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

Chen Yiran said with a bit of domineering confidence, "I want you to develop your own instruction set architecture from scratch, and make it compatible with both ends. We're going to send those damned ARM and x86 to their graves."

Olofsson leaned forward, his face finally free of the restraint of a bankrupt boss, and said with some excitement, "Building the ISA instruction set architecture from scratch? That's definitely more exciting. A completely new instruction set, which we can implement exactly as we envision."

Chen Yiran nodded: "ARM and x86 have indeed accumulated over thirty years, but on another level, in order to keep backward compatible, a lot of transistors have been wasted."

Olofsson nodded and said, "The idea of ​​using the same instruction set for ARM, only changing the microarchitecture for mobile and desktop devices, has never been successful. We can do better by using a simplified ISA with a minimalist small core array in parallel."

Chen Yiran set a specific goal: "Yes, or to put it more grandly, we want to create a new instruction set that is more streamlined than ARM and better suited for heterogeneous computing, unifying CPUs, GPUs, and NPUs!"

Qiu Ciyun nodded seriously: "I can find some manpower to do microarchitecture verification, and poach some people from HiSilicon. There must be many people willing to do big things and try to overthrow ARM, and our company's conditions are good enough."

When the conversation turned to specifics, Olofsson became somewhat uneasy: "Even with cheap labor in China, such a large project would still cost a lot of money. Although I'm not very market-savvy, I know that this kind of project isn't likely to be profitable."

Qiu Ciyun thought for a moment and said, "Actually, it's not too bad. Labor costs are really cheap in China. Even for an ISA architecture designer, you can earn 500,000 to 600,000 RMB a year. Other aspects are even cheaper."

However, there's a problem: the required personnel... even if I poached all the people from Loongson, it might not be enough. Specifically, for the ISA architecture design, Olofson, you'll need to help recruit some people in the US.

Olofsson said hesitantly, "Actually, there are quite a few people willing to come with me, but the salary in China is indeed much lower than in the United States, so I can't help it."

Chen Yiran waved his hand dismissively: "I'm so poor right now that all I have left is money, so money isn't a problem. How about this: for those willing to come to China with you, we can give you real estate. You can check out the housing prices in Beijing. And you won't be short of any benefits either."

Qiu Ciyun glanced at Chen Yiran in surprise. He hadn't expected Chen Yiran to be so generous. He did the calculations carefully and said, "If we only talk about the initial labor costs, it's not that high. It's a team of about a hundred people, plus some basic equipment."

An annual investment of over ten million US dollars would suffice. However, Mr. Chen, I must remind you that your idea will not be easy to realize. And even if it does succeed, cultivating the subsequent market will be extremely difficult, and the subsidies required will be astronomical.

That's the truth; just thinking about promoting this thing is a hassle.

ARM and x86 have been around for decades. Although the software runs on the operating system, there is an isolation layer in between. In theory, all compatibility issues can be solved by simply adapting the operating system as a whole to the new instruction set.

However, in reality, many video decoding, encryption, and graphics rendering functions in various software programs require each ISA to be compiled separately.

So a vicious cycle ensues: without software support, you have no users and can only contact various software companies for support; but since you don't even have users, the major software companies certainly won't be foolish enough to send people over to do adaptation.

ARM is a small company that thrives on relentless accumulation of code. Since the 1990s, they've employed teams to continuously submit ARM architecture support code to the Linux kernel. As a result, universities worldwide receive their free development boards, successfully dominating the market by allowing students to focus solely on their studies.

If this balance is to be broken now, ecological subsidies are necessary, with hundreds of millions of dollars in subsidies every year for more than a decade.

Chen Yiran shook his head: "You don't need to worry about the subsequent ecological issues, I will take care of those. You only need to focus on the technical issues."

Olof breathed a sigh of relief, feeling quite pleased. Working for someone else was great; he'd stick to working for Chen Yiran from now on. Only a fool would start their own business.

Qiu Ciyun believed that Chen Yiran's target was China. Many key domestic departments, such as finance, energy, and telecommunications, would certainly support a domestically developed framework if given the choice.

With sufficient technical capabilities, this market segment is readily available. Qiu Ciyun thought for a moment and asked, "Then how about setting up a new project team under Wan Guoyuan? Let's initially allocate a fixed annual investment of ten million US dollars, and after three years, if there's some success, then we can discuss issues like ecological subsidies?"

"Sure, just let me know if you need more money. Doubao's patents and technology belong to me personally; I just entrusted FlowEDA with their maintenance. I have a fixed income every year, more than I can spend."

Olofson, who had failed in his business venture due to insufficient funds, shed tears of envy. Qiu Ciyun, though somewhat uncomfortable with the young people's flaunting of wealth, felt a surge of joy at the thought that this was her boss.

He found his job at Shougang very comfortable. He didn't have to worry about any messy stuff; the boss had almost no requirements for finance, only for the technology to keep pushing forward. Any technical disagreements? They could handle both—there was plenty of money. This feeling was truly wonderful.

After seeing off his two key subordinates, Chen Yiran lay on the sofa and said to Jasmine in his mind:

"It's decided! I'll reject Jobs. I won't bow down to a mere 30 billion dollars."

Are you sure? That's 30 billion US dollars.

"It's confirmed. We've already made so many arrangements, it would be a huge loss to give up now!"

We've secured EDA, gaining access to chip design. We already have a wafer fab, securing the manufacturing side. Now, with the ISA in place, the major components are all in our own hands.

What we're working on is brain-computer interfaces, and I wouldn't feel comfortable leaving any part of it to someone else; it all has to come from ourselves.

I need to control every aspect; joining Apple would be a losing proposition.

Okay, actually I also think it's better not to join. Why give the honor to a small apple? We're the whole Fruit Garden; logically and emotionally, we should be the ones eating the apple! ʅ（´◔౪◔）ʃ

"That's ambitious! I'll give you the naming rights to this project."

From a technical standpoint, this project should be considered the foundational framework; otherwise, it could be called the Jasmine King Framework.

"Forget it, I'll do it. Let's call it the watermelon frame. For Chinese people, watermelon should be considered the cornerstone of fruit."

[Is the angle that tricky? Okay, do I need to remind you what to do next?]

"Do you need to remind us! After signing with AMD, hurry up and get them to send over the technology. We'll be able to learn a lot about graphics card architecture design. From then on, there will be no weaknesses in our chip manufacturing technology system."

Um, what I mean is, it's time to walk the dog.

"That works too."