Musk brings start-ups back to the artificial intelligence track, and top players are generously raising concerns about AI monopoly

Musk brings start-ups back to the artificial intelligence track, and top players are generously raising concerns about AI monopoly

Musk brought the start-up xAI back to the artificial intelligence (AI) track, adding another “boom” to the already boiling competition in the artificial intelligence industry. After OpenAI successfully launched the artificial intelligence chat application ChatGPT and obtained huge financing, American technology giants such as Microsoft, Google, and Apple are investing in artificial intelligence technology with great fanfare. The fierce “involution” of a small number of top players has raised questions about the monopoly of artificial intelligence giants. Musk said that the creation of xAI is to prevent the emergence of a “dominant” situation in the field of artificial intelligence.

　　Major giants accelerate their involvement in competition

Not only Musk, but also American technology giants are targeting OpenAI. Microsoft, Google, Amazon, etc. have invested in artificial intelligence technology with great fanfare. According to a report by the US technology website The Verge on the 19th, Meta has made its large language model LLaMA 2 open source, and commercial and research institutions can use it for free. This move is intended to compete directly with OpenAI’s free artificial intelligence applications. At the same time, the US chip giant Qualcomm announced that it will cooperate with Meta to provide LLaMa support for notebooks, mobile phones and headsets from 2024, so that artificial intelligence applications do not need cloud service support.

Other giants are also making strides. A generative artificial intelligence from Microsoft has broken ChatGPT’s speed record. Last month, Google released a set of artificial intelligence tools, including email, spreadsheets and various texts. Apple, which was previously considered to be slightly behind, was also reported by Bloomberg on the 19th that it was developing “Ajax” artificial intelligence technology and creating a chat robot “Apple GPT”. Apple shares jumped as much as 2.3 percent to a record $198.23 after the news was released.

“Several technology giants in the United States are currently competing fiercely around artificial intelligence.” Shen Yang, a professor at Tsinghua University and director of the Metaverse Culture Laboratory of the School of Journalism, analyzed to the Global Times reporter that the competition for large artificial intelligence models is divided into several levels of competition. The first level is the chip level, such as the cooperation between Qualcomm and Meta; the second level is the operating system level, which mainly involves Microsoft, Google and Apple. Microsoft has both invested in OpenAI and cooperated with Meta. Software competition, including social media applications, and office software applications; there is also competition at the content level.

　　”Risk of being monopolized by two or three entities”

Some analysts have noticed that the head competition in the artificial intelligence industry has become an internal game of Silicon Valley giants. Musk’s newly established xAI will recruit people with industry experience from Google, Microsoft, OpenAI and Tesla, which can not help but raise concerns that the future top AI talents will all come from the same family.

“The monopoly of technology giants is a bad thing, and the monopoly of artificial intelligence giants will be even worse.” The New York Times published an analysis article on this topic this month, arguing that a few leading companies such as OpenAI, Microsoft, and Google have a large lead in “shaping the future dominated by artificial intelligence”, which is not good news.

Currently, Google and Microsoft are ahead of other competitors, the report said. Given the enormous harm AI could do to jobs, privacy, and cybersecurity, the progress these companies have made in AI without externally enforced protections is worrisome. Beginning in the 1990s, the Internet provided a lower-cost way of expressing opinions. But over time, the channels of communication became concentrated in the hands of a few, including Facebook.

When Kyle Koschick, an employee of xAI, participated in a dialogue on social media, he bluntly stated that artificial intelligence has already emerged as a monopoly. Kyle Kosik previously worked for OpenAI. He believes that the biggest risk facing artificial intelligence is that the entire industry is “monopolized by two or three entities.” The way to solve this problem is to introduce competition.

“The benefits of competition in the field of artificial intelligence mainly include the following points: first, multiple institutions seek different scientific research development paths, which will promote the diversification of science and technology; second, multiple institutions can promote each other in the competition, and promote each institution to continuously strengthen their scientific and technological capabilities. At the same time, they can also supervise each other, and bottlenecks can be discovered in time; third, through competition, it can accelerate the finding of efficient artificial intelligence commercialization paths.” Kyle Kosik analyzed.

Artificial intelligence model training requires massive capital investment, which leads to limited institutions that have the strength to join the artificial intelligence track. TrendForce, an international high-tech industry research institution, analyzed in its report that OpenAI needs the computing power of tens of thousands of Nvidia A100 chips when training artificial intelligence products, and the price of this high-end chip is about 10,000 US dollars per chip, and the supply is in short supply.

Guosheng Securities estimated in an artificial intelligence industry analysis report that the cost of ChatGPT training is about 1.4 million U.S. dollars, and for some larger and more advanced artificial intelligence models, the training cost is between 2 million and 12 million U.S. dollars. At the beginning of this year, an average of 13 million unique visitors used ChatGPT data every day, and the corresponding chip demand was more than 30,000 A100 chips. The comprehensive initial investment of ChatGPT is about 800 million US dollars. If the cost of electricity is calculated, the amount of investment will be further expanded. In addition, cloud services are also an important area of ​​”burning money” for artificial intelligence.

In addition to software and hardware, OpenAI also has to bear a lot of expenses in the daily operation of the enterprise and talent recruitment. The annual salary of an ordinary software engineer in the company is between 200,000 and 370,000 US dollars, and the supervisor level is between 300,000 and 500,000 US dollars.

On Quora, an online question-and-answer platform in the United States, many people in the industry are analyzing the dangers of artificial intelligence monopoly. They revealed that not long after OpenAI’s ChatGPT started to catch fire in February this year, there was a short-term service disconnection, which made users around the world unable to use it. Some analysts believe that if artificial intelligence services are highly concentrated in one or two companies in the future, once the services of monopoly companies stop or go offline, it will have a catastrophic impact on the world.

　　Musk’s “ambition” has also been questioned

Shen Yang analyzed to the Global Times that Musk’s re-entry is good for the artificial intelligence industry, but the real purpose may not be as noble as “opposing the monopoly of giants”. Musk has a strong economic motivation for establishing xAI, otherwise it would be impossible to maintain a cutting-edge team. According to US media, Musk proposed a very attractive economic return for xAI recruiting talents. If Musk’s $20 billion valuation of xAI is finally realized, the R&D personnel who are now employed can earn hundreds of millions of dollars.

Some analysts believe that when Musk founded or took over a company, it often stemmed from his dissatisfaction with the operation of the company in this field. An early investor in OpenAI, Musk first forayed into artificial intelligence because he believed Google co-founder Larry Page “didn’t take AI safety seriously enough.” However, when OpenAI made a series of breakthroughs, before launching ChatGPT in 2022, Musk chose to leave OpenAI on the grounds that OpenAI had become “greedy for profits.” He also believes that, as public companies, both Google and Microsoft are subject to external directives on corporate decision-making.

Musk declared the newly formed xAI to be a privately traded company. Some netizens suggested on social media that Musk already owns a series of technology companies such as Space Exploration Technology Company, Tesla, Twitter, and brain-computer interface. With the announcement of xAI, how much business ambition does he have?

According to US media reports, Musk’s business landscape is once again expanding, from cars to space, from satellites to Twitter, and now to artificial intelligence. xAI will use Twitter data to train artificial intelligence systems and products, and the new artificial intelligence will cooperate with Tesla on the “chip frontier” and “artificial intelligence software frontier”.