After the layoffs, I became a tech godfather

Chapter 223 What defines a humanoid? Top-down adjustments.

Many things lead to the same goal, albeit through different paths.

Xiao Sha previously upgraded Skynet's capabilities, including analyzing the behavior and motivations of all traffic participants on the street, including people, vehicles, etc., initially to ensure safety.

These technologies have been used in technological white canes to help blind people judge and analyze the behavior of traffic participants around them and the actions they need to take.

And if these things are applied to intelligent connected vehicles, they can inevitably improve the decision-making capabilities of intelligent driving. "The technology is not the problem; it's just that our current technology has only reached the decision-making stage, and the subsequent execution is also crucial."

Hao Cheng had considered the issue of autonomous driving. According to the classification of autonomous driving logic, Xiao Sha's current analytical capabilities are definitely at the highest level, and it wouldn't be an exaggeration to say that he can surpass 99.999% of people.

However, analysis alone is certainly not enough; a comprehensive decision-making and execution process must also be made by combining factors such as vehicle size, mechanical quality, acceleration, and braking capabilities.

A simple example: there's a width restriction barrier in front of you. Can a car pass through?
You can't make a judgment based on just Xiao Sha's data; you need to know how wide and tall the car is.

On a highway, overtaking another car takes only a few seconds, which depends not only on the speed and motivation of other cars detected, but also on your own car's acceleration and top speed.

Only then can you make a decision about how deep to press the ignition switch and for how long.

In short, to apply Xiaosha's analytical skills to intelligent connected vehicles, it is also necessary to combine them with the vehicle's own data.

Baiju Technology does not have this data.

Xiao Sha's abilities are undoubtedly impressive, but how high a level they can perform ultimately depends on their ability to integrate their efforts.

"Why does this sound like you're implying something?" Yu Dong was driving and had been listening to Hao Cheng talk for a while, only managing to respond with "uh-huh" at the end before finally realizing what was going on.

"Let me ask you a question," Hao Cheng asked a very pointed one: "Are the intelligent driving programs of two cars of the same series and model exactly the same?"

As soon as Hao Cheng spoke, Yu Dong was startled, because he knew what Hao Cheng was going to ask, but he could only answer, "Yes."

As expected, after hearing his answer, Hao Cheng asked the crucial question: "When it leaves the factory, there should definitely be no problem, because the mechanical qualities of vehicles of the same series and model should be the same."

"But what about a year later? Let's say one car has already driven 100,000 kilometers, and another car is driven twice a year. Will they still be the same? If it's not obvious after one year, what about three years? Five years? Even with timely maintenance and replacement of parts, the difference in mechanical quality can't be exactly the same, right?"

"In this situation, if there is an accident ahead and the intelligent driving system determines that it needs to brake, its timing may be completely different."

Hao Cheng was pointing out a real problem: even if it's not two cars, but the same car, its braking distance and friction will be different on cement roads, asphalt roads, wet roads, and rainy or snowy roads.

Hao Cheng had also seen many thrilling videos of Huawei ADS's EAS all-around collision avoidance system. In these emergency situations, Xiao Yi would take over the steering wheel and swerve left and right to avoid extremely dangerous accident scenes with remarkable skill.

However, if it were an icy or snowy road, or even a wet and slippery road, but a concrete road, the situation would be completely different.

“Our current approach is to perform special calibrations for various scenarios. For example, we have developed adaptive AEB specifically for slippery roads,” Yu Dong explained regarding the current ADS strategy. “As for the strategy for intelligent driving on cement roads, asphalt roads, and muddy roads, there is currently no distinction.”

"What about the industry? Tesla's FSD?" Hao Cheng asked.

“It’s definitely the same,” Yu Dong said confidently.

Hao Cheng thought about it and realized that it made sense. If it had been done, people would have bragged about it a lot. It was such a great direction for humanoid research. Since it hadn't been mentioned, it couldn't be.

Moreover, Hao Cheng didn't believe that anyone other than Xiao Sha could accomplish this level of difficulty.

"So, the next step is to develop a solution for each vehicle and each scenario. For example, if it snows today, and a person is driving, anyone with a little experience can tell how fast they can go and how slippery the road is by the first time they step on the accelerator and the first time they step on the brake when they get on the road."

"If we combine Xiao Sha's analytical skills with what I just mentioned, and then set up dual-redundant backups for the perception, communication, power supply, steering, and braking systems, we can directly achieve driverless driving."

Yu Dong scratched his head, thinking that comparing people is infuriating, and comparing companies is even more infuriating.

But he couldn't refute it. He had been working hard for years, but Hao Cheng just talked about it. What was even more infuriating was that Hao Cheng seemed to be talking to himself, but if it really came down to it, with Xiao Sha's intelligence and abilities, it really wouldn't be difficult at all.

Who the hell can you reason with?!
What's even more infuriating is that Baiju Technology didn't treat this as a serious matter. Just look at the more than 10,000 research topics on their official website, and you'll know that intelligent driving or autonomous driving is not in Baiju Technology's plans.

"Besides the fundamental improvements in chips and AI, the two areas I value most are bioengineering and medicine."

Hao Cheng didn't hide anything: "The rest depends on our partners. For example, if you want to do autonomous driving, you can just apply to Xiao Sha to fine-tune the model and do it yourself."

Yu Dong fell silent instantly. This was truly throwing away a mountain of gold like it was made of stones!
While providing fine-tuned models isn't free, the fee is negligible compared to the huge downstream market.

"Apply! We absolutely must apply! We'll submit the documents as soon as we get back, requesting the capability to fine-tune the driverless Xiaosha model!"

One vehicle, one solution; one scenario, one autonomous decision-making; an intelligent driving or autonomous driving system that can evolve and grow on its own—that's incredibly appealing.

We must seize this opportunity before others do it.

Yu Dong is full of vigor, like a young man in his twenties or thirties, but in reality, he is already 55 years old this year.

Ren Zhengfei, the leader of Huawei, is a different kind of person who reveals a kind of unconventional wisdom, completely different from Yu Dong's sharpness.

The next day, Hao Cheng, along with Yu Dong, also visited the old man, who was already nearing eighty.

"It's clear that you're a young man with ideas," said Mr. Ren in his evaluation of Hao Cheng. "You have the desire and the ability to change many things and lead the world forward."

"I have a suggestion: relinquishing some benefits is the most effective way to control the rules. To put it bluntly, if you can make people money, they will listen to you."

"Since you have decided to share some of the results, unite some of the partners, and cooperate on some of the projects, the opportunity to change the whole society has already arrived."

In just a few months, he rose from the depths of unemployment to the very top of the tech industry. When a single decision can impact the entire industry and even the entire course of technological progress, many decisions must be made with utmost care.

Hao Cheng benefited greatly from his several hours of conversation with Mr. Ren.

Leaving aside the technical aspects, decades of life experience cannot be grasped in just a few months.

If this applies to humans, shouldn't the same principle apply to AI training?
Xiao Sha is so strong now, and his training is so rapid, but compared to others, what he lacks is this part of his past experience and wisdom.

Hao Cheng still needs time to digest the conversation with Mr. Ren that morning. In fact, without firsthand experience, he would never gain any insights, let alone AI.

"This question wasn't even mentioned in the late-night classes! I'll design a customized late-night class and ask them for advice when I get back!" Hao Cheng thought to himself.

……

Hao Cheng didn't stay in Pengcheng for long; he quickly returned to Linzhou with a lot on his mind.

Upon returning to the park, Hao Cheng immediately convened a meeting with all personnel.

"There are two aspects to this. First, we will be optimizing and adjusting our organizational structure this month."

Hao Cheng smiled and said, "Look at all of you, what kind of expressions are these? Is organizational consolidation and optimization the same as layoffs in your minds? No, at least not at Baiju Technology. Not only will there be no layoffs, but according to Xiao Sha's assessment, after the adjustments, the company's total staff will be three times larger than it is now."

"Furthermore, the number of management and senior technical positions will increase several times over. This is an opportunity for each of you. As you know, our company offers limited internal promotions if you are a suitable candidate. Okay, let's get back to the main topic..."

Many things require trying, and on the plane back, Hao Cheng worked with Xiao Sha to develop the basic framework for Baiju Technology.

At the structural level, the company is directly divided into three parts from top to bottom: scientific research, technological breakthroughs, and成果转化 (technology transfer).

Scientific Research: Register Baiju Technology Research Institute Co., Ltd., and register corresponding laboratories for chips, AI and related disciplines. Its main responsibilities are scientific research and talent training. When necessary, it can jointly train master's and doctoral students with various universities.

Many companies are already engaged in joint training programs between universities and enterprises, and there are already relatively mature approaches and routes, with clear and replicable policies in place.

Technological breakthroughs refer to the exploration of technological implementation and mass production based on theoretical research. These are consistent with scientific research, but are closer to reality. In other words, compared to the long cycle and future-oriented nature of scientific research, technological breakthroughs are technologies that can be achieved now and mass-produced quickly.

Commercialization of research results refers to transforming the achievements of technological breakthroughs into products. Examples include AI robots like Xiao Sha and children's robots that we developed ourselves, as well as collaborations with other companies on projects such as humanoid robots and autonomous driving solutions.

An organizational chart was displayed in front of everyone, clearly indicating which positions were still needed.

Hao Cheng said, "After the meeting ends, you can think about which positions you are interested in, and then let Xiao Sha assess whether you can meet the requirements of the positions."

"The worst outcome is that things will remain the same, and there will definitely be a place for you in the company, that's for sure. But I hope that all of you here can go further and fill every position we need."

"Cao Song, what are your thoughts?" Seeing Cao Song hesitating, Hao Cheng called him out directly.

“I think that in terms of technological breakthroughs and the commercialization of research results, we don’t necessarily do all of it ourselves. I have a question: how are these positions allocated for these collaborative projects, or even for these Class C collaborative projects that we tender or offer rewards for?” Cao Song asked.

"It's very simple," Hao Cheng explained. "For collaborative projects, such as manufacturing equipment and production lines for silicon-carbon fusion semiconductors, we have our own team, and we work together with our partners' teams to solve these problems."

"For Class C projects, which are entirely developed by our partners, we will still need to have a liaison. Depending on the project, there may be a small team, an individual, or a group responsible for multiple projects, but there will definitely be someone from our team to follow up."

The design of this architecture follows a route from scientific research to technological breakthroughs and then to the transformation of results. At the same time, if difficulties are encountered in the transformation of results, there is also a route that is a mechanism for feedback upwards!
When designing a specific product, if a problem is found that cannot be solved, there is a direct route to submit the problem to the relevant technical breakthrough department, and they will tackle and solve it.

If a technological breakthrough requires theoretical support, it can be submitted to higher levels. Scientific research will verify whether the theory is valid and whether the technology is feasible. If it is feasible, a research project can be initiated. If it is not feasible or the cycle is too long, it is recommended that the technological achievement be changed.

“Therefore, we definitely need people for every step of the process,” Hao Cheng said. “We have already decided to hand over the Class C projects to our partners to lead, but in today’s world, even for our Class C projects, there are very few that can be done independently.”

"Either we need more support from Xiao Sha, or we need to set up a dedicated team to collaborate. To put it simply, this model is where we provide the outline and then cultivate talent for the industry."

Cao Song understood Hao Cheng's plan, but if you calculate it this way, putting everything else aside, the projects with these more than 10,000 partners alone would require tens of thousands of people to be recruited. And the company's entire recruitment plan this time is estimated to be heading towards nearly 100,000 people.

Laboratories, R&D centers, and technical departments require expensive equipment, which is a huge expense just to think about!
“Projects and positions marked with an asterisk are the top priority; other projects and positions can be implemented gradually. The initial recruitment plan for our first batch has been finalized at 10,000 people. I have already asked Xiao Sha to compile the relevant requirements and share them with HR so that arrangements can be made as soon as possible.”

This recruitment plan involves a staggering 10,000 people, enough to practically fill the entire small town of Xiaosha. This time, the scale is truly enormous.

“This involves personnel and partners,” Hao Cheng continued. “Previously, we issued many rules in a scattered manner. Today, I want to establish the code of conduct for our partners in the form of rules and conduct a credit rating for the companies.”

Giving up benefits is the most effective way to control the rules. If you can make people money, they will listen to you. The opportunity to change the entire society is there.

Mr. Ren's words enlightened Hao Cheng, revealing that the conditions for establishing partnership rules were ripe.

Just like the current credit system, a person's breach of trust has a significant impact. Even without a breach of trust, many companies will refuse to hire you simply because you have an arbitration record.

However, a company's breach of trust has no impact whatsoever, no matter how many lawsuits it has, or even if it has lost those lawsuits. This is clearly unfair.

What Hao Cheng needs to do now is to conduct credit rating assessments of his partners and businesses within his business system.

Hao Cheng originally had concerns, worried that there might be some negative consequences.

But in reality, Hao Cheng was overthinking it. There are many rating agencies like this, both domestically and internationally. If you are authoritative, everyone will recognize you. If you are not authoritative, you can use it within your own system without any impact or need to worry about anything.

Even in the banking sector, Alipay has Sesame Credit, and WeChat has Credit Score, but it hasn't made much difference.

"We'll start with our own system, and then guide all companies to recognize and follow our standards," Hao Cheng said. "Specific rules and details will be solicited from our users and partners directly through Xiao Sha."

(End of this chapter)