After the layoffs, I became a tech godfather

Chapter 178 Production Problems Can Be Solved
Now, the biggest problem is manufacturing.

“From simple vertical through-holes to 360-degree omnidirectional mesh connections, the biggest problem is accuracy.”

This is Chu Bai’s judgment, and it is also an important reason why he had to come here this time: the application of machine vision and deep learning in the field of chip manufacturing.

Of course, before coming here, Chu Bai didn’t know about Baiju Technology’s carbon-silicon fusion semiconductor solution.

But the application of AI in chip design and manufacturing is not uncommon.

Even now, EDA tools, especially simulation tools, have begun to use AI reasoning capabilities to a large extent.

However, Huawei's chip design in the past two years has been limited by domestic process technology, and iteration has been extremely difficult.

Originally, Chu Bai came here to discuss with Baiju Technology whether it is possible to upgrade the existing machine vision and AI models to further tap the potential of current performance based on the existing industrial foundation.

"As long as there is data, then of course it is possible." Hao Cheng nodded: "It's just that it's hard to judge to what extent it can be achieved now!"

"It's not difficult to judge!" Chu Bai was excited all of a sudden: "Compared with language models, whether it is computational lithography, mask optimization based on deep learning, or even various aspects of the automation system, it is much simpler.

"Huawei has accumulated a lot of experience in the field of industrial AI. By working together, we can improve performance by at least 50%. The production problem of carbon-silicon fusion chips can also be solved!
"I have several plans in mind now, but they all need further verification. Therefore, the carbon-silicon fusion chip can also be handed over to us for experimental manufacturing until it can be mass-produced."

The "we" that Chu Bo mentioned does not only refer to Huawei, but also includes almost all the top companies in the domestic chip industry chain such as SMIC and Shanghai Microelectronics.

Under the strict blockade of America, Huawei's Kirin, Ascend, Kunpeng and other products can achieve nationwide mass production. In addition to their own unremitting efforts, this is also inseparable from the strong support of these companies.

And this is exactly what Hao Cheng meant.

Baiju Technology currently has no experience in chip design and manufacturing. Carbon-silicon fusion semiconductors are now more like an idea and concept, and the sample size of Hao Cheng's design is also relatively small.

There is still a long way to go before it can be applied to large-scale integrated circuits.

Silicon semiconductors have already gone through this journey, and quite a number of people have vast experience with it.

If these experiences can be combined with the idea of carbon-silicon fusion through Xiaosha or Moshu, then it will truly take off.

These design and manufacturing experiences are not available on the public Internet, and are definitely absolute secrets for major chip design and production companies.

It is in everyone's interest to take advantage of this opportunity to bring these companies together and form a super-powerful combination to achieve breakthroughs in chip design and manufacturing.

Hao Cheng naturally agreed.

"Then let's make a preliminary draft," Hao Cheng said, "and then gather everyone together to work out a detailed plan."

"Well, no problem!" Chu Bai has obtained a very high level of authorization, and he can make the final decision on this preliminary plan himself.

……

For two consecutive days, Hao Cheng had been receiving experts or business leaders from important fields who were invited to come. During the negotiation process, he was also building the future plan of Baiju Technology step by step.

"Gather a group of top technology companies, top research institutions and universities, and top talents, and then jointly climb to the top of science and technology, Xiaosha, what do you think?" Xiaosha has participated in all the talks in the past few days. In fact, in many areas, Xiaosha's understanding is deeper than Hao Cheng's.

But Xiaosha was stuck on this question that Hao Chengdu was very confident about.

[I… I don’t know.] After Xiaosha became more human, he even paused for a moment before adding an explanation: [My ability is limited to the present, and fusion analysis can be said to be my strong point.

[However, things are constantly changing. Today's technology may not represent next year, next month or even tomorrow. I can only analyze the problem based on the current technical level, which is obviously not appropriate.]

Hao Cheng nodded. The so-called reasoning ability of AI now is actually calculation rather than thinking. The so-called parameters refer to the breadth of the calculation. The wider the scope, the higher the computing power required and the slower the execution.

Even Xiaosha, which is based on the micro-consciousness model algorithm, is based on the mapping fusion of operations, and is still a long way from the word "thinking".

As for looking at things from a developmental perspective, it is naturally difficult to do so.

[Because of the limitations of the existing computing power card hardware, my current capabilities are only two-dimensional computing fusion. ]

Xiaosha said: [My thinking is still quite constrained. The interface data tells me that if I use the micro-consciousness chip to achieve three-dimensional computing fusion, my ability will be explosively improved.]

"Is that so?" Hao Cheng was delighted: "Isn't this coming soon?"

[The carbon-silicon fusion semiconductor is not yet a complete micro-conscious body, so it is still unknown how much of my original strength it can exert. Of course, it will definitely be stronger than it is now, that's for sure.]

"That's enough." Hao Cheng nodded: "Take it step by step. When you become more capable, you can design more powerful chips, and stronger chips can bring out your stronger level. Then you can design even more powerful chips... In this way, sooner or later you will be full of blood, right!"

[Yes!] Xiaosha seemed to be confused by Hao Cheng. After a long while, he finally came out and answered with the word [Yes].

It is not easy. If the production conditions are met, then we will design a plan here, send the layout to the wafer fab, and go into mass production directly after successful tape-out. This is the normal process for a chip company.

But now, after the design plan has come out, there is no chip factory with production capacity. It is necessary to improve the chip engineering and other upstream and downstream industrial chains together. The difficulty is simply unimaginable.

What’s more crucial is that this improvement is completely beyond our control, which makes us feel very uncomfortable.

Hao Cheng can't wait for the entire industry to come together and form a powerful combination. That way, with Xiaosha's help and Mo Shu's coordination, the entire industry can improve more quickly.

……

[Boss, there are only two and a half hours left before the press conference. You should go to the site.]

January 2025, 1, the 22rd day of the twelfth lunar month of the Year of the Dragon in the lunar calendar, is a Wednesday.

At exactly twelve o'clock, Xiaosha reminded Hao Cheng on time to go to the press conference site at the Kingdee Sports Center Gymnasium located in the nearby CCD. The first press conference after the establishment of Baiju Technology will start in two and a half hours.

(End of this chapter)