Artificial intelligence is the greatest invention of our time, but it would not have been without a strong infrastructure through which it would work. Imagine that you are following the construction of a skyscraper for several months or even years, you see only a large hole in the ground and it does not seem to be any progress, then suddenly, the iron structure appears and the columns are raised and the walls are built shortly.
It is quite similar in the world of artificial intelligence, there are basic parts built in the background for years or even contracts, and they are not visible to users, but they are necessary for artificial intelligence products properly.
According to a report published by “Business Insider”, Google has most of this infrastructure, and major companies are like Microsoft AndDead AndAmazon Some of them have, as for “”OPII“It works at an accelerated pace for its construction, and it still has a long way to cut it, and for Apple, it lacks a great infrastructure that occupies its own intelligence, which is a great challenge to its future, especially with regard to iPhone phones.
It is worth noting that these technical problems may not be always clear because it occurs behind the scenes, but this year it has emerged remarkably when Apple was forced to postpone the great update of its smart assistant “Siri”, the company is trying to change “Siri” from its roots to match the era of obstetric intelligence, but it was not ready.
The reform of “Siri” is required in the form required to create an infrastructure from scratch, and if it does not succeed in that, you may have to rely on other competing companies, or enter into high acquisition deals.
How did Google benefit from its infrastructure to develop artificial intelligence?
Need Apple To a lot of basic ingredients for developing artificial intelligence techniques, but to understand the image more clearly, it is sufficient to take a look at what Google has built over the past decades to be ready to keep pace with the progress of artificial intelligence.
For example, it was released Google Recently, the “Flow” tool, which is an artificial intelligence tool that helps creators to produce professional videos easily, but behind this simple tool, there is a huge and complex technical structure developed by Google for years, and the “Veo” model – dedicated to generating the video with artificial intelligence, is now in its third edition – it would not have had no YouTube videos that trained On it. Who owns YouTube? Yes, Google.
Google also has the “Gimenai” model, the rival of “Chat GBT”, as well as the “IMAGEN” image to generate images in its fourth edition.
If we search for the origin of artificial intelligence, we will find that the transformer technology created by Google in 2017 was the basic stone that made artificial intelligence possible, in addition to Tensor processors, which are advanced segments of Google.
But the story does not end here. Google has a treasure of data, as it sipped the contents of the Internet for decades, and it collects huge amounts of information from multiple sources, and all these data are used to train artificial intelligence models to become more powerful and accurate.
Google founder reveals artificial intelligence 25 years ago
In an interview conducted in 2000, Larry Paige, co -founder of Google, spoke about his vision of artificial intelligence, and said that “AI will be the final version of Google, who understands exactly what you want and gives you the right answer.”
He stated that Google had about 6 thousand computer to store about 100 copies of the Internet, and Paige said, “A lot of computing, and a lot of data that were not available before, and from a geometric and scientific point of view, building things to benefit from this is an interesting intellectual exercise, so I expected to be busy for some time.”
It seems that when Google became a public company as a search engine in 2004, it was already a company specialized in artificial intelligence.
In 2012, the researchers achieved a big boom in the field of artificial intelligence when they trained computers to identify and classify things by “watching them”, as Alex Creisiezky, Ele Sutskifer and Jeffrey Hinton – whose advisor at the University of Toronto – Alexnet technology, founded the company “AlexNet”, and established the company “DNN Reservation”. (DNNRESEARCH) (DNNRESEARCH) (DNNRESEARCH), acquired in 2013, acquired all intellectual property rights, including its source code.
In 2014, Google bought DeepMind, a secret artificial intelligence laboratory run by Dimis Hasbis and Mustafa Suleiman, and this step inspired Elon Musk to launch the AI ”Oben AI” as a opposite force of Google in the field of artificial intelligence.
This has been more than a decade, and now “Deep Mind”, led by Hasbis, has become an essential part of Google, and leads many important artificial intelligence projects, while Suleiman runs large artificial intelligence projects in Microsoft.
Before the “Google I/or” conference in 2016, the company used a group of journalists to introduce them to the concept of “learning the machine” – which is a branch of artificial intelligence – and Hinton and other artificial intelligence pioneers spent hours and explain how this complex technology works, and in the same year the company revealed “TPU”, which is a series of artificial intelligence segments that developed internally Graphics processing units from Invidia.
Google uses these segments in its private data centers and rents them to other companies and developers through its cloud service, and has also developed an artificial intelligence framework called “Tensorflow” to support the machine learning developers.
Google is planning to establish 75 billion dollars in artificial intelligence centers this year, and to operate all these centers, Google recently entered into deals to develop 3 nuclear stations to provide sufficient energy that occupy their positions.
Apple’s situation is difficult
In 2018, it seemed as if the efforts of Apple in the field of artificial intelligence began heading towards the right path. At the beginning of that year, Kraij Vedirigi, head of the software department in Apple, gathered his senior employees and announced the appointment of a new prominent figure, as the company attracted John Jianandria from Google to be the head of its artificial intelligence department, who was running the groups of research and artificial intelligence in Google, according to a report from “Bloomberg”.
However, Apple still lacks many basic elements, the most important of which is the infrastructure, as it does not run many large data centers, and uses Google facilities for important projects, when Apple users make backup copies of “Aqualoud”, these copies are often stored in Google Data Data centers, and with regard to training artificial intelligence models from Apple that which Apple Apple Interte is busy, as Apple has requested additional access to Google TB Yu slices to carry out exercises for models.
It is noteworthy that Apple has a lot of data, but it is careful in using it to develop artificial intelligence due to concerns related to user privacy, and it has tried to process artificial intelligence on devices such as iPhone, but these projects need a huge computing power that only data centers provide.
On the other hand, Apple was slow to attract talents in the field of artificial intelligence, and it did not allow researchers to publicly publish research papers or imposed restrictions on this, as this was an essential element in attracting vital talents for many years.