After the layoffs, I became a tech godfather

Chapter 234 A new batch of chips is about to be released.

For several days in a row, Hao Chengdu visited the chip factory in Shenzhen to learn about the trial production of silicon-carbon fusion semiconductors.

The results of the packaging test show that the electrical architecture, properties, performance and other indicators have all met the design goals and requirements.

"Next, we need to test and verify the actual installation and networking performance of these chips." Hao Cheng concluded, "The first batch, a thousand chips, is considered small-scale mass production. We will make further plans after we have evaluated the results."

A completely new architecture, or even a completely new concept, has no precedent to follow. Even if Hao Cheng has a clear idea in mind, he still needs to carefully verify every step he takes.

In the past two days, Hao Cheng has once again sorted out the entire line between metacognition architecture, mapping instruction set, and unbounded operating system.

In particular, he also had relevant exchanges and communications with Lin Wen and the entire team of the Boundless System:
"The operating system is undoubtedly the most crucial factor in determining the ultimate effectiveness of a chip based on a completely new concept. You've been researching the relevant concepts for several months now, and it's time to put them into practice."

Leading his team to join Baiju Technology was the best decision Lin Wen ever made in his life.

The first product after joining the company, Wujie System, had by then become the uncrowned king of domestic operating systems.

Whether it's a computer, server, or mobile phone/tablet, you can either use the Boundless system directly, or like Huawei, have your own system but use the core components of the Boundless system.

Lin Wen's most frequently uttered phrase is: "It's not that I'm so strong, it's just that I chose the right path."

Thus, the heavy responsibility of adapting the new concept chip and the entire metacognitive system fell on his shoulders once again.

For several months, the team did nothing. They even handed over almost all the maintenance of the Boundless System to Xiao Sha. Their only task was to understand what cognition is, what cognitive mapping is, and how to form a set of instructions that the three-dimensional meta-cognitive chip can understand.

However, several months have passed, and it's time to make a move.

Lin Wen, however, looked somewhat worried.

"What? No confidence?" Hao Cheng smiled as he looked at the hesitant Lin Wen.

"When we look at the documentation for the Metacognitive Instruction System with our eyes open, we think, 'Hmm, I understand.' But when we close our eyes and try to design an operating system architecture based on it, we feel like our minds are blank and we don't know where to start."

Lin Wen said truthfully, and it wasn't just him; it seemed like all the team members felt the same way.

"That's normal." Hao Cheng offered a few words of encouragement. He knew this process had taken many years.

Even with the help of late-night classes, he still went through such a period in learning the relevant content.

Cognitive instructions are essentially non-definite instructions. However, human thinking, with its established patterns or inertia, is more receptive to definite and clear matters.

As for patterns, it's not difficult to summarize them.

However, non-definite states are extremely challenging for a person's "fuzzy" thinking ability.

Besides talent, continuous thinking and expanding one's mental capacity are also crucial factors.

"You'll learn by doing, and Xiao Sha will help you," Hao Cheng said.

"Xiao Sha..." Lin Wen really wanted to ask, could Xiao Sha do it? Could it be that Xiao Sha is stronger than humans in terms of non-definite cognition?

“Don’t doubt this,” Hao Cheng shook his head: “AI is different from humans in this respect. It’s hard to explain. If I had to put it this way, let me put it this way: humans rely on understanding, while AI relies on memory.”

"Sometimes you can't get over many things because you have too strong a thirst for knowledge. For example, with metacognition, given the conditions and the result, your mind will involuntarily start thinking: why?"
"The reason for this is currently very complex and difficult to understand before your brain has been trained. But Xiao Sha is different; you give it the conditions and the result, and it just matches them up."

"So there's no doubt that it can get things done. And it can guide you through the process. As you follow along, you'll gradually gain a deeper understanding of metacognition, and then you'll comprehend it."

Hao Cheng didn't bother explaining to them again and again what metacognition was. Hao Cheng confirmed that it was related to whether certain neurons in the brain had formed connections. Before these connections formed a closed loop, it would be useless to explain it any further.

Just as Lin Wen himself said, you can understand with your eyes open, but you can't do anything with your eyes closed.

Therefore, I can only gradually master it during the course of my work, and I estimate it will take about a few months.

Of course, talent also plays a role.

"Indeed, back then, when I was learning advanced mathematics, I felt like I was just getting by. But after a year, I was in my element." Lin Wen also recalled, "But some people just can't do it and can never get past that hurdle."

"That's pretty much it. We'll see when you all suddenly have a breakthrough and can immediately understand everything here. At that point, it's either about doing what you understand, or just following along if you don't understand."

Hao Cheng gave the order: "Boundless System (Metacognitive Version), your timeframe is one month."

"I will be in Shenzhen for a while, but you can contact me anytime if you have any questions."

"Aside from two meetings this month, I have no other work scheduled. My primary task is to complete the development of the metacognitive architecture version of the boundless system," Hao Cheng explained.

"I feel more at ease now that you say that." Lin Wen breathed a sigh of relief. He really wasn't sure if he had to take full responsibility himself. With Xiao Sha's help and the leadership backing him up, there wouldn't be any problems.

"Okay, you guys get ready first, I have a meeting to attend to in a bit."

……

In the past few days, Hao Cheng has been updating the outside world on the official website regarding the progress of the silicon-carbon fusion semiconductor chip.

In his own words, hiding things or keeping secrets is pointless.

"What I like most about Baiju Technology is its style. They really don't treat us like outsiders. They release all the news to us directly. In other companies, wouldn't only major shareholders receive this kind of treatment?"

"What do you know? This announcement isn't just about the progress of silicon-carbon fusion semiconductors!" "What else is there?"

"You can't just look at a problem from one perspective; you have to learn to connect all the dots. For example, Mr. Hao once said in response to a netizen that the AI ​​decision-making ability and feedback to the brain in artificial organs will see a significant improvement in the future."

"When do you think about this future? If there's a turning point, wouldn't it be when silicon-carbon fusion semiconductors become available?"

"Yes, yes, Mr. Hao once went to Wang Rui's live broadcast room and made a statement at the time, I think that's what he said."

"See? I knew it would be like this! Now, try to apply this to other situations. When will Xiao Sha's improved abilities be able to solve the current problems with the artificial kidney? And what about all the fine-tuning models that our many collaborative projects are currently applying for? Won't they all be enhanced or accelerated after this chip is released?"

"But why are the children's robots being sold now still CPU-less versions? Haven't EUV lithography machines been successfully developed? Why is the production capacity of traditional chips still insufficient?"

"I don't know if you guys have noticed, but it's been four months since Chinese New Year, and no company has released a new phone yet. What big surprises do you think they're all preparing?"

A word to awaken the dreamer.

In previous years, spring was the golden season for many mobile phone manufacturers to release new models. Not to mention the countless new models released by Vivo and OPPO's budget phones, OPPO's mid-generation flagship iteration, and Vivo's Ultra version were also released in the spring every year!

This year, everyone has fallen silent, as if by prior arrangement.

Throughout the entire spring, no major brands released a single new phone.

"No way, really? They're going to abandon Qualcomm and MediaTek and make all their own chips? Is there any progress?!"

"Of course there is. The mobile phone manufacturers must have received some information to do this. Moreover, Qualcomm and MediaTek's chips must be lagging behind, and to some extent, they are lagging behind quite a bit, which is why this could happen."

"That makes sense. I bet it won't be long, maybe next month at the latest. A whole bunch of new machines you can't even imagine will be coming! When they're all made with domestically produced chips and the Boundless System, just imagine the scene, it's amazing!"

"Oh my god, I was thinking of getting a new phone soon, but it seems I'll have to wait!"

……

It must be said that some people are really quick-witted; they can always glean part of the truth from the smallest clues.

They were absolutely right; a new batch of chips is about to be released.

While making breakthroughs in silicon-carbon fusion semiconductor manufacturing, EUV, which has already completed trial production, has not stopped its progress. At the last chip summit, it was confirmed that although the share of traditional semiconductors will gradually decrease to 20%, the total volume will continue to increase. As a result, the construction speed of wafer fabs is like a rocket.

Unlike ASML's EUV lithography machines, which are bulky, have numerous parts, and are slow to debug, the debugging efficiency in China is more than ten times higher with the assistance of Xiao Sha.

If it weren't for the limitations of parts and raw materials, it would have been possible to build more than two hundred production lines in two months.

Even under various constraints, the number of production lines that have been put into operation has reached eighteen.

There is an even more critical issue here: the lithography machine and production line controlled by Xiaosha have a production efficiency that is far higher than that of traditional production lines.

Although it is still far behind TSMC in terms of scale, its production capacity can meet the needs of most users without any problem.

More than two months ago, Baiju Technology released an electronic design automation tool based on Xiaosha, which greatly improved the speed and quality of chip design for various companies.

Moreover, this batch of chips is not just for mobile phones; it also includes game cards, computing cards, server and PC processing chips designed based on the characteristics of the Boundless System, all of which have passed trial production tape-out verification and entered the mass production process.

The meeting Hao Cheng just mentioned attending is a coordination meeting—is it appropriate to release so many chips all at once? How should different manufacturers and different types of chips be arranged?
"It's a chip feast. For mobile phone chips, it's just Huawei and Unisoc. HiSilicon's chips are not currently sold externally; they are all used internally, so there is no actual conflict with Unisoc."

"In the gaming card segment, there's only Moore's Thread, so there's no conflict. As for AI training cards, the launch event was more geared towards enterprises. In the computer CPU segment, none of the three companies had much of a reputation or buzz before, so they're all starting from the same point now."

"I haven't seen your performance test results, but with such powerful tools, surpassing Intel should be no problem at all. I think it's good to release them together. Competition is inevitable; if you want to improve, you can't be afraid of competition."

"So, let's just set a time and hold a grand chip launch event, releasing them all at once. Just think about the impact of that—wouldn't it instantly break through to new circles? It would make everyone forget about Intel, AMD, Nvidia, and so on."

Everyone present twitched at the corner of their mouths, thinking to themselves: You really are just looking for trouble.

"Well, this is just a suggestion of mine."

Hao Cheng originally wouldn't have gotten involved in this, but coincidentally, there was an international semiconductor exhibition in Pengcheng in a couple of days, and these companies had all gathered there to ask for his opinion, so he just said it directly:

"How about we collaborate with the exhibition organizers, have everyone make a collective appearance the night before the opening ceremony, and then directly participate in the exhibition the next day? Wouldn't that be killing two birds with one stone?"

“That works too. Let’s talk to the organizers and ask them to give us a dedicated exhibition hall.” He Gang’s eyes lit up when he heard this, and he suggested, “Baiju Technology’s silicon-carbon fusion semiconductors can also be exhibited. We already have samples at this point.”

"Hmm!" Hao Cheng nodded thoughtfully: "Just showing samples isn't very interesting. How about this, we'll announce our design concept, application scenarios, and structural design at this electronics exhibition, and also do a short science popularization class. I'll find someone to arrange it right away."

"Hmm!" Regardless of what these people thought before, no one would refute Hao Cheng now that he had spoken.

In fact, this is a pretty good option for them.

In the past, any similar chip would have caused a sensation across the country or even the world when it came out, but with a large batch coming out, no one can guarantee that they will be the first to emerge victorious.

Putting them together is a pretty good choice.

As for the fact that the semiconductor exhibition in Shenzhen is not very influential, that is not important at all. As long as this news is released today, this exhibition will immediately become the most influential semiconductor exhibition of the year.

"So, hurry back and get ready!" Hao Cheng waved his hand, and that was it.

(End of this chapter)