Start earning money as the girl next door

If you do well, you can generally get tens of thousands at this time.If you don't do it well, you may not be able to earn the cost of the server.

What's more, subscribing to genuine novel content also costs money.

Therefore, in the entire gray piracy industry chain, only a very small number of sources come from buying genuine novels and obtaining subscription novel content. Other piracy websites are like matryoshka dolls, grabbing a few sources that have been pirated earlier.

Therefore, it is often the case that the subscription account of a piracy source is blocked, a certain piracy website stops updating, and then a series of piracy websites stop updating.

The anti-piracy ban will increase the cost of the source piracy website.

If the risk is too great and even unprofitable, the pirated websites will actually move to other genuine novel websites, reducing their ability to compete with websites that are determined to fight against piracy.

For example, some later small and medium-sized websites have always insisted on technical anti-piracy, and the results are generally good.

Headline websites like cakes and pastries were once abandoned, thinking that the counterattack would never be completed anyway, so it was better not to counterattack. After all, the IP of the top works can make up for the losses caused by piracy.

The weak resistance of pastries has further led to piracy websites, which mainly focus on the content of pastries.

As for cakes, every time something goes wrong on the website, the author group feels dissatisfied.Not long after, Pastry came up with a trick - I am against piracy!

This is to give the author group a little sweetness and pretend that the previous unpleasantness never happened.

"By the way, boss! Our overseas authorized copyrights are limited by translation costs and have never been able to scale up. Can we study machine translation of novels? First translate by machine, and then let people polish it, so as to reduce some translation costs." Shen Chunhua said.

"Well, we can do some research!" Zhao Han thought for a while.

In fact, in addition to translation, the translation of literary works also needs to have a certain degree of literary quality, so that it can be readable after translation.

Otherwise, translating the meaning will be meaningless and useless.

There are some simple machine translations on the market, but the problem is that they are too rigid and cannot meet the translation needs of novels.

Not to mention novels, just subtitles for some animation films and TV shows. The translation level is not enough. Can simple mechanical translation be enough?Not satisfied!

In this regard, Zhao Han has some ideas, assuming that he refers to the later AI research and development ideas.

For example, let the machine translate it first, and if the human feels it is not good, then polish and modify it.

The machine will remember to imitate the results of manual polishing and include the manually polished content into the database.Later, when translating, multiple choices are given, allowing humans to choose the translation that best meets their requirements when using machine translation.

Through continuous algorithm statistics and gradual optimization, machine translation is becoming more and more similar to human translation.

Later on, the so-called generative AI, Zhao Han’s level of knowledge, was that he knew there was such a thing, but he couldn’t understand it.

But Zhao Han believes that machine translation can be combined with massive human labor.For example, for novels, machine translation is first used to translate the meaning.

Produce a machine-translated version, and then ask for manual translation. The parts that the machine cannot translate well will be manually revised and submitted.

The translation results of the modified sentences are all included. In the future, similar sentence translations will have a bunch of manually modified translation results for people to choose from.The most translated results are selected and their priority is increased.

To put it simply, what Zhao Han thinks of is "artificial" intelligence.It uses massive amounts of human labor to correct the shortcomings of machine translation.

Chapter 421 The idea of ​​cultural and entertainment works translation software

noon.

While cooking in the cafeteria, Zhao Han saw Chen Jingjing, thought for a moment and said, "Jingjing, to what extent can artificial intelligence be achieved now? For example, what about novel translation?"

Zhao Han actually has a limited level of concern for academic issues. He mainly knows that if translation tools want to do well, they must have some exposure to AI.

For example, iFlytek, which does relatively good translations, later also had the ability to build large AI models.

Of course, later on, a bunch of companies around the world made large-scale models, and many started to open source them to prove their own research.

This shows that the threshold for large model software is not too deep later.

The main threshold is hardware, such as Nvidia's GPU.and massive power resources.

Later, AI companies used NVIDIA's GPU training, mainly because NVIDIA self-developed a series of AI semi-finished technologies.Buying NVIDIA hardware gives you a series of technical added value.In addition, GPU training AI saves much more power than CPU.

Saving electricity bills was the main reason why GPU was later chosen for AI training. It was either that it could not be based on CPU, or that it could develop a hardware architecture by itself to train the so-called AI model.

The rush to use Lao Huang's GPU is essentially a panic buying of graphics cards, fearing that you will be late and not be able to catch up with the trend of AI.

But generally speaking, Zhao Han felt that that kind of mentality was terrible and was entirely Lao Huang's conspiracy.

In a sense, Zhao Han also quietly remembered the GPU.In the future, I will look for opportunities to build some GPUs. At least, there will be two rounds of popularity for GPUs. One is a short-lived bubble and a meaningless power-consuming mining wave. Mining is undoubtedly a stupid behavior of human machines. It wastes electricity to allocate the so-called Virtual currency does not contribute value, but consumes precious human resources.

However, during the virtual currency bubble, GPU manufacturers made a huge profit.

When the mining boom is over, many people think that Lao Huang is dead. Repent to the gamers and have a better attitude. As a result... the AI ​​​​trend is coming. In a sense, Lao Huang's luck, or the path of technological development According to his prediction, he can be regarded as the chosen one in the industry, and his luck lies with Lao Huang.

In this regard, Zhao Han is determined to cut off some of Hu's luck later.It is estimated that it will be difficult to surpass Lao Huang, but... at any rate, we will take advantage of the opportunity to make some arrangements in the hardware field. The main thing should be to cut off some of Lao Huang's opportunities!

"The artificial intelligence industry actually existed decades ago. It has been cold for several times. However, after the 21st century, after the performance of computer hardware improved, a hundred flowers bloomed. The mainstream theory is the simulated neural network theory. On this basis, derived Various technical routes. Several times in history, it was because of national-level investment in machine understanding of natural language that they all ended in failure. Yingjiang and Neon both used national investment and quickly entered artificial intelligence research after failure. Cold winter. Even now, there are some applications, but it is still a big problem for machines to understand natural language."

Chen Jingjing said, "In general, artificial intelligence is much better than that of the last century, but it is still very rudimentary. I think that in ten or twenty years, it may not be able to achieve the level of science fiction movies. As for machine translation, Some degree of artificial intelligence is used. However, with the current level of machine understanding of human natural language, it is difficult to achieve the level of human translation."

Zhao Han secretly thought that if the level of artificial intelligence in science fiction movies were to be measured, it would probably be too much in 20 years.

Zhao Han doesn’t know much about technology, but generally speaking, he understands the truth. When artificial intelligence begins to enter the so-called large model era, in fact, it has spread from the academic industry to many levels of productivity.

The sooner we accumulate technology in this area, the better.It's not very useful right now, but it will definitely improve the efficiency of various productions in the future.

Translation is the most common application.

Later, music, painting, and even code, etc., can already assist production and produce a lot of efficiency improvements.

After the Internet dividend came to an end, large models almost became another popular competition track for later major manufacturers.

If you wait until then to develop it, it will be too late!

Therefore, Zhao Han has the consciousness to accumulate some relevant technologies.

"What about literary translation?" Zhao Han said, "Similar to the Jingjing Pinyin input method, when you enter a word, different result options will be displayed in order.

If translation is made into a similar product, I translate novels and give 5 translation results for one sentence for users to choose.The first five sentences are all machine-provided translations.

If a translation user is dissatisfied with 5 translation results, the user can modify it himself and upload the translation results he is satisfied with, and the system will collect the user's translation results.

Then, the next time you encounter this sentence again, the translation results launched by the machine include both machine translation and user-modified translation.

Users can like and dislike all results.

Likes increase the ranking priority of translation results, giving priority to the results with the most likes from users.The downvoted results continue to lose priority, and this kind of translation result is not recommended.

We require the translators hired by the company to use this set of tools to translate and continuously improve the translation results.

In the future, as there are more translation users and constant corrections are made, the results of this translation software will increasingly show better translation results. "

"When you put it like that, it's not very artificial intelligence. It involves too much human participation, which is far away from machine learning." Chen Jingjing shook her head.

"It doesn't matter what method it is, as long as it is a useful tool." Zhao Han said with a smile, "A lot of labor costs are invested in the beginning. After perfecting the tools, a lot of labor costs can be saved in the future. If this kind of productivity software can be made, then, Domestic novels can be exported to the Yingjiang market on a large scale and earn overseas profits! Now, they cannot be exported on a large scale, mainly because the labor of translation is too expensive."

"Okay! Although, having said that, the cost of the research and development process may not be very low." Chen Jingjing thought for a while.

"It doesn't matter, the money spent on R&D will eventually yield huge returns." Zhao Han said with a smile, "If our translation tools are gradually iterated to the point where they are easy to use, do you think how big this market cake will be in the future?" ?Not just novels, but also countless translations of movies, TV shows, animations, games, etc., may also use our tools.

If this tool is done well, it can actually be exported to the whole world as a productivity tool! "

"Okay, it sounds like this project does have certain market potential." Chen Jingjing couldn't help but think about it, "The video website is almost finished. The next step is to make translation software."

"To what extent have video websites achieved?" Zhao Han asked curiously.

"Internal testing BUG stage." Chen Jingjing said.

Chapter 422: To make a big move into e-commerce, we need to wait for Taobao to make mistakes next year

"How long will it take to complete the video website?" Zhao Han asked.

"It should be online in July." Chen Jingjing said.

"Okay, we will notify Yu Weiwei when the time comes and we will cooperate with the launch of the domestic and eagle sauce markets." Zhao Han thought for a while and said, "It's time to collect the Internet video copyright."

Regarding the online entertainment video domain name and company, they have actually been registered.

It mainly uses the Wenjing Media Company registered in Hong Kong City as a platform to invest in two different branches in the Yingjiang region and the domestic region.

The online entertainment video company over at Yingjiang is trying to seize the fortune of Netflix. It pays because it is afraid of your market. The Yingjiang market is more mature and can achieve profits earlier.

Domestic online entertainment videos are mainly hoarding copyrights in a low-key manner.

The development models of the two markets are different, and the reason is very simple. The main reason is that the payment price of Eagle Sauce is expensive, so it is enough to offset the cost of servers and bandwidth.

The video industry in the domestic market is still immature, regardless of the advertising model of free video viewing.Still using the paid video payment model, the entire industry will not be able to achieve profitability in the next ten years or so.

The main reason in the early days was that the server bandwidth of video websites was very expensive. Neither the domestic Internet advertising traffic monetization model nor the payment model could offset the expensive bandwidth and server costs.

Later, the cost of bandwidth servers gradually decreased, but the copyright price rose to sky-high prices. At one time, the copyright purchase price of domestic film and television dramas was comparable to Netflix's copyright purchase price.

The revenue earned was not as much as that of Netflix, but the cost soared to the same level as Netflix. This is also the reason why domestic video websites continued to lose money for more than ten years after entering the genuine era.

While Zhao Han and Chen Jingjing were chatting in full swing, Chen Wenwen also walked in from the door of the cafeteria. When he saw the two of them, he smiled and said, "What are we talking about?"

"About translation software, it can translate novels. There are also things about video websites." Chen Jingjing summarized it briefly and comprehensively.

"Oh. I don't know much about technical things, but I have some opinions on e-commerce."

Chen Wenwen said enthusiastically, "The mall run by Jingjing Reading Network now earns about 3 yuan a day, and the profit is no less than [-] yuan. Since the novel website can add an e-commerce module, why can't it be added on Weibo? Add e-commerce to it? Or even add e-commerce to Jingjing Pinyin?"

"Well, regarding Jingjing Pinyin, don't be aggressive in adding more things for the time being. After all, it can easily be misunderstood by users as rogue software."

Zhao Han said, “It needs a slow process, and it must not be too radical. After all, it is just an input method. If too many things unrelated to the input method are added, users will be disgusted.

Now, there is already an advertisement on the user sign-in page, and you can get money for signing in every day to guide users to sign in.

By the way, it can have a drainage effect on Weibo.From now on, when the traffic on Weibo reaches a certain level, the traffic of Jingjing Pinyin Input Method will be directed to Jingjing Reading Network in turn.Even, Wenwen Service Network transports traffic. "

"Can Weibo join the mall?" Chen Wenwen asked expectantly.

"Yes, but not before the time." Zhao Han said with a smile, "Now that Weibo is engaged in e-commerce, it is easy to attract attention and arouse the vigilance of Taobao and other e-commerce websites. First, pretend to have no ambitions, just to make some cheap money. Advertising fees, paralyze Taobao. There are conditions for doing e-commerce. Our Weibo user traffic is much higher than Taobao. The gap is huge. Let’s test it a little bit.

You can start by finding some publishing houses and music companies to sell physical books, music albums, etc.We don’t seem to emphasize it ourselves and let nature take its course when entering the e-commerce field.

After initially having some users, using Weibo to purchase and receive goods, and initially establishing the impression that Weibo can be used for shopping.

Then, with lightning speed, we quickly increased our efforts to enter e-commerce, and quickly relied on our traffic advantage to make the e-commerce pie bigger.

Moreover, our best opportunity is to wait until Taobao makes a mistake! "

"Can Taobao make mistakes?" Chen Wenwen was a little unsure.

The main reason is that at this time, Taobao has surpassed eBay in terms of daily visit traffic, total number of products, and even daily transaction volume.

Probably by 2006, Taobao also surpassed eBay in registered users.

Since Taobao completely crushed eBay in 06, it felt that the three-year free agreement had arrived. Could it be that... it could be tested for monetization?

Although this attempt did not lead to death, it also caused turmoil in Taobao in the short term.

"You will definitely make mistakes! Taobao is not a fairy, and as the market continues to expand, it will gradually become lost and arrogant!" Zhao Han said with a smile, "Even if it is arrogant and foolish, it will make up for it. However, this actually gives a new opportunity to latecomers. a chance!"

"Is it because Yingying Literature made a mistake and left an opportunity for the online literature market to reshuffle?" Chen Jingjing thought for a while.

"It's probably like this!" Zhao Han couldn't help but smile, "If a person has a goal that he wants to surpass and has the potential to surpass, then he is the most efficient. But once he surpasses the goal and achieves first place, If there is no clearer upgrade target for the next step, then a short period of time may be the time window to make a mistake! Taobao made a mistake, probably because it has completely surpassed eBay. After its position is firmly established, decision-making will not be as cautious as before! "

Taobao made a mistake, that is, it came up with a plan to attract wealth without obtaining the consensus and support of a large number of store owners.

The so-called plan to attract wealth is nothing more than Taobao thinking that the three-year agreement with merchants has arrived, but it did not directly announce the commission, but launched a trial of spending money to buy advertising space.

This plan is much more restrained than later Taobao selling advertising spaces and express trains.

However, with the launch of the Wealth and Treasure Plan, some large merchants and companies can rely on their financial advantages to buy advertising positions.Once you buy advertising space, your sales will be high.After the sales volume is high, the products searched by users will naturally be ranked first, squeezing out the interests of small and medium-sized businesses.

Suddenly, a large number of small and medium-sized Taobao merchants went on strike to protest.

Some sellers even organized themselves and went to Taobao headquarters to protest.

As a result, Taobao's plan to attract wealth was hastily launched for a month, but ended without any success and was quickly removed from the shelves.

In addition to the protests from merchants, the bigger threat comes from Tencent’s Paipai.com.

On September 2005, 9, Tencent launched Paipai.com.

Around 2006, when Taobao launched the "Zhaocaijinbao" program to sell advertising space, a large number of merchants protested.

Paipai.com took the opportunity to win over a group of merchants, claiming that it would always be free and would not sell advertising space.

Taking advantage of a wave of merchants fleeing from Taobao, Paipai.com became the third largest C2C e-commerce platform after Taobao and eBay in less than a year after its establishment.

Chapter 423 The e-commerce industry has just begun

By around 2007, Goosechang's Paipai.com had 5000 million registered users, surpassed eBay in traffic, and ranked second in the industry.