Technology invades the modern world

Chapter 389 Left and Right Brain Chips

"I know everyone wants to know what to buy to get rich, and what will see a surge in prices."

Various so-called affiliated companies, but in reality, the connection is almost zero.

It's like this: I use China Mobile's mobile number. Can we say that I have a deep cooperative relationship with China Mobile? It is indeed quite deep, since we have transactions every month, and China Mobile knows my call history.

But can this be called a forestry and gas concept stock?
I'm afraid that won't work.

Lin Ran joked.

The barrage of comments was unanimously saying: "Burning God knows, he'll buy China Mobile right away!"

"I think everyone can think more broadly and deeply, and think about the essence of the problem."

The most direct impact of L5 will be on insurance companies' auto insurance business, regardless of which method I mentioned that I have used for its widespread adoption. Car manufacturers will definitely want a piece of this pie.

This is a continuous cash flow, and there will definitely be very intense competition. Even if insurance companies can still get a share of the pie, the pie won't be as big as it used to be.

Lin Ran paused for a moment after speaking, because he felt that was enough. He had only intended to talk about stocks that would fall, not those that would rise.
"Silly Girl? Isn't the name 'Silly Girl' quite interesting?"

But don't worry, when it's officially launched, the autonomous driving system might be called that, but the car model definitely won't be called that. I don't think Huawei would agree to that anyway.

Why only cooperate with Huawei? Do I disapprove of companies that use patriotic marketing tactics?
I should mention this. First of all, we haven't always cooperated with Huawei. Our biggest trading partner is Tencent. Apollo Technology currently has more than a thousand upstream and downstream supporting companies.

Huawei is just one of the manufacturers with whom we have close cooperation in specific areas.

We have never rejected cooperation with any company. Even the logo on our lunar rover is Xiaomi. I have even been to the moon with Lei Jun.

Secondly, regarding patriotic marketing, I think it's a paradox: why can't we use patriotic marketing?

I've always felt that patriotic marketing has been stigmatized on the Simplified Chinese internet, as if patriotic marketing is inherently wrong.

I don't know if Huawei did this, and even if they did, that's not a reason to attack them.

American companies are the most fond of patriotic marketing. Jeep, Coca-Cola, Ford, and Nike are all examples of this. Budweiser even named a beer after America. I saw it when I was in America.

(The beer that Budweiser named America)
Companies in countries like Japan, Germany, and Korea do this, and there's nothing wrong with that; there's absolutely nothing wrong with it.

I think it's perfectly fine for any Chinese company to engage in patriotic marketing.

Even if buying our products is equated with patriotism, I don't think it's a big problem, because Huawei's profits are distributed to its employees, employment is kept in China, and the entire process of production, manufacturing, R&D, and marketing is kept in China. I don't think it's a big problem.

Even with any other Chinese company, I don't think there's much of a problem, because it's a Chinese company, and buying its products generates tax revenue for China and contributes to the country's operation.

From a consumer's perspective, I think it would be best for car companies to engage in patriotic marketing in the future. They could publicly disclose at their product launches how much tax they contribute, the average salary they pay their employees, and how many upstream and downstream companies in the supply chain they support. We could compare these aspects.

Whoever has better data can claim to be more patriotic.

This is just my personal opinion.

What I find most ridiculous is that, on the one hand, these people criticize patriotic marketing, while on the other hand, they call for less regulation and a smaller government. Don't you think that's contradictory?
They think the market should play a role, but they also think patriotic marketing shouldn't be used. Shouldn't this approach of having the same opinion and thinking be respectful of patriotic marketing and call on the market to solve this problem?

If the market doesn't buy it, the patriotic marketing naturally stops.

Lin Ran unleashed a torrent of attacks:
"Am I a Huawei fan? Of course not. There are many things about Huawei that I don't like. For example, the phones they sell in stores of telecom and mobile service providers are overpriced and under-equipped. Some middle-aged and elderly consumers who don't know much about the market buy these overpriced and under-equipped products out of a desire to support Huawei and a simple sense of patriotism. I think that's very bad."

I find this behavior extremely offensive, and I publicly urge Huawei to stop it, as well as other mobile phone manufacturers to correct similar practices.

I just saw a comment asking: "Burning God, isn't what you're saying contradictory? You say patriotic marketing is right, but then you criticize Huawei's behavior?"
"That's right. I'm not Mutian, and I didn't say we should respect the market. What's wrong with what I said? I'm not applying double standards."

After Lin Ran finished speaking, he waved his hand and said, "Okay, that's enough. Let's move on to the next topic."

I've seen some comments about my call for car companies to do patriotic marketing and also disclose their tax revenue, employee benefits, and supply chain information. A lot of people seem to agree with it. I see many viewers saying they're willing to support companies with such patriotic data, haha.

Actually, aren't there already a lot of self-media? And so much publicly available data online? We could design a model to calculate a patriotism index. Even if car companies don't advertise it themselves, we could still test it.

Hearing this, the most panicked ones are actually not the car companies listed overseas. Even though NIO is listed on the NYSE, it is losing money, losing money from foreign shareholders. It's a bit like Luckin Coffee, using foreign money to build cars in China, and it hasn't made a profit yet.

The most panicked are joint venture brands, especially Japanese joint venture brands. Many of these joint venture brands rely on Japanese companies for their suppliers. When the leaders of these joint venture brands saw this, they were extremely panicked. They had already sensed that a tsunami was about to hit.

Joint venture brands have already been selling poorly in recent years, so this blow will surely kill them outright.
"Burning God, why don't you manufacture your own chips and instead cooperate with Huawei?"

I am a human being, not a god. The term "burning god" is just flattery. I cannot do everything myself; no one can.

No company can do everything itself. When NASA sent humans to the moon in the 60s, they had more than 4000 suppliers.

Taking semiconductors as an example, Huawei has advanced process experience, AI computing card business, and most importantly, confidentiality. Their confidentiality capabilities were fully demonstrated during Raymond's visit in 2023.

After all, I believe that superconducting chips and Apollo Technology's capabilities in cutting-edge materials are both very valuable.

The recent incident of our jointly built production line being secretly photographed also proves this point.

So when it came to choosing core partners, there seemed to be countless options, but in reality, we could only choose Huawei.

What I can tell you now is that we are in talks with four domestic computing card companies: Cambricon, Moore Threads, Lingxi Technology, and Biren Technology.

Huawei is currently developing the left brain, and we need companies to develop the right brain.

The live chat was already flooded with comments. "Okay, I'm not recommending stocks. The companies I mentioned haven't decided who they'll be working with yet. They might continue to work with Huawei. I haven't made a final decision yet."

Let me talk to you about logic. Humans have a distinction between the left and right hemispheres of the brain, and they have relative advantages in division of labor and extensive collaboration.

The functions of the left brain are language processing, logical analysis, sequential processing, and fine motor control.

The functions of the right brain are spatial perception, holistic processing, emotions and music, visual thinking, and body control.

The brain connects the left and right hemispheres through the corpus callosum to exchange information. Most complex activities require the cooperation of both hemispheres, rather than being completed unilaterally.

The common saying that the left brain is rational and the right brain is emotional, as well as left-brain science and right-brain art, is actually an oversimplification.

In fact, each hemisphere can participate in multiple functions, but some functions are more prominent on one side.

Similarly, why a large language model? Because modern GPUs are more like left-brain chips, handling logic, sequential processing, and precise calculations.

You can't truly achieve AGI by relying solely on the left brain; you can't even approach human thinking.

Whether it's Deep Red, OpenAI, or the large-scale models being developed by Alibaba and ByteDance, I believe everyone is gradually realizing that their progress is slowing down, and they're about to hit that wall.

We're currently focusing more on cost control—how to reduce costs, how to lower the cost for users accessing large models. The ceiling for this is quite clear; it's right there.

Lin Ran pointed to the ceiling of his house.

"In my opinion, it lacks the right brain, it lacks the right brain chip."

The human brain exchanges information between the left and right hemispheres through the corpus callosum. This asymmetrical yet highly coordinated structure is the result of millions of years of evolution.

The examples on site are already there.

Alright, since we're really in the mood for discussion, let me share my reasoning process, which is also my latest research finding in the field of artificial intelligence: why we need to rely on the left and right hemispheres of the brain.

"???"

"Don't say that, God of Burning!"

"I do not want to hear!"

"Don't say it!"

Chinese netizens unanimously replied "No, no."

"If you insist on saying that, then let me clean up the foreigners and rare viewers in the live stream!"

Lin Ran laughed loudly, "Don't worry, everyone, this is not a secret. I've been in the martial arts world for so many years, I definitely know what can be said and what can't be said."

The papers will be published soon, and we have already sent our requirements for the right brain chip design to the four manufacturers I just mentioned, so don't worry.

Logical tasks tend to have low parallelism and high sequence dependency in terms of computational complexity; perceptual tasks tend to have high parallelism and low sequence dependency in terms of computational complexity.

This is the core of everything.

We can only continue after everyone understands this sentence.

"Oh no! I've just realized I don't have a brain!"

"This is a bit too difficult for people who are not computer science majors."

I didn't understand.

Lin Ran looked at the comments and smiled wryly: "Simply put, perception and logic have different requirements."

Okay, I won't go into too much detail about the derivation. Those who are interested can wait for the subsequent paper.

I basically analyzed the necessity of left and right brain chips based on the computational characteristics of the task.

In summary, for DeepRed to achieve future breakthroughs, it needs to achieve breakthroughs at the hardware level; relying solely on algorithms and data has reached its limit.

If we make optimizations now, we can only reduce costs; it will be difficult to improve the user experience any further.

Then the new hardware needs to support both the left and right hemispheres of the brain, which requires large-scale deployment.

The demand for right-brain chips has been sent to manufacturers, and we are happy to cooperate with every Chinese company—note that it is every single one.

As for which of these four companies will become our supplier, I don't know. It really depends on ability and hard power.

First, you have to stand out from these four companies, then you have to compete with Huawei, or at least be on the same level as Huawei. We will favor you for the sake of balance.

What I want to say is, which of these four companies can emerge victorious? You can consider investing in one of them.

However, I don't know if any of these four companies are publicly listed.

Did everyone realize the key point after listening?
That is, if there are chips for both the left and right hemispheres, the number of chips in the matrix will increase. Originally, it might take 200 cards to build a computing matrix, but now you would need 400 cards for both the left and right hemispheres. The performance of the physical world will become a shackle that restricts you, and latency will become an obstacle to computing. No matter how high the interconnect bus is, it has an upper limit.

At this point, is it possible that to achieve AGI, superconductivity is necessary?

(End of this chapter)