Sihai network

The artifact of one key generating a thousand character long text is on fire and laughs after the ac

After reading the full text, I just want to say that if it's really "bullshit"...... learn python, you can also design similar programs?

I believe that everyone, like Xiaobian, would like to know what is sacred to develop this program.

Originally, the program was just to help a student who was difficult to apply for a six thousand word withdrawal from the student union. Unexpectedly, it was broken by netizens. So what is the generation principle of this program?

In short, the whole article comes from the fixed database arrangement and combination, and then uses the theme you give as the key word, and replaces it in a timely manner, and then adds the famous sayings of celebrities, which can be improved by forcing the case, not only achieving the goal of "no nonsense", but also asking questions everywhere.

The database includes argumentation, famous sayings and introduction. Each sentence ends with "this does not forbid me to ponder". These inanimate sentences, together with simple algorithms, form a gorgeous article. It's really "gathering is nonsense, scattering is all over the sky".

In addition, the developer also admits that there are many loopholes in this program. In order to prevent the content of the article from being too repetitive, the next step is to add more verbose words, as well as the content in Masan's "meeting fans". At the same time, in order to meet the needs of handwritten articles, the next step is to add the function of handwritten direct rendering of pictures. Netizens also have different opinions on "nonsense article generator". It's interesting to know that there is a discussion on this topic.

Is there any comparison between GPT and dogshit article generator?

The answer is No.

The reason is that the "GPT generator" is used to generate some Chinese text quickly for GUI development test text rendering, while the "GPT" text generator aims to generate high-quality coherent text, "GPT" is characterized by high speed, slow speed and high quality of the model, which does not meet the design requirements of the "GPT generator".

The implication is that the difficulty of the two algorithms is different, but they can't be used in the same day. The article generator is used for entertainment, and it's not appropriate to use it in work and life.

He's not the only one who thinks about it

Lei Feng learned that there is a hidden function in Microsoft word called Rand trick.

As long as you enter = rand () in the new page of word, and then enter two numbers in brackets at will, separated by [,], word will automatically generate a paragraph of text.

For example, if you type = rand (6,6), the following text will appear automatically. In addition, someone has set up a website.

On the web page, as long as you enter the author's name, the computer can automatically generate a "SCI level" computer science paper for you. In the paper, abstract, introduction, literal review, experience result, discussion, conclusion, reference, diagram are all available. Except that no one can really understand the content, "paper" is just like a picture. In addition to the web version, there are also apps.

The same is true of a popular online novel generator app in the past few years. Through this software, users only need to input the basic elements into the software, and then they can get a complete network novel. The software classifies the materials into several categories, including outline design, character setting, dialogue description assistance, Xianxia skill generator, sect power generator and so on, which is also enough.

After these are filled in, the software can automatically write a novel according to the traditional pattern of network novel, and also can automatically generate the corresponding dialogue according to the character. You are the only one who can't do it without them. But the content of both the nonsense generator and the RAND trick in word is based on the random call of the existing content library, which is totally different from the AI algorithm.

Therefore, Lei Feng's editor reminds everyone here that it's impossible to be lazy. It's Wang Dao who writes well.