After the layoffs, I became a tech godfather

Chapter 51: Having skills means being willful

"Xiaosha, Xiaosha, upload [the personalized information acquisition algorithm based on the client and assistant AI] to the open source community of the Open Atom Open Source Foundation for me."

This algorithm model is fundamentally contrary to the original platform recommendation algorithm. In theory, the user's assistant replaces the user himself and represents the user's own ideas.

But, on the other hand, assistant AI is also developed by companies. For example, Xiaosha is developed by Baiju Technology. Will the development company put its own private goods into the assistant AI?
Of course Hao Cheng did not smuggle anything in the small sand, but he had to have others believe it!

Moreover, the recommendation and acquisition of information is indeed a very sensitive issue.

If the platform recommends this model, it can only control its own platform, but the information acquisition based on assistant AI is equivalent to controlling the entire channel on behalf of the user.

Such a thing is even more sensitive, and Hao Cheng naturally has no intention of controlling or manipulating public opinion.

After careful consideration, he decided to make the entire algorithm public directly in the Open Atom Open Source Foundation open source community.

Of course, what is open source is only the algorithm for information acquisition. This algorithm works together with the assistant AI to actively acquire information based on the rules set by the user and in combination with the user's behavioral habits and preferences.

This fundamentally eliminates any possibility of precisely incorporating personal opinions into a single viewpoint.

As for inaccurate private goods, you don’t have to consider them. First, their targeting is limited and the effect will definitely not be good.

Second, if you really want to smuggle, you have to start with the output of the model itself. Otherwise, you might ruin the model you have trained carefully. It is definitely not worth it for a private gain.

Of course, there is a biggest paradox here, that is, if you open it up, will others use it?

Even if we put aside the issue of smuggling, paid streaming itself is a huge profit, and the platform will definitely not give it up.

What's more, traffic is the right to speak, and the "fourth power" that can make things happen with a wave of the hand is also very tempting.

It is fundamentally unrealistic to ask the platform to actively use and connect with this thing!
Therefore, this is just Hao Cheng’s first move.

“For the second step, Xiaosha, please help me write an article titled “Who is controlling my preferences? The darkness behind big data and AI recommendations!””

Although it was Saturday, Hao Cheng still didn't get to rest and was forced to start working again: "The evidence is what you monitored today. The things that attacked us, from the source to the platform to the users, from production to processing to streaming, reveal the whole chain to them!"

"Also, send a survey to all Xiaosha users: Do you want your information channels to be controlled by various platforms? If you have encountered such a situation yourself, provide relevant information and evidence and you will get a one-month membership for free!" Hao Cheng finally explained to Xiaosha.

……

Hao Cheng felt that he was worrying a lot, but in fact, these should have been the responsibilities of the supervisors.

But in the field of technology, supervision is actually in a state of extreme polarization: on the one hand, there is a one-size-fits-all blockade and a one-size-fits-all wall.

But at the same time, in some areas, supervision has been out of place for a long time.

For example, for websites, apps, algorithms, etc., all the regulatory authorities can do is just to file them. What can filing do? It’s just to find someone if there is a problem!

What about before the problem occurs? Regulation is completely out of place.

Why? Because the senior talents in the industry are all in the enterprises. The regulatory authorities have no idea what others are doing and how they are doing it. How can they regulate when they can’t even understand it?

Take the most common and widely used streaming algorithm for example. There is no way to monitor whether it is fishy or working for foreign forces. Once there is a problem, find someone to ask. The question is big data, the judgment and decision made by the algorithm itself, and we don’t know. As for whether it is true or false, who can verify it?

You can't say that people have a bad attitude. As soon as you make a request, they will make corrections according to your requirements and get away with the problem this time, but they will dare to do it again next time.

The current state of the domestic public opinion environment is closely related to the lack of supervision.

Otherwise, no matter how much a few media outlets talk about it, if they don’t give it traffic, it won’t be able to make any headway.

"So, Mr. He," Hao Cheng told He Gang when he called again to ask, "Have you seen the algorithm I uploaded to the open source community?"

"You really want to help them dig their graves!" He Gang looked at the algorithm and had an in-depth discussion with the engineers. The conclusion he came to was just one word: Absolute!
Combined with the end-side AI assistant, the information the user wants can be filtered based on the user's preferences. This is the normal usage of information filtering and acquisition based on the assistant AI.

But if it is expanded and not limited to one user, it can also simulate a large number of users and be packaged into a tool.

That's not easy. We can directly detect the streaming algorithms of major platforms. Once they want to cause trouble, such as cooperating with foreign forces to create public opinion, we can detect the abnormality at the first time and even get the evidence directly.

"It's absolutely amazing! The algorithm is indeed amazing!" He Gang sighed, "But if we want official approval, it's hard to say whether it will work. Even if it does, it will take a very long time. Many problems are not as simple as you think."

"Mr. He, it doesn't matter whether the authorities take action or not. What matters is doing the right thing." Hao Cheng's thoughts were not simple at all. How could he not know that the current situation was not a problem of one party, but was the result of a combination of many factors?
But this is his third step: "At least they should think twice before attacking me again, right?"

Officially, Hao Ben didn't have too much hope. This step was first and foremost a deterrent - I have nuclear weapons in my hands, so you should be careful in what you do.

Then, with a little more time, Hao Cheng would be able to prepare a big gift for them.

"A big gift? What are you going to do?" He Gang was full of curiosity.

"Just wait and see!" Hao Cheng said with a smile.

"Okay, you..." He Gang was talking when he suddenly thought of something and said in surprise, "You didn't suddenly come up with this idea after that group of people visited you yesterday, did you?"

When Hao Cheng called me yesterday, he was still telling me about the algorithm. From the sound of his voice, it seems that he has a complete plan now. It can't be that he just came up with it at the last minute, right?

"Similar ideas have indeed existed for a long time, but if we talk about whether it was implemented, it was indeed implemented yesterday." Hao Cheng told the truth.

"I...I...Oh my God!" He Gang said "I" for a long time but couldn't think of a suitable adjective. Finally, he blurted out: "You have skills, but you are willful!"

He immediately contacted experts from the technical department to study this algorithm. To know how sophisticated this algorithm is, just look at the performance of the experts. I guess they are still immersed in it!

According to the experts, "It can be seen that this is just an ordinary algorithm, but its sophistication is unprecedented. It is not an exaggeration to say that if you hear the truth in the morning, you can die in the evening.

"But I really can't die now, because I still have Xiaosha. I can guarantee that Xiaosha's training algorithm and model are definitely more sophisticated than this.

"We can understand this thing after looking at it, it still belongs to the realm of the human world. But Xiaosha's thing, even if we imagine it, we still have no idea what it is."

And this algorithm that was highly praised by experts was actually created by Hao Cheng overnight in order to deal with public opinion attacks!

I’m afraid it’s impossible to imagine what level Xiaosha is at just by thinking about it!
(End of this chapter)