Reborn in 2010, the AI ​​in my mind is too powerful.

Huang Shengbin's phone call came at such an unusually convenient time.

After all, Chen Yiran and Huang Shengbin had only met twice at various gatherings. Although they had some common ground as fellow Chinese, this phone call at this time seemed to be on behalf of his father delivering the leather jacket.

"Yiran, you're so smart, you can probably guess why I called you."

"Of course I could guess. But didn't you say you didn't want to take over your dad's business? What about your dream of becoming a bartender?"

"I'm not stupid. Even if I don't want to take over the business, it depends on whether my dad agrees. How about we have dinner together tomorrow night? I'll do you a favor. My dad agrees that if I get you to come out, he'll give me some money to open a bar in Taiwan."

"No way, you've been a rich kid since you were little, and you've never saved up any pocket money. You even have to ask your dad for help to open a bar?"

"My family has only become wealthy in the last few years, okay? Just tell me, can I come out?"

"Okay, send me the time and place. I'll definitely go to your bar."

Chen Yiran felt that Lao Huang was a bit different; inviting him through his son gave it a more traditional Chinese feel.

The following evening, Chen Yiran drove to the address Huang Shengbin had given him. It was a mansion in the Los Altos Mountains. To be honest, the mansion's style was a bit outdated. Aside from its size, it didn't quite fit the image of a wealthy man. Of course, Huang's net worth wasn't as exaggerated as it would later become.

Huang Shengbin stood at the door waiting for him. After a brief greeting, he led him through the living room into the backyard. Huang Renxun was standing in front of a barbecue grill, grilling steaks, wearing a very cool pink apron.

This style is too incongruous. In his mind, Chen Yiran imagined Lao Huang wearing a leather jacket.

"You must be Chen Yiran, right? Please have a seat. How well-done would you like it?"

"Seven points is fine. Do you need my help?"

"No, no, you're a guest, please sit down. My wife is with my daughter visiting universities, so it's just the three of us men for the next few days. Let's have a drink later."

The steak sizzled on the grill, and Huang Renxun chatted with Chen Yiran while flipping the meat, starting by talking about the feelings of the Chinese community.

"You're amazing! To achieve this at 18. I was still in high school when I was 18, in Oregon, and I was the only Chinese student in the whole school. Later, I went to Stanford for my master's degree, where there were more Chinese people.

However, there were only three Chinese students in our cohort, so the three of us helped each other. Later, when I worked at Nvidia, half of the best engineers we recruited in the early days were of Chinese descent, and it's the same now. I don't care about recruiting external talent like foreigners do; I prefer to hire the children of my employees.

We Chinese Americans, while we don't necessarily have to cooperate, should still try our best to help each other, right?

Although Huang's company is indeed the tech company with the highest density of Chinese people in Silicon Valley.

But Chen still couldn't believe a word of this. Lao Huang's sense of personal identity had always been pragmatic; he would only bring it up when he needed to.

It was obvious they were trying to fool this young man and take advantage of him. Although Chen Yiran was a soul from Zhongdeng, he had never actually spoken to a big shot of this caliber in his two lifetimes. Not knowing how to brush them off, he simply adopted the persona of a technical genius with very low emotional intelligence and pretended not to hear them.

Although Huang Shengbin had only met Chen Yiran twice, he knew that Chen Yiran was not a social idiot, and he found it a little funny to see Chen Yiran pretending to be blank and not responding.

Old Huang, being a seasoned veteran, wasn't embarrassed at all by the awkward silence. He quickly grilled the steaks and divided them between the two of them.

After a brief chat, the group changed the subject and Old Huang revealed his true intentions, stating his purpose for the day.

"Your DeepFlow is a great tool, not just because it's fast. It's because it allows more people who don't use CUDA to write CUDA programs. With that Python layer of encapsulation, the barrier to entry and flexibility of CUDA have undergone a qualitative change, which is very beneficial for NVIDIA in promoting CUDA."

CUDA is a general-purpose parallel computing platform and programming model launched by NVIDIA, enabling collaborative computing between CPUs and GPUs. In a little over a decade, it will be incredibly powerful, an indispensable tool for the AI ​​industry, covering more than 95% of AI developers worldwide.

However, CUDA, which is only 6 years old now, can only be considered to have a promising future.

Chen Yiran asked with a smile, "So you want me to keep CUDA as my priority?"

"Of course, logically speaking, it would be the best solution for us to invest in your company. Although I know you're probably not short of money and might not be looking for funding right now, I still have to give it a try. How about selling me 5% for $1500 million? That's a direct valuation of $3 million."

"Microsoft offered me $500 million yesterday, cash or stock, I could choose."

"Um... Microsoft, you can't trust them, they can mess up anything they buy."

"Of course I won't sell; I really don't need financing right now."

"That's right, so I won't talk to you about investment, at least not tonight. I can give you two conditions now. First, NVIDIA's CUDA toolchain team will help you with the underlying optimization of DeepFlow, which is at the kernel level. Your team of fourteen people is not enough, my engineers can help you with that."

Secondly, I can give you access to the new hardware architecture parameters and beta drivers six to eight months in advance, without waiting for the public release. I just hope to get a small favor.

DeepFlow publicly states, such as on your GitHub page and in subsequent documentation, that its optimizations prioritize CUDA and NVIDIA GPUs. It's not exclusive, but optimizations for NVIDIA cards are prioritized.

Chen Yiran felt insulted; both conditions were meaningless.

Nvidia's team will definitely use DeepFlow no matter what. If you don't use such a useful tool, others will take the lead and overtake you. This first condition is just to fool idiots.

Getting the new parameters in advance and asking him to prioritize Nvidia cards isn't the most basic thing; this second condition is just fooling idiots.

However, after some thought, Chen Yiran felt he had no choice but to agree. Although he could potentially create a new powerhouse with his full support for other graphics cards, that would take too much time. It would be better to just cooperate with Nvidia.

To put it bluntly, the current cooperation between the two sides is a powerful alliance, and Lao Huang's point is that he hopes for a powerful alliance rather than a powerful confrontation.

Chen Yiran made a decision: he would first write down Lao Huang's inhuman conditions in his notebook, and then, when he became successful, he would chase after N cards to beat them up.

"Alright, I think Nvidia is indeed the best choice. I agree to prioritize Nvidia cards."

"That's great. Here's what we'll do: get ready, and I can give you a keynote address at NVIDIA's GPU Technology Conference in September. You can talk about any of your technologies, and if there's anything that needs to be commercially demonstrated, you can showcase it then."

"Sure, thank you so much. I might actually have some results to promote in September."

After dinner, they chatted casually for another half hour before Chen Yiran said goodbye. Huang Shengbin saw Chen Yiran to the door. Fortunately, none of the three drank alcohol, and Chen Yiran drove straight home.

As Chen Yiran started working, he calculated the time and decided that he still needed to give his employees a little prank when he got back. If he could find SMIC to conduct some testing in July, it would be best to have a commercial version released in September so that it could be showcased at the NVIDIA Technology Conference in September.

If all goes well, we might be able to secure some orders, especially from domestic companies, which we can prioritize negotiating with.

It's also a good time to go back to China for a visit and see how various companies are actually developing.

Chen Yiran was thinking about returning to China for an inspection tour, and some people had already begun to consider how to keep him in the country.

On Laguna Street in San Francisco, there is an unassuming beige building that houses the Chinese Consulate.

The five-star red flag at the entrance flapped loudly in the sea breeze, and the hall was always filled with all kinds of people: those applying for visas, those renewing passports, and various Chinese seeking help.

These are the most basic tasks of a consulate, but the consulate in San Francisco, compared to consulates in other regions, has many non-basic tasks.

For example, Dong Jianlong on the second floor holds the position of Science and Technology Counselor, but his work is far from basic. He needs to constantly keep an eye on the Chinese science and technology talent in the Bay Area, maintain contact with people at Stanford and Berkeley, and help domestic universities find overseas partners.

The 42-year-old graduated from Beihang University and worked at the Ministry of Science and Technology for ten years. Although he was not a front-line researcher, he had a basic understanding of various research situations, especially in the field of science and technology and computer science, which is why he was selected to be stationed in San Francisco.

On his desk were several documents, the top one being the arXiv daily briefing printed out by his assistant that morning.

This is part of their daily work. Every day, they roughly scan all the foreign papers with Chinese backgrounds. For those with Chinese nationality, they are roughly screened by name.

The second page of today's briefing has a line highlighted in yellow, which is the annotation about Chen Yiran in Andrew Ng's paper.

There were also several newspaper clippings, all about Chen Yiran making a profit of 200 million US dollars by selling software to Microsoft.

Of course, this wasn't the first time Dong Jianlong had seen this information. If he only started paying attention to a genius who earned 200 million US dollars at the age of 18, it would be a serious dereliction of duty.

However, Chen Yiran's previous research focused on compilers, which was very impressive, but it wasn't part of the national development projects. So they simply kept an eye on it and continued to gather information.

Later, when Dong Jianlong learned that he had started an EDA company, he was extremely shocked. After all, this was not a promising direction.

However, Chen Yiran did nothing, so according to procedure, he did not make any contact.

Who knew that Chen Yiran's company, which had only been open for a little over a month, would start making such bold claims. Although it had only been three days, he received information that Google, Microsoft, Nvidia, and all the big names in Silicon Valley had begun testing DeepFlow.

There's no need to wait any longer; the current results have already proven Chen Yiran's abilities.

Moreover, although it's not fair to say that Chen Yiran's family background is perfect, her parents are both employees of state-owned enterprises, and their political background checks were also very good.

After making money, Chen Yiran immediately bought houses for his parents. The key point is that he bought two houses, one of which was likely intended for Chen Yiran to live in himself.

All indications suggest that he would not resist returning to China to develop his career.

It's time to call the prodigal son home. Dong Jianlong has decided to call Chen Yiran's company tomorrow.