On January 12, local time, Microsoft founder Bill Gates' podcast "Clear My Questions" invited a hot and heavy guest: Sam Altman, CEO of OpenAI. Two leading figures in the science and technology community exchanged views on "the impact of AI on human society."
▲Poster of this talk show. According to Bill Gates’ official WeChat public account
, both Gates and Altman agreed that the development speed of AI is too fast and is daunting to some extent, but Altman encouraged the audience to be brave in career development. Comfort zone, don’t be afraid of unemployment caused by AI.
Looking forward to the future, Altman is very optimistic. "For example, if AI can triple people's efficiency, it will not only mean that people can do three times more things, but they can have the energy to think about higher-level things." Abstract problem." He also believes that with the development of AI, human beings will enter the "post-scarcity" era. Without the constraints of resource scarcity, human beings can freely choose their lives, such as considering which galaxy to settle in.
Talking about the development of AI
The current reasoning ability is like "primary school students doing homework"
But in the future, the ability will be exponentially enhanced
Altman admitted in the program that the current gpt-4 model (the language model released by OpenAI for its chatbot chatgpt) His reasoning ability is extremely limited and he can only make the simplest reasoning similar to that of a primary school student. In the future, as reasoning capabilities increase, AI's capabilities will increase exponentially, and the entire capability growth curve will be "very steep," and "we are currently at the beginning of this continuous, long-term curve."
▲ Data map: Altman
Altman believes that logical reasoning ability is the key point that allows the function of AI to make a leap from quantitative change to qualitative change. For example, the current AI model can complete a part of many people's work tasks, but it is still difficult to completely replace a position. In the very near future, perhaps AI can complete a complete complex task independently, such as "Write a program for me." The user only needs to provide very light prompts and supplements during the process, and then the AI can complete the tasks assigned by the boss like an employee.
As AI's reasoning capabilities continue to develop, users can directly tell AI to "help me operate and manage a company" and let AI act as a professional manager. After a while, humans can directly command AI to "help me discover new laws of physics."
Altman believes that the application prospects of AI in the next five to ten years are extremely exciting. Looking back at that time, all today's AI models and application scenarios will look "stupid."
Talking about future applications
The development momentum of AI is so fast that it is "daunting"
But its progress will eventually benefit mankind
Bill Gates mentioned in the interview that the development speed of AI makes him feel "a bit daunting". "Unlike previous technological advances, AI can improve very quickly and has no upper limit. It can reach human level in many fields of work. Even if it cannot do unique scientific research, it can also make customer service calls and sales calls. I think We do have some concerns - although that's a good thing - but it's going to force us to adapt faster than ever before," he said.
In this regard, Altman said that he has confidence in human beings' adaptability. However, he also admitted that in the past, great technological revolutions often "eradicated" original types of work on a large scale within a few generations, but at least society had relatively ample time to digest and absorb the shock. Nowadays, AI is advancing at an unprecedented speed, "which makes me feel a little scary."
▲Program Live
However, Altman remains optimistic about the future. He believes that AI technology will eventually benefit mankind, especially improving the lives of the poor. . Energy costs and "smart costs" are the two largest cost expenditures in the material world, and the development of AI can make smart costs infinitely closer to zero.
He took OpenAI's model development as an example: "Gpt-3 is the model we have launched the longest. In the more than three years since its launch, we have reduced the cost by 40 times.As for gpt-3.5 version, we have reduced its cost by nearly 10 times. gpt-4 is a new product, so we don't have that much time to reduce costs yet, but we will continue. I think we have the steepest cost decline curve of any technology that I know of, better than Moore's Law for chips. "
Gates also agreed that the reduction in costs will greatly increase "the fairness of technology" and allow more poor people to improve their quality of life.
Looking forward to the "post-scarcity" era
there will always be new things to do
Human beings will never Lack of ways to obtain satisfaction
In addition to the capabilities and application scenarios of AI, Gates also took the initiative to talk about his confusion about "the meaning of human life" in the show. In this regard, Altman believes that when AI technology is mature enough, If it can help people solve the constraints of lack of resources, then the way human society is organized and the purpose of life will change accordingly.
▲Information pictures and pictures according to Tuchong Creative
When Gates mentioned the philanthropy he was proud of, he said, "I Very good at researching malaria and eradicating malaria, and very good at gathering smart people and devoting resources to it. If one day an AI robot says to me, ‘Bill, you go play pickleball, I can eradicate malaria, and you are just a slow-thinking human being,’ then I will be faced with a philosophically puzzling thing. How should we organize society next? What is the significance of our commitment to education in remote areas? "
Altman extracted the concept of "scarcity" from Gates's dialogue. He said: "Our thinking is so dependent on scarcity. In life, we know good teachers, good doctors and good entrepreneurial ideas. It's very rare and work hard for it. I do wonder, if an entire generation in the future grew up without this kind of scarcity, what they would think about the philosophical question of how to organize society and what humans are to do. Maybe then they'll come up with a solution. I'm worried that my thinking has been so influenced by the scarcity framework that it's hard for me to even think rationally about this issue.
Altman added: "Although we are giving up something in a sense, we will have something smarter than us humans." If we could enter this post-scarcity world, we would find new things to do. Maybe instead of solving the malaria problem in the third world, you're choosing which galaxy you'd like to live in. I believe we will never lack problems, we will never lack ways to achieve satisfaction, and I think the only way out is to keep going. We must do AI, it will definitely happen, and it is now an unstoppable technological process because its value is too great. I'm very, very confident that we're going to make it, but it's going to feel really different then. ”
Red Star News reporter Zheng Zhi
editor Zhang Li editor-in-chief Li Binbin
