As for the result?
The answer obtained naturally made Jobs and Cook vomit blood, and made them as uncomfortable as if they had eaten a lump of unidentified yellow matter.
Because after the gathering of many R\u0026D giants from Apple to analyze the Geek Assistant.
Apple reluctantly discovered that the artificial intelligence Siri, which they had spent 7 years developing in the United States.
Was indeed inferior to the Geek Technology's artificial intelligence Geek Assistant, which was developed in 6 months!
How could they not vomit blood after learning this result?!
Of course, they naturally did not forget the predecessor of the Geek Assistant, the intelligent voice engine developed by Xunfei Company.
But even so, the research and development time did not exceed two years.
Therefore, how could the Apple staff not be so angry as to vomit blood after learning the truth?
However, the reason why Apple was so angry as to vomit blood was not only due to the problem of artificial intelligence Siri itself.
It also included other aspects of the problem, and it was those problems that limited the performance effect and ability limit of artificial intelligence Siri!
For example, the Geek Assistant is backed by the Geek chat software, which supports text chat and voice chat.
Therefore, when global users use it, they are actually accumulating useful training data for the Geek Assistant.
For example, if you want to have a powerful artificial intelligence, in addition to preparing excellent hardware equipment and artificial intelligence algorithms and other programs, the most important thing is training data.
For artificial intelligence voice recognition function, the best training data is the voice data when users actually use it.
Therefore, because the Geek Assistant is backed by Geek chat software, it naturally has a large amount of voice data for the Geek Assistant to train.
This is also the reason why the Geek Assistant has such a high recognition rate as soon as it comes up!
The second factor that made Apple vomit blood after this was Geek Encyclopedia and Geek search engine.
Since January, Geek Technology has been frantically buying data from other encyclopedia platforms, and it has also been frantically burning money, inviting users across the Internet to fill and maintain the Geek Encyclopedia database together.
Therefore, now, 6 months later, after burning hundreds of millions of dollars, Geek Encyclopedia currently has as many as 40 million knowledge entries, ranking first in the world.
This rich entry naturally allows Geek Technology to have a rich Geek Encyclopedia database, when users ask the Geek Assistant.
The Geek Assistant can preferentially retrieve the data that has been manually edited and reviewed from the Geek Encyclopedia database and directly answer the questions raised by users.
Of course, even so, it is impossible to answer all the questions of users.
At this time, another key weapon, the Geek intelligent search engine, should come into play.
In terms of this Geek intelligent search engine, Geek Technology is also very willing to burn money.
Therefore, the Geek intelligent search engine crawls a large amount of data on the Internet every moment and saves it.
These data crawled on the Internet are finally combined with the data of Geek Encyclopedia.
The superposition of each other makes the experience of the Geek Assistant excellent, and it can almost cope with the daily questions asked by users.
And the most important thing is that when the Geek Assistant answers questions, it will also include where the data comes from, and the key data sources are also specially marked in red.
This greatly proves the authenticity and security of the data, and it can be said that the Geek Assistant has suddenly become a practical Internet data search tool.
This is something that many mobile phone artificial intelligence assistants in the previous life could not do, after all, the mobile phone companies in the previous life were not willing to burn money like this, and they did not have this condition.
Of course, the shortcomings of the Geek Assistant cannot be said to be non-existent.
The first shortcoming is the GG problem, when users ask the Geek Assistant, accompanied by the appearance of relevant data.
That is, in the three data source locations below the phone screen, one or two locations are GG locations.
So when users ask the Geek Assistant, it is always inevitable for users to see the graphic GG with the 'GG label'.
Of course, users can also ignore these GGs.
This is like they used to use search engines to inevitably see GGs appearing on the data page, and then ignore them.
So when users use the Geek Assistant and ask it questions, it is inevitable to see GGs, which is an invisible shortcoming.
However, Geek Technology also has to eat, it is impossible to open a charity hall.
When users call the cloud artificial intelligence Geek Assistant, they have to use cloud computing power, which generates usage costs.
And just maintaining Geek Encyclopedia and Geek search engine is a considerable expense.
So the wool is naturally taken from the sheep, and the cost of supporting all this is naturally borne by users watching GGs.
As for what are the other shortcomings of artificial intelligence?
The second shortcoming is naturally that the supported languages are too few, such as Korean, Japanese, French, German, etc. are currently not supported.
These all need Geek Technology to slowly make up for, and ultimately allow the Geek Assistant to support more languages.
In addition, the third shortcoming is that the reaction speed of the Geek Assistant when answering user questions is still a bit slow.
After all, the current solid-state drives are still a bit too expensive, and Geek Technology naturally prefers to use cheap mechanical hard drives when storing a large amount of data.
The reading speed of this mechanical hard drive is naturally not as good as that of solid-state drives, and ultimately it is a little slow here, and a little slow there.
As a result, the Geek Assistant is not so fast when answering questions, and it is normal to take two or three seconds to answer a question.
Especially for overseas users, because the total amount of international optical fiber bandwidth is limited, and the amount of data transmitted at all times is so much.
At this time, it is natural to need to queue up and even review information, which also leads to signal delay problems.
So it is normal for the Geek Assistant to take three or four seconds to answer questions from users in the United States or Europe.
And this point can actually be solved by establishing Geek Encyclopedia data sub-libraries and Geek search engine data sub-libraries in Europe and the Americas.
But it also takes time to complete the construction of this data sub-library, it is impossible to solve it immediately, so the current Geek Assistant does have a certain answer delay problem.
But even so, with the perfect experience of the Geek Assistant, users in the United States and Europe are still willing to wait for these three or four seconds.
So Google discovered that it was inexplicably shot while lying down, its search engine business was doing well, but as a result, the business was snatched by the Geek Assistant.
And because asking questions to the Geek Assistant saves a lot of time compared to using its own search engine, the GGs seen by users are ultimately less than traditional search engines such as Google.
This also led to the global Geek smart phone system users very much like to use the Geek Assistant, treating the Geek Assistant as an intelligent search engine to use.
So Google vomit blood found that with the passage of time, the number of daily active users of the Google search engine showed a cliff-like decline!
This is really a disaster from the sky!
Therefore, Google really hates Geek Technology, and is forced to make an important decision.
