Since OpenAI released chatGPT last year, after about half a year of brewing, AI has once again ignited people's attention and topics, and has also brought a stimulus to stocks in related industries. As the national team of voice AI intelligence, HKUST Xunfei has accumulated more than ten years of AI technology and voice technology. After nearly half a year of research and development, it launched the "Spark" model, and released it to the public on May 6. .
Previously, the perception of AI was still in the foolish way of interaction. I always thought that it was difficult or even impossible to break through the foolish way of interaction, because it was beyond my imagination. A cold How can a machine have logical thinking, and it is a thinking with strict logic like human beings.
But after chatGPT came out, to be honest, I was shocked. This might be the impossible AI intelligence I imagined. He can really have a conversation, and he can answer any question in an orderly and thoughtful way, instead of the purely foolish answers of "what's your name?", "my name is Xiaozhi". But this is not what shocked me the most. As a programmer, he can help you write codes, modify codes, and implement sql functions. This is shocking and terrifying.
I don't know how other people feel about chatGPT or other products with similar functions. From my own feeling, the era of AI intelligence is really coming. It is not like the previously hyped blockchain, VR, Bitcoin, etc., which are far away from us ordinary people, and we don’t even know how to use them, but chatGPT, or a similar large model, is around us and uses us You can communicate with him by registering an account on your own computer, and he has indeed come to us ordinary people. As my title says: AI is finally within reach.
1. Have you ever used this conversational AI? What are your thoughts or feelings about this type of AI?
Yes. Since chatGPT was released last year, everyone in the Internet circle has been discussing it enthusiastically. My company has also developed some services, calling the interface of chatGPT, and then providing them to the company's personnel for use.
To be honest, I was really shocked. As I said above, he completely washed away the fact that I thought AI could not have strict logic. Colleagues also enjoy using chatGPT for daily work assistance. As a big data engineer, when I encountered codes that I couldn’t write, I used to search on csdn or other websites, and then I had to check it myself, but chatGPT will return you the answer directly, so you don’t need to search again. Saved a lot of time.
For myself, after such a period of use, I deeply realized that the era of AI intelligence may really be coming. Using this general big data model, coupled with some interfaces for special scenarios, it may not only be used in Our Internet circles are buzzing, and he will be present in every aspect of society. Just like Xunfei's integrated applications in education, medical care, and cockpit at the "Spark" conference, it is actually embedded in our daily life.
And I have a hunch that when the big model understands human semantics to such an extent, the era of robots that people envision will slowly enter our society. This kind of robot is fully capable of realizing the main functions that people imagined robots can accomplish. : Understand the meaning of people's speech and give precise feedback, even physically execute the commands given by people. For this aspect, we will wait and see.
2. What is your attitude towards the proposition "Xunfei Xinghuo's large model will surpass chatgpt?" Briefly explain why
I look forward and look forward to it.
People always say that Chinese Internet companies can only do one thing: copy. Even copying is just copying the upper-level application, and the underlying infrastructure does not even know how to copy. In fact, from the perspective of the development of the Internet, we cannot deny the fact that, to some extent, the Chinese Internet often has a low self-esteem.
But as far as I know, Xunfei is really starting from scratch to develop China's own voice industry, and has become the industry leader. Liu Qingfeng, chairman of Xunfei, is a passionate entrepreneur. Every time he listens to His speeches are passionate and tireless. I have a lot of admiration for these entrepreneurs who have really gone through all kinds of difficulties and hardships.
As a new infrastructure in the future, the large model is important in terms of national significance, and it is also important for passionate entrepreneurs like Liu Qingfeng. They will not waste such a once-in-a-lifetime opportunity. They spent 6 months It can be seen from the "Spark" large model. Combined with their technology accumulation over the past ten years, we have reason to believe and should believe that "catching up with or even surpassing chatGPT" will not be an impossible task. So I look forward and look forward to it.
3. What significance do you think this type of artificial intelligence has for today's society?
Significant.
As I said, one of the reasons why I was noncommittal about AI before was AI’s ability to understand semantics. How to make machines understand complex human language and text expressions has always been a difficult problem. We used to laugh at foolish questions and answers. It is said that it can be realized by changing the question mark into a period and removing the "do, eh". But now, chatGPT has completely made a qualitative breakthrough in its understanding of semantics. All kinds of affairs in people's lives do not need to understand human language or text expressions, and solve the problem of semantic understanding, and many artificial intelligence in the social sense will be easily solved.
In addition to the education, medical care, robots, etc. mentioned above, people's lifestyles, living content, working methods, and work content will all undergo subversive changes, just like the changes to human social life when the Internet rose more than 10 years ago. Same.
It’s just that our group of people born in the 80s and 90s in the Internet age will be at a loss, expecting and at a loss when facing the AI era like our parents born in the 60s and 70s! This is a very real problem. The development of human society is moving forward, and new things are emerging one after another. For us, whether we still have the energy to learn from him and understand him is actually a question mark.
When such a powerful function of AI takes away your job, can humans still look at AI rationally, and even the conflict between AI and humans is triggered, and finally comes to the fore, what kind of scene will it be?
People in the world only hear the newcomers laughing, but don't care about the old ones crying!
4. For this type of artificial intelligence, what do you think is the future prospect?
As I said above, although some people have warned against the development of AI, including Musk. However, personal strength cannot hold back the torrent of the times. People’s fear of unknown events is far lower than human curiosity about unknown events. The era of AI will inevitably come. Judging from the current development situation, the benefits are: Self-evident, the downside, only time will tell.
To make a digression, I never thought that the era of AI would come so soon, just like planetary immigration, which exists in imagination. With the failure of Musk’s starship rocket, I am more convinced of planetary immigration. will come early.