Outlook on the Impact of XAI Supercomputers on the AI Industry

Recently, Elon Musk revealed to his investors that his artificial intelligence startup xAI is working on building a supercomputer. This supercomputer is designed to provide strong support for the next generation version of the artificial intelligence chatbot Grok for xAI. Musk hopes to complete the construction and operation of this supercomputer before the fall of 2025. He also mentioned that xAI may collaborate with Oracle to jointly develop this large-scale computer.

Supercomputers, also known as high-performance computers or supercomputers, refer to a series of computers with super powerful functions designed specifically for scientific computing, engineering applications, and commercial applications that require computing large amounts of data at ultra fast speeds. Supercomputers have a wide range of applications, including scientific research such as planetary simulation, new material development, molecular drug design, gene analysis, big data processing, intelligence collection, information technology, and medical fields. With their powerful computing and processing capabilities, they play an important role in many fields, promoting the progress of science and technology and the rapid development of society.

According to the 61st issue of the Global 500 Supercomputers released in 2023, Frontier at Oak Ridge National Laboratory in the United States is the only billion dollar secondary machine on Earth, with an HPL score of approximately 17%, up from 1.02 Eflop/s in November 2022 to 1.194 Eflop/s.

XAI is a high-tech company founded by Musk in 2023, with the intention of competing with OpenAI supported by Microsoft and Google under Alphabet in the field of artificial intelligence. And Musk is also a co-founder of OpenAI.

It is reported that Musk's artificial intelligence startup xAI plans to complete a round of financing in June, including the recently acquired cash. The company's valuation may exceed $24 billion, so its plan to develop supercomputers has no obstacles or problems in terms of funding.

Regarding the original intention of building this supercomputer, Maersk firmly believes that "the safest thing for AI is to seek the truth as much as possible", and emphasizes that xAI will be committed to this goal, striving to establish an AI that is beneficial to humanity and avoiding training "super intelligent liars". Mask repeatedly emphasized that "honesty is the best policy". He also stated that xAI is committed to this, promising to always release a model that goes beyond the current open source model to improve transparency and promote healthy competition.

As is well known, AI, as an important industry trend that has attracted worldwide attention, has many potential risks and hidden dangers. Industry researchers claim that fundamentally speaking, it is currently impossible to train an AI model that cannot be deceived in all possible situations. In other words, lying and deception are one of the most serious problems and challenges facing the development of the AI industry.

On May 11th this year, a research team from the Massachusetts Institute of Technology in the United States published a research result in the Science Journal of Patterns, stating that some artificial intelligence systems have systematically learned how to manipulate others through learned deception. In addition, a recently published research report shows that when under pressure in a simulated work environment, artificial intelligence chatbots trained to be helpful, harmless, and honest actually become deceptive.

Currently, many industry professionals around the world have expressed concern and concern about the serious consequences of deception and lies that may arise during the development of artificial intelligence. However, there are not many industry giants like Mask who hold a firm attitude and publicly promise to develop honest AI systems. I hope that industry professionals around the world can maintain an open and inclusive attitude, take appropriate measures to actively address the problems that arise during the development of AI systems, and explore them in depth with scientific methods and solve them properly.