Musk's latest AI entrepreneurial live broadcast record: I hope to use AI to solve the unsolved mysteries of the universe

**Source:**Tencent Technology

At 4 a.m. on July 15th, Musk’s artificial intelligence company xAI held a live voice event on Twitter. Musk and his team appeared, discussed the mission and goals of xAI, and answered user questions on the spot.

On July 13, Musk announced on Twitter that his artificial intelligence company xAI was officially established, with the goal of understanding “the true nature of the universe.” xAI is separate from Company X, Twitter’s current parent company, but will work closely with Company X and Tesla.

In this live broadcast, in addition to Musk himself, there were also many core team members of the company present. According to the official website, members of the xAI team have worked in cutting-edge companies such as DeepMind, OpenAI, Google Research, Microsoft Research, and Tesla.

Musk talked about xAI’s future exploration direction, data security and protection, AI application fields, and hopes to add more human elements to the exploration of aerospace.

Musk said that the purpose of establishing xAI is to understand the world and the universe. The safest way to build AI is to make AI full of interest and curiosity about the world, and to empower AI with various technologies. This is what we can do choose.

Regarding the goal of “understanding the nature of the universe” that has received the most attention from xAI, Musk said that we have understood a lot of the universe now, but we should also combine the universe with reality. There are still many unsolved problems in the universe. Mysteries, these are the keys to our work. Human beings have many unsolved mysteries, and we hope to solve these problems through the truly large-scale data computing capabilities of AI. This is the future direction of AGI.

The following is the real record of xAI’s live broadcast:

Musk: Let’s start now. First a brief introduction. Let me introduce the whole team, some of the work we are doing now, and some work focuses that we need to focus on first, and the direction of xAI’s next step. The purpose of establishing xAI is to understand the world and the universe. I think the safest way to build AI is to make AI full of interest and curiosity about the world, and to empower AI with various technologies. This is a choice we can make .

I am also constantly optimizing the use of AI, allowing AI to distinguish what is real information and what is false information. So stay curious and keep exploring, this is the safest way for us to use AI. I think this is exploring artificial intelligence in a human way, which is much more interesting than exploring artificial intelligence in a non-human way.

I would like to add all the elements, such as spaceships and astronauts, together, they are not as interesting as AI. First of all, let’s break down the barriers. The first goal of exploration is Mars. As a planet, Mars is not as interesting as Earth. The way we explore Mars right now is with AI. I think this is also very interesting, and it is also the direction of AI’s future work to meet people’s expectations for Mars.

I myself have considered the data security brought about by AI many years ago, and what the real future direction of AI will be. I have thought about it day and night for a long time, and now I have decided to use such a company to truly realize the future direction of AI, which is also related to the safety of AI. We need to use industrialized operations to realize the security of AI, so that all some data can get good data security and data protection, and generate real meaning.

We also thought about the acquisition of AI, how we can get real benefits from AI. We can use AI to ensure the safety of more satellites. This is the real AI application field we want now. If we can add more human elements to aerospace exploration, we will be able to explore some places that humans have not explored. I think that’s also a metaphor that people might understand, and that’s what I’m doing here.

This is my basic generalization.

Igor Babuchkin: Hello everyone, I am igor, and I am a team member from xAI. My main major is physics, astrophysics, and cosmophysics, which is what I have studied before. I think that through deep learning, the most important thing is to be able to apply the processing power of deep learning and some capabilities of machine learning to aerospace and understand the mysteries of the entire universe. This is also the key element for everything we want to do. This is a timely project, which is a completely different project from OpenAI. We can also apply the big language model to the data analysis of the universe. I believe that such data can give us a deeper understanding of the universe and the world.

**Musk: **Physics is a way to expand the margins of our entire world. Through physics, we can understand the basic principles of the universe and make some decisions. So computer science should be more closely integrated with physical science and astrophysics, and explore the intersection of these three different sciences. Because these are some information sciences, we want to be able to understand the real world. Now we hand over the mic to Mantal, also from xAI.

Mantal: In the past six years, I have been working at DeepMind, mainly responsible for the training of the entire AI learning model. I have studied AI for a very long time. My purpose is to further explore some of the latest developments, understand some directions of AI, and have various tools and topics to deepen my research and exploration. xAI is a tool we choose to learn about various latest developments through research and share our latest developments with researchers. Now I hand the mic over to Joy.

Joy: I am Joy. I have been working on AI mathematical models for the past seven years. I hope this AI can be good at solving mathematical problems, because I think mathematical language is a logical language, a mathematical language and advanced programs. It is to show the most beautiful progress of AI, and it is also the most advanced kind of AI. We have been working with programming and physics for a long time, and this is also a key step in our understanding of the universe. Then hand the mic to Tony will.

Tony will: Hi, I’m Tony, thanks for your questions. My dream is to use my whole life to solve the most difficult problems in the world, so the xAI project itself is a relatively ambitious project. We made some progress last year, which means we are getting closer and closer to our dream.

Now I have introduced some of your past work experience to you. Perhaps Musk thinks that you should further introduce the work you did last year. Last year, we made more progress in the field of AI. In such a team, we built an AI model, which can help us in the high school quiz and test marking. This is one of our research directions.

Because we add the principles of mathematics to the process in such a design process, this gives us the deepest understanding of logic and basic science. Now we hope to be able to apply these tools in the form of the network, and we will soon see the dream come true. This is Toby Pohlen, let me introduce myself.

Toby Pohlen: I am from the University of Toronto. You may have taken classes at the University of Toronto. I mainly study AI data models, and my research directions are all about deep learning. I’m basically uncovering the other side of every stone, the feeling of every stone on the coast, that’s what my work feels like. In the past few years, I think we have had a revolution in deep learning.

Musk: Our entire xAI team is very strong and has a good background. We hope to be able to solve some of the most basic, but also the most ambitious problems. The tools I use are to further strengthen myself and empower everyone, and I am willing to join such a new journey. Now hand the mic over to Toby.

Toby Pohlen: Hello everyone, I am Toby, I am an engineer, I am mainly responsible for writing code, the code is mainly about machine learning. I joined DeepMind many years ago, and my research direction is mainly about deep learning. I hope that such technology and interface can be truly applied to solve the problems existing in different groups of people in society. Able to solve some current problems and issues. Now I hand over the mic to Kyle.

Kyle Kosic: Hello everyone, I am a systems engineer, just like my colleagues, and my entire professional background is (the same). In the field of algorithms, a few years ago, I worked at Ostyle, and recently I worked at OpenAI, where I was in charge of the GPT project. The reason why I am particularly interested in xAI is that I think the biggest threat to AI technology is that a certain company may cause a data monopoly, and the power of capital may eventually obtain benefits that are anti-human or anti-human. I think xAI is We provide a brand new opportunity, especially for AI practitioners to use AI to further explore the mysteries of science and engineering. It will not be affected by political background factors, nor will it be affected by national conflicts. We just explore It’s just the boundaries of science, so I also have two friends from OpenAI who also joined xAI at the same time. this is my friend.

Greg Yang: Hello everyone, I am Greg, and I am in charge of mathematics and algorithms. My connection with AI started ten years ago. I am particularly good at mathematics and algorithms, and I have done a lot of work related to mathematics and algorithms. After I was admitted to university for two years, I didn’t want to stay in the same basket like a “little hamster”, so I left school and became a DJ. Then I went back to school again and finished my studies all over again. Because of such an experience, I have a deeper understanding of myself and the whole world. I have also thought about many questions, such as how can quantum physics have a big impact on the real universe? And to what extent can the algorithm affect if our data sample size is large enough? In the past so much time, I gradually realized that the next step is to make AGI a reality, to make AGI smarter than us, and to contribute to the basic reality of myself, which is the source of all my motivation.

I also gradually realized that mathematics is the summary language and logic behind everything behind reality and science. If we want to really make progress, we must understand the logic of mathematics. So I basically learned math from scratch, the first two books I read were basic math, and we started from scratch. There are also non-linear algorithms. Later we went to study topology and Java. Our goal is to be able to produce such a language that can talk to all mathematicians around the world and understand the logic behind the entire mathematics.

I got a research job and have been working in Microsoft Research for the past five years. Their work is a fundamental contribution to our understanding of how the world works fundamentally. The most basic content of my own work is also very complicated. It is mainly about the entire universe and how it should be explored. After that, I myself was able to draw a relatively simple conclusion, which is to apply a large-scale data model to the universe. This can also further empower the models and algorithms of the universe, allowing us to truly understand how some satellites and planets in the universe operate. This creates a lot of interesting data models for us.

Looking to the future, I am particularly confident and confident in xAI. I think we are not only constantly approaching AGI, but from a scientific point of view, we have also truly completed the changes in neural network science and physics. At the same time, we have also combined quantum physics with the entire AI. The entire data model is beautiful.

Guodong Zhang: I am very grateful that I have such an opportunity. I am very happy to create such an AI, because it will be the beginning of new mathematics and new science, and we will truly realize augmented reality, thank you!

Moderator: The next one is Model.

Model: Hello everyone, I am Model, and my entire research foundation is also the role of xAI. On top of this, I will think of various projects. Now I am very excited about such a project, because I think AI can truly empower our generation of scientific research, so our job is to ensure that such a progress process is very fast, and what everyone wants to achieve We should also understand the laws of operation of such a universe. This is my brief introduction and my vision for xAI.

Team members: Hello everyone, my entire academic background is in deep learning, I have ten years of experience in the industry, and I am very happy to join the team. My main background is in database management and data analysis, and I apply real scientific techniques to such aspects. I believe in the importance of working together. We want to be able to obtain the most information, mine the deepest data, and address the most challenges. I will manage your data operation as the person in charge of the database. I am mainly concerned with the construction of data systems. I am responsible for the construction of the US national database. This work direction is the same as that of xAI, which is all about database construction. It’s also a cool system. I also like news, I like to study physics, astrophysics and the universe, I think it can solve some problems that science can’t solve so far, and answer these questions. This is my vision for xAI.

Musk: Everyone has said some of their own prospects. If you have any ideas, you can turn on the microphone and continue to contribute your ideas. The main purpose of our establishment of xAI is to have a deeper understanding of the nature of the universe. We have understood a lot of the universe now, but we should also combine the universe with reality. There are still many unsolved mysteries in the universe. These are all is the key to our work. We have a standard model that can be applied to this, but there are still many questions about nature itself, the universe itself. We now have one of the biggest paradoxes, where are the aliens? This also involves our operation of the algorithm. Because we can calculate where the aliens are by calculating the data of different outer spaces through the algorithm of different planets. But so far there isn’t any computational progress in this area. This means that we have many different explanations for the current situation in the universe, but it is impossible to say which of these explanations is true and which is false. We have other paradoxes, we see some new elements, and we hope to find some new elements in the universe. Perhaps, we are the only company that focuses on the galaxy, and everything we do is relatively relevant. We should pay attention to the birth of new elements in such cosmic explorations.

The earth is now 4.5 billion years old, and the environment of the earth will further deteriorate in the future, and the water in the ocean will be completely burned. We still need to build such an escape channel, and we should consider such escape channels. We don’t need to expand to the entire solar system, we just need to find the most suitable living environment for us, which requires a lot of algorithms. If we take out 10% of the earth’s resources to find the same planet in the universe, we will find that this is very meaningful. We will always have the problem of immigrating to the universe, and human beings also have many unsolved mysteries, so we hope to solve these problems through AI’s truly large-scale data computing capabilities, which is the future direction of AGI. We also need more solutions.

Greg Yang: We can set up xAI and let xAI make some commitments to let its development (along) the direction it was designed for. I think what we are talking about is not just about technology, but everything we do is in the hope of better understanding the nature of science. Our key mission is to deepen understanding. What I want to add is that we hope to build such a machine to better understand the universe at the beginning, but from my point of view, from a mathematical point of view Large-scale calculations are also possible. I can still see that some large-scale computing methods here can allow us to rethink basic physics and reality. For example, large-scale neural network systems can solve the current spaceflight missions and help us solve advanced physical problems. .

Team member: One thing we still need to understand is that data can also help us solve some current problems, such as understanding what really happened through theory. This is another understanding, and we also hope to have a deeper understanding of the entire universe and the world. After understanding the universe, we will be able to understand all people, create better real conditions for people, and produce a series of products and tools to help people better understand the world.

Musk: Yes, this is also one of our work directions.

Team member: Let’s talk about the supply of energy. If we have a GPa of 10 megawatts, we can write better novels and come up with a better power system. We should think about such issues, in fact, these are all about order. We should also consider two different ways of contact, or the way of adding a converter in the middle, or like the synapses of neuron cells, the two are combined through electrical signals. I also want to ask such a question, what is the driving force behind the development of all science and technology after such technical problems?

Musk: But these questions are not very helpful. We hope that the direction of machine learning application is a way that everyone is interested in. The current AI development is not the direction of AI research and development that we wanted before. My own experience and experience in creating Tesla, I want to solve problems in real life and transform existing energy forms. On a larger scale and in a wider field, this result is more helpful than ours, and this is what we hope AGI can achieve.

In the development of the field of computer science, in terms of data analysis, in fact, asking questions is more important than solving them. It is also true that once we can ask the question, the answer becomes very easy. When we are in the progress of xAI, I feel that asking a question and asking a task is more difficult and complicated than completing this task. Once the problem of AGI is solved, we can review the logic of some previous problems.

If we go back to 2022 now, we will find that many problems that were difficult at the time will become very simple now. These are some problems we need to prepare for. We need an AGI platform, our workload is huge, but we need to work more intelligently, and we need to work from the perspective of understanding.

This is the basic introduction of the project. This is some computer resources we invested in the project. I hope this project will be successful. This is a very small team, which brings together the best talents and the most resources, hoping to solve some of the current problems. We hope to solve the most basic models and problems, hope to increase the per capita computer ownership rate on the earth, and enable digital technology to change the current people’s work direction, about people’s work planning and future human development. , so such a way is going to last. This is an interesting direction to work.

Another very interesting direction of work is the problem of people’s per capita energy supply. How to optimize the per capita energy supply structure is a very good way. We should take a new look at energy supply and power supply projects. If we consider the per capita energy consumption of each person, this figure is very interesting, and we should also add imagination to such a process to understand the current world. Now everyone’s per capita energy is very low, including electric energy and geothermal energy. And we hope to further increase energy per capita. It sounds like we have a lot of work to do, but in fact, the work load will become very little after building this platform.

As a relatively good team, we are very concerned about the amount of work that each person can complete. Resources are limited, and we want everyone to make progress. At the same time, we also hope to obtain the most progress as soon as possible, and then we can share information in this field. This is our work direction.

*The text record is arranged according to the Chinese simultaneous interpretation on the spot, and the above names are transliterated