As for whether the internet companies of his previous life would purchase data for their generative AI?
The answer is both yes and no.
Because the price of that data was too expensive, and the data was still being generated in real-time, in the long run, no one could afford to burn that much money, to pay that much money.
Therefore, each internet company still chose to do its own thing, which ultimately led to the creation of data silos. These data silos limited the upper limit of the development of generative AI in China, limiting its intelligence and breadth of knowledge.
This also led to generative AI in China sometimes giving answers that made people laugh because they lacked common sense.
Conversely, the situation was different overseas. The closed-loop application ecosystem overseas was not as severe as in China, at least there were no restrictions on information crawling.
Ultimately, the amount of 'high-quality text data' overseas far exceeded that of China, which also made the AI of overseas technology companies incredibly intelligent.
It's like you've only read a hundred books, how can you compare to someone who has read ten thousand books?
The breadth of knowledge is completely incomparable.
Thinking of this, Gao Nian thought for a moment and said:
"Lei Bu Si."
"Hmm?"
Lei Bu Si turned his head to look at Gao Nian, not knowing why Gao Nian suddenly frowned deeply.
"The essence of artificial intelligence is algorithms, but more important than algorithms is data!"
"Data?"
Hearing Gao Nian's words, the people present all showed surprised expressions.
They knew that the essence of artificial intelligence was algorithms, and it could even be said that the computer industry was built on algorithms.
But the theory that data was more important than algorithms was the first time they had heard of it, and they didn't expect data to be so important.
However, whether the people present knew it or not was not very important, as long as they followed Gao Nian's orders to execute it.
So Gao Nian continued in the surprised gazes of the people:
"Next, we need to do a good job in data storage. We need to use the advantages of Gui Province to establish a data storage center in Gui Province.
This data storage center will mainly store text and voice data generated in real-time by users when playing games and using Geek chat software and other products."
Hearing this, many people present frowned deeply, and CFO He Tong even frowned on the spot and said:
"Boss, the cost of storing these daily chat text data and voice data is too high. Users are generating information every moment.
Our product range is global, there are so many netizens around the world, and the chat data and voice data generated every moment is quite amazing.
If we want to store it, and store it for a long time, the cost will be too great.
Is this data really useful?"
However, facing He Tong's question, Gao Nian not only did not retreat, but slightly shook his head and said:
"It's useful, and it can even be said that this data is the best wealth."
The details of life are reflected in all aspects, and whether generative AI is intelligent is reflected in these aspects. After a pause, Gao Nian continued:
"In addition to the permanent storage of text and voice data of our own internet products.
Next, we also need to establish our own encyclopedia, to build a Geek Encyclopedia with the most world information.
In order to speed up the growth of 'Geek Encyclopedia', we can spend money to buy materials and articles that Wikipedia or QianDu Encyclopedia have already compiled, and fill them into our own encyclopedia.
In addition, in the next two years, we will also take out at least 200 million US dollars as an incentive, so that domestic and foreign users,
Strive to compile reasonable and qualified encyclopedia articles.
In short, whether it is big or small things, or trivial things or current affairs in daily life,
Or common sense questions, we must compile articles.
We want to establish the most detailed encyclopedia, detailed to decades or hundreds of years ago, how much were the wages of domestic and foreign workers, how much were the prices, etc., all must be compiled and recorded.
Therefore, my goal is to have Geek Encyclopedia have at least 100 million encyclopedia articles within two years!"
"Hiss!"
Hearing Gao Nian's words, the people present instantly gasped, not expecting Gao Nian to make such a big deal.
Because if Geek Technology really does this, it would not be an exaggeration to spend several hundred million US dollars in the next two years.
Burning several hundred million US dollars in two years just to create an encyclopedia with 100 million articles, is the cost too high?
However, Gao Nian's words did not stop. Gao Nian continued when people were surprised:
"In addition, we also need to acquire well-known forums at home and abroad such as Tianya and Mao Pu.
Then transform them all into internet products similar to Tieba, and then all the posts and replies posted by users on them must be saved.
As for those that cannot be acquired, we will directly adopt the form of search engine crawlers to crawl and save their data materials.
In addition, we also need to collect and save the data of domestic and foreign internet news websites, thesis websites, and digital libraries.
In short, we must find ways to save all internet text information to form an extremely large digital database."
After thinking for a while, Gao Nian then said:
"Simply crawling data directly may cause media controversy, so we need to establish a search engine business and develop the Geek Search Engine.
In this way, we can not only maintain our own operations through the GG profit of the search engine, but also reduce the economic pressure of storing data.
The development of the search engine will be the responsibility of Li Jun and Ni Guanghai. If you need to poach talent, poach talent, and if you need to buy technology patents, buy technology patents."
Hearing Gao Nian's words, the people present frowned deeply.
Because they did not understand what was the use of spending so much money to collect and store these text and voice data.
After all, QianDu Search and Bone Song Search would not store data so crazily. Even if they store data, they store it conditionally.
It's almost too exaggerated for Geek Technology to save even the chat data in games.
Although Gao Nian had previously explained that this data is the key to the power of artificial intelligence, is this data really useful, is it really worth investing so much, and can it be recovered in the future?
"Gao Nian, is it really useful to spend so much money to store this data? Can we really get our money back?"
Lei Bu Si couldn't help but frown and ask Gao Nian.
After all, he is also a major shareholder of the company, and he must ask about this kind of behavior that wastes a lot of the company's money.
"Of course it's useful, and it's very useful, because these are all intangible assets.
Whether the artificial intelligence we launch in the future is powerful or not, intelligent or not, depends on this seemingly garbage data.
1
The biggest feature of generative AI is that it requires a lot of data, and the more data it has, the smarter it will be.
For example, the chat data of Geek chat software, although there will definitely be a lot of garbage data in it.
But their chat topics are actually all aspects of life or current affairs.
Among these, the more artificial intelligence sees and learns, the more it will understand the world.
Ultimately, when generative AI answers questions, not only will the answers be more accurate and there will be no fabricated phenomena, but it can also become more intelligent.
This is why it is necessary to save these "garbage data" that should be deleted.
