficiency problem, I suggest that the project should be made into a form of international cooperation like the EAST project.

Everyone who can dig it, let's dig it - even this project, we can let them enter the core layer.

Our purpose is to lay the foundation as quickly as possible.

Chen Nian's strategy was carefully considered. After all, according to his memory, even in his last life, until around 2020, there was no breakthrough in the field of quantum computing.

But talent is increasing day by day.

Incorporating all these people into one's own team can solve a technical availability problem, thereby minimizing the source requirements for analysis in the system.

In essence, it's still for wool.

Hearing his words, Chen Guo was silent for a moment, and there was a buzzing sound from the receiver, and then he answered:

Understood, I will arrange this.

The progress of the project approval, I will follow up with you in real time.

Now the strongest talent is on Laomei's side, we may still have to offer something in exchange - it depends on the decision of the superior, you don't have to worry about it.

clear.

Chen Nian replied briefly.

worry?

He doesn't worry.

In fact, he is very clear that whether it is nuclear fusion or room temperature superconducting technology, it cannot be hidden in the end.

The most that one's own side can maintain is a few years, or even only two or three years of first-mover advantage.

Taking out a little thing in exchange will not really affect the overall situation.

But if you can get something good in exchange, it's a bargain.

A few days later, in the laboratory of the American Quantum Computing Project, AQCL group.

Davis rubbed his eyebrows tiredly, looking at the thick stack of report documents in front of him with a melancholy expression.

In the past two months, he has watched these files slowly grow from a thin layer at the beginning to a pile higher than his computer screen.

He wanted to review and deal with them one by one-in fact, he did exactly that in the early days of the project.

However, with the increase in capital investment and the explosive emergence of results, he has also fallen into the predicament of lack of energy.

In the beginning, I couldn't finish reading the report and had to work overtime.

Gradually, overtime work could not make up for the natural gap in time, and the reports piled up.

Up to now, he has completely given up on the original way of working, and will only pick and choose relatively critical progress.

This will undoubtedly cause him and the project team to miss many important details and potential directions, but so what?

I really can't do it myself.

Not only him, but the entire project team has been under extremely high pressure. He can even feel that the carrying capacity of the AQCL team has been squeezed to the limit. It may be just a small disturbance that will cause a comprehensive collapse.

But no one dared to stop.

Because the sponsor of this project is the military.

They are notoriously utilitarian, notoriously strong, and notoriously unscrupulous.

Who knows if they'll actually pull out their guns if the project is delayed?

No one dared to bet.

Davis sighed deeply, then, he looked at the assistant on the side, and said:

Our next plan is to keep the focus of our work on optimizing the quantum algorithm as much as possible. You need to make this clear to everyone.

While stacking hardware devices may allow us to make good progress in the short term, at the end of the day, the upper limit of the hardware is there.

We can expand to 16qubit, 30qubit, or even hundreds of bits, but the limitations of the algorithm will inevitably make us stagnate.

At that time, if the performance of the result fails to meet expectations, there will still be a big accident.

I see.

The assistant nodded firmly, but immediately, his expression became a little ugly again.

He opened his mouth and said:

We may need to make some resource allocation adjustments in the short term, and that's bound to cause a sharp backlash.

As you know, many people in the project team are actually opportunists. They don't care about real 'progress', but only want to please the military.

But their methods of currying favor with the military are too low-level.

Davis interrupted the assistant.

Only making effective results is the best 'please'.

Don't worry too much, I will take the pressure off.

Seeing Davis' expression, the assistant couldn't help feeling a little sad.

The man had given enough—but who could understand?

He sighed secretly, picked up the progress report on the desktop that Davis had reviewed, turned around and walked out the door.

But at the moment when he opened the door, a huge thrust suddenly came from the door handle.

He stumbled back and fell to the ground, and two tall figures appeared at the door.

Davis?

We are DHS.

Come with us.

DHS interrogation room.

Opposite to Davis sat two senior detectives, with a bright incandescent lamp unique to the interrogation room overhead.

From the time the door was broken into to when he came to this interrogation room, the whole process did not exceed two hours.

Obviously, the other party came prepared, and the purpose was him, the person in charge of the AQCL project.

When he was blindfolded and taken into the car, countless terrifying thoughts flooded his mind. For a moment, he even wanted to kick the door and jump out of the car to escape like the agents in those movies.

But the pistol on his waist made him take back the idea of ​​taking risks. He could only comfort himself that this was just a routine inspection.

After all, I really haven't done anything that affects national security, and I have strictly enforced the secrecy regulations.

No matter how you check, you should be invincible.

However, as the other party's interrogation deepened, Davis felt that the matter became more and more serious.

.Are you sure that you went home immediately after leaving the laboratory on the night of the 15th of last month, and that you had no other activities on the way?

One of the detectives asked indifferently.

I'm pretty sure my car broke down twice in the car park that night and my co-worker Richard fixed it for me, so I'm pretty impressed.

Davis replied firmly.

But I have a record of your drive-thru consumption at McDonald's. Did you buy anything there before you left the lab and went home?

.Should have bought it, but I don't remember it clearly.

Why is it unclear? Your wife was on a business trip at the time, and no one will prepare dinner for you. Buying dinner at the drive-thru should be a routine activity. Why did you deliberately hide it?

I didn't hide it on purpose, I just didn't think it was an important thing.

Davis frowned. Judging from the level of detail in the inquiry, something big must have happened.

Okay, now you know, this is an important thing.

Now, I need you to remember well, who was the waiter who served you that night?

A black person, I don't know - I'm more familiar with the red-haired waitress, but she wasn't there that day.

You remember it very clearly, why did you say you couldn't remember it just now?

I'll say it again, I don't think this is an important thing, I just remembered more details after you reminded me!

Davis' tone became a little angry. He was really fed up with this kind of interrogation without respect.

Before the other party answered, he continued to question:

Now, stop your tricks.

I want to know, what happened to make you so vigilant?!

Tell me the answer!

Only when I know the answer, can I help you better.

This is the best solution for both of us!

Looking at Davis, who was already a little hysterical, the two agents looked at each other.

After a moment of silence, one of them said:

Well. You do have that right.

The answer is: Huaxia has launched a quantum computing project, and it is expected that it will openly recruit talents from all over the world in the short term.

We seriously suspect that the AQCL project has been leaked, which has led them to rely on their superior resources and compete with us.

Huaxia, quantum computing project?

How can this be

Shouldn't their attention be all on derivative technologies of superconductivity and nuclear fusion? ?

No wonder.

No wonder DHS is acting so sensitive.

This is indeed a serious problem.

——

No, leaking secrets is not the most serious situation.

In case, it's really just a coincidence.

According to the projects that China has done before, doesn't this mean that they are already sure of winning in this field? ?

The huge efforts and huge resources invested in the AQCL project

Does it really make sense?