Reborn 2005: The Internet Myth

The party in the evening was a great success and the atmosphere was very lively.

Business white-collar workers have put aside their usual rigor and solemnity, and programmers are not as unsociable as they imagined.

The four people who joined Ultron are also very popular. They all work for the same boss and have different business scopes. Cross-collaboration will be unavoidable in the future. It will be beneficial to both parties to get familiar with them in advance.

Wang Zheng was sitting in the corner with a bottle of beer, drinking slowly and leisurely. He had fried three large pots of crayfish tonight... He was really a little tired.

Cheryl came over and sat next to him with a glass of cocktail, and said, "Justin, how do you feel?"

He picked up the wine bottle and touched it lightly with Cheryl: "Very good. Thank you Sherry, it's you who made Meta grow so fast."

Cheryl took a sip lightly and said with a smile: "It is the credit of everyone here, especially you, otherwise we would not have such a relaxing and enjoyable job."

Wang Zheng was modest: "I am just more favored by fate."

Cheryl was stunned for a moment and said: "It's almost time, Justin, let's make a toast."

Wang Zheng nodded, picked up his beer and walked to the center of the venue, shouting "Toast". The music faded and people's eyes gathered to look at their boss.

Wang Zheng looked around and finally raised his beer and shouted: "To create a miracle! Cheers!"

More than thirty people shouted "Create a miracle" in unison, laughing and clinking glasses with the people next to them.

Around 10 o'clock, the party officially ended.

The employees spontaneously packed up all the party items and said goodbye and went home.

Wang Zheng took a shower and lay on the bed thinking about it. In fact, he felt quite accomplished.

But the more he blended in, the more he missed his hometown and classmate Ye Zi.

I couldn't sleep, so I simply got up and turned on my computer, logged in to the domestic Huoyun server, sent a message to Ye Zi, and handled today's work.

Then continue learning.

After copying the browser, Wang Zheng has become very proficient in the Cangjie language.

But being proficient doesn’t mean you can write whatever you want.

People who have not received systematic training may not be able to understand why it is written in this way if the source code is given to them.

There is a logical problem here.

For example, Wang Zheng knows about artificial intelligence, blockchain, and deep learning algorithms. Who doesn’t know something about this 20 years later?

But he couldn't write it - he didn't even know how to implement it in code.

I really just don’t know about this thing, so I don’t even know where to start.

He can write blindly and randomly, let alone artificial intelligence. He can't write anything like the artificially retarded version of Siri in China.

There is growth, at least after the period of browser imitation, he finally understood the relationship between the so-called algorithm, code, and program.

The algorithm is the steps to solve the problem, the code is the pen used to solve the problem, the program is the result, and the programming language is equivalent to the piece of paper.

His current situation is equivalent to already knowing the answer. He has paper and pen but no steps, so he cannot solve the problem.

However, Wang Zheng never thought about becoming a fat man in one bite. It would be very difficult for him to be able to read and understand a high-level language without even finishing high school and relying on self-taught English.

Not to mention that his junior high school level mathematics would not be able to teach even an elementary school child twenty years later.

If you can’t learn it, can you still copy it?

Regardless of the steps, copying 100 points is better than anything else.

Now, whether it is Huoyun or Meta, there is an urgent need for image recognition technology and image compression technology, including video content recognition technology and video compression technology.

The purpose of image recognition and video recognition is to reduce the content review workload of Meta community and Xiaoyunshu community websites, while also improving a better experience for users.

At present, the American League is a little better, not so exaggerated. But in China...it's hard to describe, even bikinis have to be blocked.

We underestimate the psychological endurance of the people of Daxia.

Although the community does not have the video playback function yet, this will happen sooner or later. It is necessary to prepare for a rainy day.

Pictures are okay, but video recognition is really hard to solve. Wang Zheng still remembers a piece of news, probably from the Douyin platform, which had a video review team of about 2,000 people.

The cost in China is okay, but in the United States... it's just too expensive.

There is also image and video compression technology, which is also very important.

The clearer the picture, the larger the volume it will occupy and the slower it will load. In a community like Meta, the more pictures there are, the more storage will be taken up, which is a huge cost.

Not to mention that if videos are uploaded in the future, the amount of storage required will be astronomical.

It is unrealistic for Wang Zheng to solve the problem by himself, but he is willing to do it.

The software in the USB flash drive plus Cangjie compiler can help him directly get the steps to solve the problem.

Wang Zheng opened the USB flash drive and found the software "Yunwang Internet Cafe Management System".

There is a program here, which is the face comparison module, which Wang Zheng originally bought.

Face comparison is already an advanced technology compared to photo comparison.

If you can compare faces, can you compare photos?

So he dragged the face comparison module into the Cangjie development kit. Soon, the program running virtual machine was on the left, and the source code converted to Cangjie language was on the right.

No matter what language the program is written in, as long as it can ultimately be recognized as machine language, the computer can understand and execute it.

Theoretically, Cangjie language can run under win system.

However, Wang Zheng has not yet started to learn the format of the win system, but this has little impact.

Hang this program on a server running Hongmeng system, and ask the web server to transmit the received uploaded photos to this server. After comparison by the program, those that pass will be returned and released, and those that fail will be rejected.

This idea is still reliable.

As for the amount of data being too large, just build more high-performance servers.

Wang Zheng was not lazy and took the source code directly. After all, these things are external things, and knowledge is his own.

Learn more, you won't lose anything.

It's still imitation, but much faster.

And with this imitation, Wang Zheng's understanding of what kind of logic programming actually is will become more and more thorough.

There are more than 300,000 lines of code, and the more Wang Zheng writes, the clearer he becomes.

Which codes analyze face positions, which codes extract facial features, which codes are used for image comparison, and which codes are used to output results.

These alone are enough to benefit Wang Zheng endlessly.

Not to mention anything else, each of these items can generate a patent. A combination of two or three, or a combination of three or three, can result in at least twenty or thirty patents being applied for.

As for facial recognition technology, the patent authorization alone is enough to last a lifetime.

Moreover, with the wisdom of future generations in hand, he no longer has to reinvent the wheel. When developing new technologies, he can apply them directly when they are needed, saving a lot of time.

I continued to work until early in the morning, but my liver couldn't move anymore, so I turned off the monitor and fell asleep smoothly.

There are many advanced algorithms waiting for him later.